Article

6min read

Conversion is a conversation: how GenAI is reshaping e-commerce and travel

Generative AI is now creating a massive shift in how online buyers discover and compare products. That’s why we’ve put together a roadmap to help you understand the 2026 consumer in our e-book, The Spontaneous Shift: Consumer E-Commerce Trends for 2025. For e-commerce and travel brands this represents a real paradigm shift in how digital marketing works; one they will need to adapt to, and fast, if they want to survive.

Meet your customer’s new shopping assistant

While Google search is still the dominant force in discovery, our research shows that other channels are rapidly gaining ground. And the fastest growing channel of all is Generative AI (like ChatGPT, Google’s Gemini, or Claude from Anthropic). Use of Gen AI tools by online shoppers in the discovery phase has grown 75% in just the last year, rising from 8% in 2025 to 18% as it quickly moves from a novelty to a utility.

But while more people are choosing to use AI, the level of uptake differs across generations. Perhaps not surprisingly, younger shoppers are more comfortable using AI tools to find what they need. 32% of Gen Z now say that they find AI helpful in their online buying journey. Millennials are not far behind, with 30% of them seeing AI as a helpful assistant, but this number falls to 13% of Baby Boomers.

What’s clear is that the online behavior of consumers is changing and at pace. Younger generations especially are no longer content with scrolling through a list of blue links. Instead, they’re asking questions – and expecting answers. They aren’t just typing keywords like “red sneakers” into their browser; they’re giving AI prompts, like “Find me a pair of red sneakers that look cool and are under $80.”

This is a profound shift and one that has big implications for e-commerce and travel brands. Firstly, it means consumers no longer need to visit multiple websites to compare different products and services. They can do all of that and more by simply asking questions to an AI interface: “Is this sustainable?” “Will this work for my specific use case?” “How does this compare to their competitor’s offer?” The transaction is quickly becoming a conversation.

What this signals is a move away from carefully crafted product detail pages (PDPs) and towards much more dynamic interactions. If your brand can’t answer these questions in the conversational spaces where they’re being asked, either an on-site AI chatbot or a third-party platform, you’re effectively invisible. You don’t exist in an online consumer’s decision-making loop.

The new frontier of optimization

That in turn presents brands with a unique challenge. When online buyers perform a Google search or search directly on your website, you can see their query. But when they ask a question to a third-party AI platform, that happens inside a metaphorical “black box”. You can’t see what question they asked, and you can’t (currently) buy an ad to influence it.

This “invisible intent” requires a new kind of optimization. Because you’re no longer just writing for humans or typical SEO algorithms; you’re also writing for the Large Language Models (LLMs) of Generative AI. Fortunately, the core SEO principles of E-E-A-T (Experience, Expertise, Authoritativeness, Trustworthiness) can still provide brands with a competitive edge.

This is because LLMs are designed to recommend what they interpret as being the most credible answer, not just the one with the most (or best) keywords. If you want GenAI to recommend your products or services, you’ll have to give it proof that your brand is the authority.

To stay visible in this conversational future, that means your content strategy needs to move from volume to value. And if LLMs are to place any trust in your user reviews, they’ll need to be authentic. You also need to ensure that product data is structured with schema markup so it can be “read” by these new gatekeepers. It needs to be rich enough to answer the nuanced questions of online shoppers, not just fill a spec sheet.

Changing the face of travel

Human interaction has always been at the heart of the travel industry. And travelers have traditionally been happy to receive personalized recommendations or resolve any issues they might have by speaking with a real person. But this change in online buyer behavior from keyword search to asking questions of GenAI means the travel sector is perhaps even more exposed than most to this disruption. 

Our research shows that 36% of travelers have now used AI-powered tools, chatbots, or virtual assistants to book travel or solve issues and found them helpful. Another 32% haven’t tried using these yet but are open to doing so. Again, Gen Z are the most open to AI, with 49% saying they have used AI tools and found them helpful when booking travel. And while this figure is 41% for Millennials, it falls to 21% of Baby Boomers.

Focus on quality

This is good news for travel brands looking to embrace on-site AI chatbots to help visitors plan travel. But at the same time, there’s also nothing to stop those same online buyers from doing that on a third-party platform. A prompt like “Plan a 5-day family-friendly trip to Lisbon in May, staying near the city center with a budget of $2000,” can generate a complete itinerary with flight suggestions, hotel options, and activity booking links. All without visitors ever going near a brand’s website.

That means there’s a real possibility for travel brands of being reduced to a single item in an AI-generated plan. And if they lose the ability to showcase a unique brand experience on their website, travel companies like hotels and airlines risk being compared merely on price and basic features alone. Loyalty programs and creating exceptional customer experiences are likely to be more important than ever.

Travel is also time sensitive. A third-party AI platform is unlikely to recommend a hotel or a flight if it can’t access real-time availability, accurate pricing, and clear policies. This requires travel companies to be laser-focused on the quality and accessibility of their data. Again, the best way to ensure that GenAI recommends your brand to online buyers is to have rich, well-structured information that is instantly available.

Conclusion

The world of e-commerce and online travel is undergoing a real paradigm shift in how online buyers discover and compare products. And the rules are still changing in real-time. Ensuring that your data is as authoritative as possible and easily accessible to GenAI is currently your best bet to remain visible in this new environment. The only viable way to do this is to build a culture of continuous experimentation; testing, learning, and iterating towards your “better”.

Takeaways for e-commerce and travel brands

  • Test your GenAI visibility: Go to ChatGPT or Gemini and ask questions about your product category. If your brand doesn’t show up, you have an optimization gap to fill.
  • Develop AI-friendly product data structure: Implementing comprehensive schema markup across your site is no longer optional, it’s the price of entry.
  • Create AI-friendly content: Develop clear, factual, easily digestible content that directly answers questions your customers might have. This makes it much more likely that it will be used by an LLM.
  • Track GenAI adoption monthly: Monitor the amount of traffic to your website coming from third-party Gen AI tools and how this affects sales.

Article

4min read

Real-Time Personalization in 2026: How to Meet Customer Expectations

Personalization has long been a goal for brands looking to create more relevant and engaging experiences. Traditionally, this has meant using static rules and segments – showing different content to new visitors, returning customers, or VIP shoppers.

While this approach has value, it often falls short when customer needs and interests change quickly.

Why Real-Time Personalization Matters More Than Ever

Today’s customers are unpredictable. They jump between devices, change their minds mid-session, and expect brands to keep up. It’s very common for shoppers to change their intent during a single visit.

For example: a user who starts out searching for a small purse for everyday fashion [category: bag] might begin looking for black sunglasses for an upcoming beach trip [category: accessories] – all in one session. Real-time personalization adapts to user needs in the moment.

Traditional personalization is based on fixed segments and rules.

It can’t keep up with rapid shifts in customer needs and preferences. Traditional personalization treats users the same way throughout their visit, missing crucial signals and opportunities to engage.

The Challenge: Static Personalization’s Limits

Most brands have invested in standard personalization. It’s a great foundation, using segments like “new visitor” or “VIP” to tailor experiences. But static rules are slow to react. They’re built on who a customer was, not who they are right now or how they are behaving.

The result? Missed chances to upsell, cross-sell, or simply delight your customers with relevant content. In a world where attention spans are short and competition is fierce, that’s a risk brands can’t afford.

The Solution: Real-Time Personalization with AdaptiveCX

Real-time personalization is about responding to customer intent as it happens. Instead of relying solely on historical data or static segments, real-time solutions monitor user actions (clicks, searches, time spent on pages) and adjust the experience instantly.

So, how can brands personalize in real time in 2026? The answer lies in solutions like AdaptiveCX – technology designed to sense, interpret, and respond to customer intent as it happens. Adopting real-time personalization doesn’t mean abandoning what works. Think of it as evolving your strategy – layering real-time capabilities on top of your existing efforts.

Here’s what real-time personalization looks like:

  • Continuous monitoring: Every interaction is tracked as it happens.
  • Intent recognition: The system identifies when a user’s interests or goals change, even within a single session.
  • Instant updates: Content, offers, and recommendations are refreshed on the fly to match the user’s current intent.

Example:
Let’s say there’s a user who has a typically of using one brand to rent a car for business travels and now this user wants to use the same brand to explore family vacation rental options. With real-time personalization, the website immediately shifts to highlight family-friendly deals and relevant upsells, ensuring the experience stays relevant throughout the session.

How to Personalize in Real Time in 2026

  1. Use behavioral data: Go beyond static segments by analyzing real-time actions – what users are clicking, searching, and viewing right now.
  2. Integrate across channels: Ensure your personalization engine works seamlessly on your website, mobile app, email, and even in-store.
  3. Prioritize intent: Focus on what users are doing in the moment, not just who they are or what they did in the past.
  4. Test and refine: Continuously experiment with different approaches to see what drives the best results.
  5. Respect privacy: Be transparent about how you use data and give users control over their personalization preferences.

The Impact of Real-Time Personalization

Brands that implement real-time personalization see clear benefits:

  • Higher engagement and conversion rates
  • Increased customer satisfaction and loyalty
  • More effective upselling and cross-selling
  • A stronger competitive position

Moving Forward

Personalization in 2026 is about more than just knowing your customers. It’s about understanding and responding to what they want in the moment. Real-time personalization helps brands stay relevant, capture more opportunities, and deliver experiences that truly resonate.

Ready to see how real-time personalization can work for you? Explore our resources, request a demo, or connect with our team to learn more about AdaptiveCX and how it can help you deliver what your customers want, instantly.

Article

6min read

Cleared for Takeoff: A Guide to Airline Website Optimization

An experience that starts with turbulence won’t sell a smooth flight. The journey begins on your screen, not the runway. Long before the boarding call, before the stress of the security line, and before a passenger ever experiences the comfort of their seat, they experience your website. That digital interaction is the true start of their journey with you, and your first opportunity to impress your brand ethos upon them.

For the modern traveler, a clunky, confusing website isn’t just an inconvenience; it’s a red flag. It creates friction and anxiety at a time when they are looking for reassurance and ease. The experience you provide online must be a direct reflection of the seamless, efficient, and hospitable service you provide in the air.

By thinking holistically about the digital experience, and optimizing your website to match user expectations, you can turn a stressful process into a smooth takeoff, converting browsers into loyal passengers. 

Your brand experience begins before the runway

The core challenge for any airline is to translate the feeling of a well-serviced flight into a digital format. When a passenger is onboard, your team is trained to anticipate needs, offer comfort, and provide clear information. Your website must be held to the same high standard. Every page, every button, and every form is a touchpoint that either builds confidence or creates doubt.

Think of your homepage as your virtual gate agent, welcoming travelers and guiding them effortlessly to where they need to go. Is the booking engine intuitive? Is information about baggage fees easy to find? Can a passenger change their seat without a headache? These aren’t just usability questions; they’re brand questions. A frictionless digital experience tells your passengers that you value their time and are committed to making their entire journey, from booking to landing, a positive one.

girl on her phone

Business or pleasure? Crafting journeys for the ‘why’

Not every passenger arrives on your website with the same mission. A business traveler flying to a conference has a vastly different intent than a family planning their summer holiday. The first needs efficiency and directness; the second is looking for inspiration and value. A one-size-fits-all website serves neither of them well.

The power of tailoring the journey can be seen in the work we did with Evolve Vacation Rental. They recognized that a user arriving from a high-intent Google search (“how to rent my vacation home”) was ready for a different conversation than a user who clicked a brand-aware ad on Facebook. Working with us, they tailored their landing pages to match that intent, and spoke to each user’s specific needs.

Airlines can apply the exact same logic. A visitor searching for a specific flight route and date is on a mission. Their journey should be streamlined, with clear calls-to-action for booking and relevant upsells like seat upgrades or lounge access. In contrast, a visitor arriving from a broad search like “flights to Spain” is in a discovery phase. Their landing page should be rich with inspirational content: fare sales, destination guides, or travel ideas. By segmenting your audience by their ‘why,’ you can create a more relevant and effective journey for everyone.

Make self-service a first-class experience

Flying comes with a checklist of necessary tasks: booking, checking in, amending details, and adding extras. The more of these actions a passenger can complete themselves, quickly and easily, the less stressful their experience will be. Your goal should be to make your website’s self-service functionality a first-class experience.

This often comes down to simple, intuitive design. The travel agency Iberojet, for example, questioned whether the tabs on their homepage were in the most logical order for their users. By running a test and switching the order of their “Holiday Packages” and “Travel Circuits” tabs to better reflect user behavior, they increased clicks on the “Search” button by a huge 25%.

Similarly, consider the lesson from SuperShuttle, who found that their social media icons weren’t getting much engagement. By redesigning their footer to make the icons more prominent, they saw a massive increase in traffic to their social pages. Sometimes it really is as simple as making things clearer.

 For airlines, the takeaway is clear: make your primary calls-to-action for essential tasks like “Check-In Online” or “Manage My Booking” unmissable.

Turn loyalty from a program into a perk

For many passengers, the true value of an airline loyalty program isn’t just about accumulating miles for a future trip. It’s about the immediate, friction-reducing perks that make the travel day smoother: priority boarding, preferred seat selection, or an extra baggage allowance. The challenge is that the sign-up process itself can often feel like a point of friction.

To encourage sign-ups, the process needs to be as seamless as possible. The travel company OUI.sncf wanted to increase account creations and realized that the traditional sign-up form was a barrier. By running a test and adding a social login option via Facebook, they made the process significantly easier. The result was an 18% increase in account creations and a 9% lift in transactions from logged-in users.

Once a passenger has an account, the real value emerges. By securely saving key details like passport numbers, frequent flyer information, and meal preferences, you can transform a ten-minute booking process into a two-minute one. This is how loyalty moves from being an abstract program to a tangible, valuable perk that passengers appreciate every time they fly.

From lost connection to brand connection

Even on the best-designed websites, users can sometimes get stuck. A complex fare rule, a confusing payment page, or a simple user error can lead to frustration and an abandoned booking. The most forward-thinking airlines anticipate these moments and have a plan to turn that potential friction into a positive brand interaction.

We saw this principle play out perfectly with travel agency Havas Voyages. Their team implemented a smart strategy using exit-intent technology. When the site detected a user was about to leave a page, it didn’t just let them go. It proactively triggered a pop-up that offered an appointment with a human travel planner at the nearest physical agency. They saved the lead by offering a different path forward.

For an airline, this principle can be modernized with live chat or AI-powered chatbots. Imagine a user hesitating on the payment page. A small chatbot window could appear, asking, “Having trouble? I can help with common questions about payment or baggage.” This automated, helpful intervention can solve problems instantly. For more complex issues, the chatbot can escalate the conversation to a live agent. This transforms a moment of potential failure into an opportunity to provide excellent, real-time customer service, reinforcing your brand’s commitment to a stress-free journey.

Curious about how your website can be better optimized for people? We’re here to help you build up brave ideas into brilliant results. Let’s talk about what you want to achieve.

Article

4min read

What’s New in Feature Experimentation & Rollouts?

We’re continuously improving Feature Experimentation & Rollouts (FE&R) to make your experience smoother, faster, and more reliable. Here are the latest updates now available:

Data Explorer: Now available in FE&R

The Data Explorer, already available for Web Experimentation, is now accessible in Feature Experimentation & Rollouts. You can now explore and export your server-side data directly from the platform, without relying on external or technical tools like Postman.

With FE&R Data Explorer, you can:

  • Query and export metrics (conversion rate, revenue, custom goals, NPS, action counts, etc.)
  • Access and export raw hits (SDK events, API calls)
  • Apply filters and dimensions (campaign, environment, feature, user properties, time range, etc.)
  • Define custom date ranges and result limits
  • Generate API payloads directly from the UI

This feature was one of the most requested improvements by our clients. It increases transparency, improves trust in data collection, and brings stronger consistency between client-side and server-side capabilities within One Platform.

👉 Access Data Explorer directly from your FE&R interface.

Live Hits: Real-Time QA for Server-Side

When implementing or debugging a campaign, waiting for aggregated reports can slow you down. With Live Hits, you can now visualize incoming SDK events in real time and instantly verify that your implementation is working as expected.

No more guesswork. No more waiting.

Live Hits significantly improves QA workflows and reduces implementation friction for technical teams.

You can quickly confirm:

  • That events are properly triggered
  • That traffic is correctly allocated
  • That variations are being served as expected

👉 Access in Reporting → Button next to the filter.

Server-Side Reporting: Evolution

AND / OR filter operators

You can now switch between AND / OR operators for multiple values within the same filter type.

This gives you much more flexibility when segmenting your data and running deeper analysis directly inside your reports.

More precise segmentation means more accurate insights.

RevenueIQ*: now available for Server-Side 💰

*RevenueIQ is AB Tasty’s patented statistical engine that transforms experiment results into clear, reliable revenue projections — helping teams move from CRO to true Revenue Optimization by confidently quantifying uplift and ROI for every variation.

RevenueIQ, already available and widely adopted on client-side experimentation, is now live for server-side reports.

You can now project the financial impact of your experiments directly from your server-side reporting.

Available in any report with a transaction goal under the “Revenue stats” tab, RevenueIQ provides:

  • Uplift per visitor: estimated revenue increase per visitor once deployed
  • Uplift per month: projected monthly revenue impact
  • Confidence intervals: lower, median, and upper revenue scenarios
  • Revenue chances to win: probability that the variation will increase revenue

This means clearer decisions, fewer inconclusive experiments, and a direct view of real business impact.

You can confidently quantify and share the ROI of every server-side experiment.

Conclusion

These latest enhancements to Feature Experimentation & Rollouts are designed to empower your teams with greater transparency, flexibility, and actionable insights. With real-time data access, advanced reporting options, and robust revenue projections, you can confidently optimize your experiments and drive measurable business impact.

Article

5min read

Reaching New Experimentation Heights: How Air Europa Put Mobile First

Air Europa wanted to both raise the bar for customer satisfaction and improve business performance. The solution? Putting mobile at the heart if its digital experience. To do so, it partnered with AB Tasty to deliver an online buying journey that’s as smooth and enjoyable as the perfect flight.

Founded in 1986 as a charter airline, Air Europa is now Spain’s third largest carrier. As well as serving a number of domestic routes within Spain, it flies to more than 40 destinations across Europe, the Middle East, Africa, and the Americas. The airline became a member of the SkyTeam alliance in 2007.

You can find out more in our case study The Sky’s the Limit: Air Europa’s Experimentation Success Story.

Making the mobile journey easy

With increasing numbers of travelers now booking and managing their journeys entirely on smartphones, optimizing the mobile experience for buyers is a strategic priority for any travel company. But unlike the e-commerce sector in general, booking air travel is rarely a one-click affair. The travel experience is multi-faceted, and each component increases the complexity of checkout.

Air Europa wanted to create a seamless, intuitive journey for mobile that boosted both conversions and ticket value, all while scaling experimentation at speed. As Digital Customer Experience Manager, Jaume Comas, explains,

Air Europa is an airline that always keeps its focus on the client. From the fleet to the digital product, we always strive to offer the best experience to users. We’re also currently in a growth phase in which optimization is a key piece.”

To do this, the team focused on streamlining both the purchase and post-purchase experiences – ensuring that every step, from selecting a flight to managing bookings after purchase was intuitive and easy on mobile devices.

Start with data-driven insights

Working with AB Tasty, Air Europa first developed a comprehensive experimentation methodology, placing strong emphasis on both qualitative and quantitative data. Jaume Comas says,

We always start with a very clear methodology in which research is very important. We start by conducting interviews, surveys, moderated tests, prototyping, and we’ve put a lot of focus on behavioral analysis.”

The insights gathered from the research phase were transformed into test hypotheses, prioritized based on business KPIs, and systematically tested. Initially, the primary focus of these tests was optimizing the booking flow on mobile, specifically in two key areas:

1. Streamlining the booking form

Air Europa’s booking form was entirely redesigned to minimize friction and improve chances of completion. By reducing the number of steps involved, reorganizing content, and updating field styles, the booking process became clearer and more intuitive for visitors. This led to a noticeable improvement in the conversion funnel, with more travelers completing their bookings using the new form.

2. Enhancing the flight summary experience

To give visitors greater confidence in the booking process and more control, Air Europa also revamped the flight summary page. Key details, such as baggage allowance and the selected fare for each segment, are now clearly displayed. Travelers can also change flights without having to navigate hidden menus. This increased transparency resulted in a significant uplift at this step, achieving a 100% chance to win in testing.

The results speak for themselves

In 2025 alone, Air Europa ran approximately 500 tests with AB Tasty, maintaining around 18 active campaigns at any given time. These resulted in significant improvements in both user experience and conversion rates. 

Key to Air Europa’s testing strategy was using AB Tasty’s mutually exclusive experimentation feature. This enables multiple tests to be run simultaneously on the same URL without data interference. This increased the speed of implementation by up to five times, enabling faster decision-making and deployment of winning variations.

Key uplifts:

  • Booking Flow: A +9% increase in conversion from the flight summary step to the passenger form.
  • Passenger Form: A +5% increase in conversion at the passenger form stage. 
  • Customer Satisfaction Score: Improved from 81% to 87%.
  • Component Testing: On the search results page alone, 120 tests were conducted, with 40 grouped for streamlined development and deployment.
  • Implementation Speed: The mutual exclusive experimentation feature increased the speed of implementation by up to 5x.

You’ve got a partner in AB Tasty

Working with AB Tasty has transformed Air Europa’s approach to digital optimization. And by building a culture of experimentation and leveraging advanced testing capabilities, the airline has achieved measurable uplifts in conversion and user satisfaction.

But for Air Europa, experimentation is a journey, not a destination. It plans to expand its CRO methodology across all digital products, aiming for customer experiences that are even more seamless and personalized in 2026 and beyond. As Jaume Comas says,

AB Tasty is very user friendly. Everyone from our developers to product owners and designers can enter the tool and be autonomous and independent. Thanks to AB Tasty’s good work and advice, we’ve managed to establish both a very clear methodology and a highly trained, professional team to take on all the experimentation projects to come.”

Key Takeaways

  • Data-driven culture: success was based on a rigorous research process, and a clear, prioritized experimentation backlog. This ensured every test was aligned with business objectives.
  • Executive support: Leadership buy-in was key, providing the resources and strategic focus needed to build a culture of experimentation within the company.
  • Team autonomy: AB Tasty’s intuitive dashboard empowered team members of varying expertise to analyze results independently and make informed decisions.
  • Continuous optimization: Regular, high-frequency testing created a culture of continuous improvement and learning.
  • Reliable results: The confidence intervals and data reliability provided by AB Tasty’s platform meant that Air Europa could trust experiment outcomes and make impactful changes quickly.

Article

6min read

KPIs, Custom Metrics, and Evi: Supercharge AI Insights in AB Tasty

Our platform is above all designed to help you drive your business forward. That’s why in AB Tasty you can set up a range of metrics straight out of the box that have meaningful business impact. And our agentic AI, Evi, will help you analyze test results fast and provide you with valuable insights your team might otherwise miss.

We put the key in KPI

AB Tasty provides comprehensive reporting tools to evaluate the performance of your campaigns, focusing on the achievement of specific KPIs, called goals. These can be set up both at the account level and campaign level right out of the box. There are six different types of goals that you can chose to define as a KPI. These are:

Action tracking: Clicks, dwell time, iframe clicks, element visibility, scroll rate.

Page tracking: Visits to a particular page or group of pages.

Browsing metrics: Revisit rate, bounce rate, exit rate, pageviews per session.

Transactions: Average order value (AOV), purchase rate, total revenue.

Datalayer Goals: Tracking based on variables in your website’s data layer

Custom tracking: Trackers you’ve created with custom code on your website.

You shouldn’t get bogged down trying to define too many KPIs. And less is definitely more in this case. Try to identify 3 – 5 KPIs for each major business objective. If you Start with KPIs that are directly tied to business outcomes, they’ll help drive your business forward. 

You measure you with custom metrics

These out of the box metrics are a great starting point. But no two businesses are the same, and your customer journey is unique. Standard KPIs might not capture interactions that are critical for you. By measuring what makes your business unique, you get insights that your competitors don’t have access to.

In AB Tasty, you can create Custom Trackings that are directly linked to your website’s DataLayer. This lets you create personalized metrics based entirely on your unique data. You can also create custom trackers using JavaScript, giving you the flexibility to measure almost any interaction on your website. These custom tracking events can even be used to replicate goals from other platforms, like Google’s GA4.

Real-time reporting happens now

We know that speed is critical to your business. For optimization to be effective, the feedback loop between action and insights needs to be as short as possible. Real-time reporting empowers your team to spot problems or winners within minutes, not days, enabling them to make confident decisions quickly.

That’s why AB Tasty offers real-time reporting capabilities that activate automatically when you push a campaign live. During the most critical initial phase of a test (the first 1,000 unique visitors or the first 12 hours), data for each goal and variation is updated every five minutes. We also provide a Live Hits monitoring tool which allows you to track event data at any time. This allows you to make faster, smarter decisions based on up-to-the-minute data.

Test with Confidence

Understanding the story your statistics are telling you is obviously crucial to making the right decisions. At AB Tasty, we use a Bayesian statistical model that provides you with intuitive, actionable results you can understand. We use two key figures to help you make more confident decisions:

  1. A 95% confidence interval: This gives you a likely range for the true value of a gain. If the confidence interval for the gain is [2% 8%], we are 95% confident that the true uplift from this variation is between 2% and 8%. The remaining 5% represents the margin of error.
  2. The chance to win: This is a direct probability that tells you how likely it is that a variation is better than the original. A 98% chance to win means there’s a 98% probability that the variation is the true winner.

Move at the speed of evidence with Evi

Having data and reporting from your tests is one thing, but analyzing those is another. Until recently, this has often been a time-consuming process and sometimes involved a little guesswork. But by integrating agentic AI into the reporting process, you can analyze data fast and receive valuable insights that your team might otherwise miss. 

Evi is AB Tasty’s AI-powered marketing agent designed for evidence-based decision making. It transforms your data into clear, actionable strategies for repeatable, measurable results, ensuring every step you take is grounded in evidence.

With Evi, your team can:

  • Greatly accelerate the reporting process, enabling you to analyze campaign data within a matter of clicks not hours.
  • Extract deeper insights, all driven by actual website data using built-in AI analysis.

Evi features two separate AI agents accessible from the reporting page for each campaign in AB Tasty: Evi Analysis and Evi Explore.

Evi Analysis

Tired of spending hours sifting through data tables and colorful charts and wondering what they all mean? Evi Analysis will analyze your campaign data and deliver clear, actionable insights. It highlights winning variations and breaks down why they drive transactions so you can feel confident in your next move.

Simply enter your questions and Evi Analysis will process the underlying metrics, statistical significance, and objective performance to deliver clear, concise answers backed by campaign data. All in a matter of clicks. Use case examples might include:

  • Explaining the winning variation.
  • Challenging your hypothesis.
  • Giving you the best CRO practices based on your campaign results.

Evi Explore

Want to know if your tests will actually drive revenue? Evi Explore, powered by our own patented metric, RevenueIQ, makes it easy to interpret the results of campaigns that use a transactional goal (i.e. a goal that tracks purchases or revenue). 

Evi Explore gives you a clear, statistically sound view of the revenue impact of each test variation before you launch. Rather than simply relying on traditional metrics like conversion rate or average order value (AOV), RevenueIQ combines these into a single metric to give you a direct view of revenue per visitor and per month. 

This means no more inconclusive campaigns, no more ‘conversion rate vs AOV’ dilemmas, and a significant reduction in ‘undecidable’ tests. Teams can now project the revenue impact of a campaign before full rollout with confidence intervals for best- and worst-case scenarios. This gives you the confidence to make faster, more profitable decisions. And because none of our competitors currently offer a comparable metric, by using AB Tasty you receive insights others won’t have.

Article

5min read

From Messy to Manageable: Organizing Experiments with Folders & Buckets

If you’ve ever opened your company’s experimentation dashboard and felt overwhelmed by the sheer number of campaigns, you’re not alone. As businesses grow, so do the number of teams, projects, and experiments running at any given time. Suddenly, what started as a handful of tests can turn into a maze of overlapping campaigns, making it tough to find what you need. 

That’s where Folders & Buckets come in. These two simple features can make a world of difference in how you manage, secure, and scale your experimentation efforts. Here’s how they work, why they matter, and some tips for getting the most out of them.

Too Many Experiments, Not Enough Organization

Picture this: your marketing, product, and development teams are all running their own experiments. Maybe you’ve got a few hundred campaigns live, or maybe it’s closer to a thousand. Either way, it’s easy for things to get messy. Important tests get buried, people accidentally edit the wrong campaign, and sometimes experiments even overlap – skewing your results or causing confusion.

This isn’t just a headache for your data team. It can slow down your whole organization and make it harder to get clear, actionable insights.

Folders: Your Experiment Filing Cabinet

Folders are exactly what they sound like: a way to group and organize your experiments in a way that makes sense for your business. But they’re much more than just a visual aid – they’re a powerful tool for access control and workflow management.

How Folders Work

  • Custom Organization: Structure folders by team (e.g., Marketing, Product), by project or sprint, by product line, or even by page type (e.g., Homepage, Checkout). The choice is yours.
  • Granular Permissions: Assign users to specific folders with different roles – viewer, editor, or admin. By default, new users see nothing until they’re granted access, minimizing risk and keeping sensitive experiments secure.
  • Flexible Access: Users can be given access to multiple folders, with different roles in each. This is perfect for organizations where people wear multiple hats or collaborate across teams.

Why Folders Matter

  • Clarity: Users see only the experiments relevant to them, reducing clutter and confusion.
  • Security: Sensitive or high-impact experiments are visible only to authorized users.
  • Agility: As teams grow or projects shift, folders and permissions can be reorganized on the fly – no need to start from scratch.

Pro Tip: Many organizations use folders to mirror their internal structure, but you can also get creative – organize by campaign type, business objective, or even experiment status.

Buckets: Keeping Experiments in Their Own Lanes

While folders help you organize and control access, Buckets (sometimes called “traffic repartition”) are all about managing how user traffic is allocated across experiments. Think of buckets as traffic lanes on a highway – each experiment gets its own lane, so there’s no risk of collisions.

How Buckets Work

  • Traffic Allocation: By default, you can create up to 10 buckets, each representing 10% of your total user traffic. Assign experiments to specific buckets to ensure they don’t overlap.
  • Mutual Exclusivity: Experiments in different buckets never see the same users, so results are clean and reliable.
  • Planned Flexibility: While the default is 10 buckets, future updates will allow you to customize the number of buckets and the percentage of traffic allocated to each.

Why Buckets Matter

  • No Overlap: Run multiple experiments at the same time – on the same page or feature – without worrying about interference.
  • Reliable Results: By keeping experiments mutually exclusive, you avoid skewed data and can trust your insights.
  • Enterprise-Ready: Especially valuable for organizations with multiple teams running simultaneous experiments.

Why This Matters for Your Team

Folders & Buckets aren’t just “nice-to-have” features – they’re essential for any organization looking to scale experimentation without losing control. They help you:

  • Stay organized as your program grows.
  • Keep sensitive experiments secure and compliant.
  • Empower teams to work independently without stepping on each other’s toes.
  • Deliver reliable, actionable insights by preventing experiment overlap.

As digital experimentation becomes a core business function, tools like Folders & Buckets are what separate the leaders from the laggards.

Ready to Get Organized?

If you’re struggling with a cluttered experimentation environment or worried about experiment overlap, it’s time to explore what Folders & Buckets can do for you. Customer Success Manager for more information, and see how easy it is to bring order – and results – to your experimentation program.

Experiment boldly. Organize smartly. Grow faster.

Want to learn more? Check out our documentation (folders/buckets) or contact us for a personalized demo.

FAQs about experimentation organization

How can I stay organized when running lots of experiments?

AB Tasty is built for big teams running many experiments. We offer our users a clear folder structure to group experiments by team, project, product line, or page type, and apply granular permissions so people only see the campaigns relevant to them. This reduces clutter, limits mistakes, and keeps your experimentation environment manageable as you scale.

What are folders in an A/B testing or experimentation platform?

Folders act like a filing cabinet for your tests: you can group experiments in ways that match your organization (e.g., by team, sprint, product, or page type) and assign viewer, editor, or admin roles per folder to control who can see and edit each campaign.

What are buckets in an experimentation platform?

Buckets (or “traffic repartition”) are a way to divide user traffic into separate lanes. Each bucket gets a portion of traffic (e.g., 10%) and experiments assigned to different buckets don’t share users, which keeps tests mutually exclusive.

Which A/B testing solutions help teams stay organized at scale?

The most effective solutions offer both structural tools (like folders with role‑based access) and traffic management features (like buckets for mutual exclusivity), like AB Tasty. Together, these help large organizations keep experiments secure, organized, and analytically sound as their programs grow.

Article

11min read

Frequentist vs Bayesian Methods in A/B Testing

Table of content

When you’re running A/B tests, you’re making a choice—whether you know it or not.

Two statistical methods power how we interpret test results: Frequentist vs Bayesian A/B testing. The debates are fierce. The stakes are real. And at AB Tasty, we’ve picked our side.

If you’re shopping for an A/B testing platform, new to experimentation, or just trying to make sense of your results, understanding these methods matters. It’s the difference between guessing and knowing. Between implementing winners and chasing false positives.

Let’s break it down.

AB testing Bayesian vs frequentist methods

What is Inferential Statistics?

Both Frequentist and Bayesian methods live under the umbrella of inferential statistics.

Unlike descriptive statistics—which simply describes what already happened—inferential statistics help you forecast what’s coming. They let you extrapolate results from a sample to a larger population.

Here’s the question we’re answering: Would version A or version B perform better when rolled out to your entire audience?

A Quick Example

Let’s say you’re studying Olympic swimmers. With descriptive statistics, you could calculate:

  • Average height of the team
  • Height variance across athletes
  • Distribution above or below average

That’s useful, but limited.

Inferential statistics let you go further. Want to know the average height of all men on the planet? You can’t measure everyone. But you can infer that average from smaller, representative samples.

That’s where Frequentist vs Bayesian methods come in. Both help you make predictions from incomplete data—but they do it differently, especially when applied to A/B testing.

What is the Frequentist Statistics Method in A/B Testing?

The Frequentist approach is the classic. You’ve probably seen it in college stats classes or in most A/B testing tools.

This is one of the main Frequentist vs Bayesian A/B testing comparisons: Frequentist statistics focus on long-run frequencies and fixed hypotheses.

Here’s how it works:

The Hypothesis

You start by assuming there is no difference between version A and version B. This is called the null hypothesis.

At the end of your test, you get a P-Value (probability value). The P-Value tells you the probability of seeing your results—or more extreme results—if there really is no difference between your variations. In other words, how likely is it that your results happened by chance?

The smaller the P-Value, the more confident you can be that there’s a real difference between your A/B testing variations.

What is the Bayesian Statistics in A/B Testing?

The Bayesian approach takes a different route—and we think it’s a smarter one for many A/B testing scenarios.

Baye's Theorem formula

The Bayesian approach allows for the inclusion of prior information (‘a prior’) intNamed after British mathematician Thomas Bayes, this method lets you incorporate prior information into your analysis. It’s built around three overlapping concepts:

The Three Pillars of Bayesian Analysis

  • Prior: Information from previous experiments. At the start, we use a “non-informative” prior—essentially a blank slate.
  • Evidence: The data from your current experiment.
  • Posterior: Updated information combining the prior and evidence. This is your result.

Here’s the game-changer: Bayesian A/B testing is designed for ongoing experiments.  Every time you check your data, the previous results become the “prior,” and new incoming data becomes the “evidence.”

That means data peeking is built into the design. Each time you look, the analysis is valid.

Even better? Bayesian statistics let you estimate the actual gain of a winning variation—not just that it won—making Frequentist vs Bayesian methods in A/B testing very different from a decision-making perspective.

Bayesian ProsBayesian Cons
Peek freely: Check your data during a test without compromising accuracy. Stop losing variations early or switch to winners faster.More computational power: Requires a sampling loop, which demands more CPU load at scale (though this doesn’t affect users).
See the gain: Know the actual improvement range, not just which version won.

Fewer false positives: The method naturally rules out many misleading results in A/B testing.

Frequentist vs Bayesian A/B Testing: The Comparison

Let’s be clear: both methods are statistically valid. But when you compare Frequentist vs Bayesian A/B testing, the practical implications are very different.

At AB Tasty, we have a clear preference for the Bayesian a/b testing approach. 

Here’s why.

Gain Size Matters

With Bayesian A/B testing, you don’t just know which version won—you know by how much.

This is critical in business. When you run an A/B test, you’re deciding whether to switch from version A to version B.

That decision involves:

  • Implementation costs (time, resources, budget)
  • Associated costs (vendor licenses, maintenance)

Example: You’re testing a chatbot on your pricing page. Version B (with chatbot) outperforms version A. But implementing version B requires two weeks of developer time plus a monthly chatbot license.

You need to know if the math adds up. Bayesian statistics give you that answer by quantifying the gain from your A/B testing experiment.

Real Example from AB Tasty Reporting

Let’s look at a test measuring three variations against an original, with “CTA clicks” as the KPI.

AB testing dashboard showing an example of transaction rates and growth metrics across 4 variations with performance trend graph.

Variation 3 wins with a 34.1% conversion rate (vs. 25% for the original).

But here’s where it gets interesting:

  • Median gain: +36.4%
  • Lowest possible gain: +2.25%
  • Highest possible gain: +48.40%

In 95% of cases, your gain will fall between +2.25% and +48.40%.

This granularity helps you decide whether to roll out the winner:

  • Both ends positive? Great sign.
  • Narrow interval? High confidence. Go for it.
  • Wide interval but low implementation cost? Probably safe to proceed.
  • Wide interval with high implementation cost? Wait for more data.

This is a concrete illustration of how Frequentist vs Bayesian methods in A/B testing lead to different levels of decision-making insight.

When to Trust Your Results?

At AB Tasty, we recommend waiting until you’ve hit these benchmarks:

  • At least 5,000 unique visitors per variation
  • Test runs for at least 14 days (two business cycles)
  • 300 conversions on your main goal

These thresholds apply regardless of whether you use a Frequentist or Bayesian method, but Bayesian A/B testing gives you more interpretable outputs once you reach them.

Data Peeking: A Bayesian Advantage

Here’s a scenario: You’re running an A/B test for a major e-commerce promotion. Version B is tanking—losing you serious money.

With Bayesian A/B testing, you can stop it immediately. No need to wait until the end.

Conversely, if version B is crushing it, you can switch all traffic to the winner earlier than with Frequentist methods.

This is the logic behind our Dynamic Traffic Allocation feature—and it wouldn’t be possible without Bayesian statistics.

How Does Dynamic Traffic Allocation Work?

Dynamic Traffic Allocation balances exploration (gathering data) with exploitation (maximizing conversions).

AB Tasty traffic allocation interface with slider controls and pie chart showing test split between original and variations.

In practice, you simply:

  • Check the Dynamic Traffic Allocation box.
  • Pick your primary KPI.
  • Let the algorithm decide when to send more traffic to the winner.

This approach shines when:

  • Testing micro-conversions over short periods
  • Running time-limited campaigns (holiday sales, flash promotions)
  • Working with low-traffic pages
  • Testing 6+ variations simultaneously

Again, this is where Frequentist vs Bayesian methods in A/B testing diverge: Frequentist statistics are not naturally designed for safe continuous monitoring and dynamic allocation in the same way.

Bayesian False Positives Explained

A false positive occurs when test results suggest version B improves performance—but in reality, it doesn’t. Often, version B performs the same as version A, not worse.

False positives happen with both Frequentist and Bayesian methods in A/B testing. But here’s the difference:

How Does Bayesian Testing Limit False Positives?

Because Bayesian A/B testing provides a gain interval, you’re less likely to implement a false positive in the first place.

Example: Your test shows version B wins with 95% confidence, but the median improvement is only 1%. Even if this is a false positive, you probably won’t implement it—the resources needed don’t justify such a small gain.

With Frequentist methods, you don’t see the gain interval. You might implement that false positive, wasting time and energy on changes that bring zero return.

Gain probability using Bayesian statistics

The standard rule of thumb is 95% confidence—you’re 95% sure version B performs as indicated, with a 5% risk it doesn’t.

For most campaigns, 95% confidence works just fine. But when the stakes are high—think major product launches or business-critical tests—you can dial up your confidence threshold to 97%, 98%, or even 99%.

Just know this: whether you’re using Frequentist or Bayesian methods, higher confidence means you’ll need more time and traffic to reach statistical significance. It’s a trade-off worth making when precision matters most.

While this seems like a safe bet – and it is the right choice for high-stakes campaigns – it’s not something to apply across the board.

This is because:

  • In order to attain this higher threshold, you’ll have to wait longer for results, therefore leaving you less time to reap the rewards of a positive outcome.
  • You will implicitly only get a winner with a bigger gain (which is rarer), and you will let go of smaller improvements that still could be impactful.
  • If you have a smaller amount of traffic on your web page, you may want to consider a different approach.

Conclusion

So which is better—Frequentist or Bayesian?

Both are sound statistical methods. But when you look at Frequentist vs Bayesian methods in A/B testing, we’ve chosen the Bayesian approach because it helps teams make better business decisions.

Here’s what you get:

  • Flexibility: Peek at data without compromising accuracy.
  • Actionable insights: Know the gain size, not just the winner.
  • Maximized returns: Dynamic Traffic Allocation optimizes automatically.
  • Fewer false positives: Built-in safeguards against misleading results.

When you’re shopping for an A/B testing platform, find one that gives you results you can trust—and act on.

Want to see Bayesian A/B testing in action? AB Tasty makes it easy to set up tests, gather insights via an ROI dashboard, and determine which changes will increase your revenue. 

Ready to go further? Let’s build better experiences together →

FAQs

What’s the main difference between Bayesian and Frequentist A/B testing?

When you compare Frequentist vs Bayesian methods in A/B testing, Frequentist methods test whether there’s a difference between variations using a P-Value at the end of the experiment. Bayesian methods estimate the size of the gain and let you update results continuously as new data comes in.

Can I peek at my A/B test results early?

With Bayesian A/B testing, yes. The method is designed for ongoing analysis. With Frequentist methods, peeking early creates misleading results because it effectively turns one experiment into multiple experiments.

What is a false positive in A/B testing?

A false positive occurs when test results suggest version B improves performance, but in reality it doesn’t. Bayesian methods help limit false positives by showing the gain interval, making it less likely you’ll implement a variation with minimal or no real improvement.

What confidence level should I use for my A/B tests?

95% confidence is standard for most marketing campaigns. For high-stakes A/B testing, you can increase to 97%, 98%, or 99%—but this requires more time and traffic to reach statistical significance, regardless of whether you use Frequentist or Bayesian methods.

How long should I run my A/B test?

At AB Tasty, we recommend running tests for at least 14 days (two business cycles) and collecting at least 5,000 unique visitors per variation and 300 conversions on your main goal. These benchmarks help both Frequentist and Bayesian approaches produce reliable insights.

What is Dynamic Traffic Allocation?

Dynamic Traffic Allocation is an automated feature that balances data exploration with conversion maximization in A/B testing. Once the algorithm identifies a winning variation with confidence, it automatically sends more traffic to that version—helping you maximize returns while still gathering reliable data using Bayesian methods.

Article

6min read

Agentic AI for Experimentation: Hype vs. Reality

The use of agentic AI for A/B testing is literally a game-changer for marketing teams, making it much easier to scale testing and experimentation programs. But with some companies making bold claims about what their AI can do, it can be hard to know just what to believe. So let’s look at how the competition’s AI really stacks up against ours.

Jump to comparison table

Move at the speed of evidence with Evi

First, let’s look at our agentic AI. Launched in November 2025, Evi is AB Tasty’s AI-powered marketing agent designed for evidence-based decision making. It transforms complex data into clear, actionable strategies for repeatable, measurable results and ensures every step you take is grounded in evidence.

But Evi isn’t just one tool. Evi is a suite of intelligent AI agents integrated throughout the entire AB Tasty platform, all optimized for specific tasks.

AgentFunctionDescription
Evi IdeasIdeationScans pages and generates data-backed ideas for new tests based on visual and contextual input. It uses AB Tasty’s proprietary data and UX principles.
Evi HypothesizeHypothesis creationUses a checklist of essential elements to help you create well-structured hypotheses with clear objectives. Assigns quality scores, highlights gaps and suggests edits.
Evi ContentVisual editorTurns natural language prompts into precise on-page edits (HTML/CSS/JS) with no coding required.
Evi AnalysisPost-test analysisAnalyzes campaign data, delivers clear, actionable insights. Highlights winning variations and breaks down why they drive transactions.
Evi FeedbackQualitative analysisAnalyses Net Promoter Score (NPS) and Customer Satisfaction (CSAT) feedback. Quickly identifies key themes and provides actionable insights from customer comments.
Evi ExploreRevenue insightsPowered by our patented metric, RevenueIQ, provides real revenue projections per visitor / per month with confidence intervals. Let’s you see what each test is worth before you launch.
Evi FormulaCatalog attributes (R&M)Self-serve tool for creating catalog attributes using natural language prompts.

So how does Evi compare to the AI used by our main competitors?

Evi vs. Optimizely Opal

Opal is the name of Optimizely’s suite of AI tools integrated throughout their platform. It’s not a standalone product, but rather a collection of different AI agents threaded across their entire product suite, which, along with Experimentation, also includes CMS, CDP, and Commerce.

Indeed, most of Opal’s AI agents are actually focused on CMS, CDP, and Commerce rather than Experimentation. One potential drawback for customers is that many of these AI features are tied to using the entire Optimizely tech stack. Rather than talk about Opal’s features for other areas, let’s look at what AI features Opal does have specifically for Experimentation:

FeatureDescription
Test ideationGenerates ideas for new experiments based on URLs and brand tone.
Variation editorAI-assisted creation of test variations based on Google Gemini.
Campaign creationCreates containers for both web and feature experimentation.
Variable suggestionsSuggests flag variables and variations in feature experimentation.
Chat-based data explorationAllows conversational exploration of test data.
Results summarizationSummarizes test results and provides directional guidance.
Experiment advisor agentsThese include a personalization advisor, experiment planner, and results summarizer.
Experiment scorecardScores experiments from the analytics interface.

Opal’s AI agents that are used specifically for testing and experimentation are very comparable to those of Evi. Both have dedicated AI agents for ideation, editing, and analysis. However, Evi also includes our proprietary RevenueIQ analysis and can leverage AB Tasty’s other AI features, EmotionsAI and Wandz for targeting and segmentation.

Some of Opal’s features also appear to be standard statistics features that have been rebranded as AI (e.g. multi-armed bandits and sequential testing).

Key differences between AB Tasty’s Evi and Optimizely Opal

  • Price: Evi’s AI features are included in all contracts at no additional cost. Opal is a paid add-on that costs around US$30,000 extra.
  • Speed: Evi’s AI editor is based on OpenAI and proven to be faster than that of Opal, based on Google Gemini.
  • AI Targeting: Evi can leverage our other AI features, EmotionsAI and predictive targeting (Wandz) for AI-based segmentation. Opal has nothing comparable.
  • Revenue Analysis: Evi Explore is based on our patented RevenueIQ metric for ROI projections. Again, Opal has no equivalent.
  • Experimentation focus: Evi is 100% focused on testing and experimentation. Most of Opals AI agents are designed for CMS/CDP/Commerce.

Evi vs. Kameleoon PBX

Kameleoon PBX (Prompt-Based Experimentation) is an AI-powered tool that allows users to generate A/B tests directly from natural language prompts. It is positioned as an all-in-one AI agent for test generation, fully integrated with Contentsquare.

Here is a list of PBX’s key AI features:

FeatureDescription
Prompt-based test generationUsers write prompts describing what they want to test, PBX then generates the necessary code/variation.
Contentsquare integrationTurns behavioral insights from Contentsquare into A/B tests.
Automatic site analysisUpon URL integration, 3 to 4 AI agents analyze the site’s structure (HTML/CSS/JS) giving prompts strong contextual awareness.
Figma integration(Currently a Beta feature) Allows the uploading of mockups from Figma, reducing errors in banner creation and saving time in implementation.
Code generationGenerates HTML, CSS, and JS code for experiments.

Unlike Evi’s suite of specialized AI agents, PBX is a single generalist AI agent. This provides you with a limited amount of control and can make it hard to iterate. Kameleoon claims that by using PBX specifically, its customers can build tests faster, more accurately, and at less cost per test. But the reality is that these improvements aren’t specific to PBX, all vendors with agentic AI see similar positive impacts for their customers.

Key differences between AB Tasty’s Evi and Kameleoon PBX

  • Price: Evi’s AI features are included in all contracts at no additional cost. Like Opal, PBX is a paid add-on.
  • Usage: All AB Tasty customers can make unlimited use of Evi, while use of PBX is based around credit quotas.
  • Architecture: Evi is a suite of different specialized AI agents. PBX is a single generalist agent.
  • Speed: Evi has a prompt response time of around 30 seconds, compared to up to 3 to 4 minutes for PBX.
  • Advanced segmentation: Evi can leverage AB Tasty’s other AI features like EmotionsAI for advanced segmentation. However, like Opal, PBX has nothing comparable.
  • Structured output: Evi supports structured output (JSON, rollback, versioning), while PBX makes no mention of whether this is the case.
  • Production quality: Evi is fully production-ready, while some customers have reported that PBX has QA issues and webperf impact.

Full comparison: Evi vs. Opal vs. PBX

Architecture, Pricing, Performance

CriteriaAB Tasty EviOptimizely OpalKameleoon PBX
AI architectureMulti-agent systemSuite of AI features across platformSingle generalist AI agent
Agent specializationTask-optimized agents (Ideas, Content, Analysis)General-purpose AI toolsOne-size-fits-all, prompt based
Structured outputJSON structure, rollback, versioningLimitedNo mention
PhilosophyEvidence-based, grounded in proprietary dataPlatform-wide AI integrationPrompt-to-test generation
Included in priceYes, all contractsPaid add-on (~US$30K)Paid add-on (+25% list price)
Usage modelUnlimitedCredit-basedCredit-based/quotas
AvailabilityAll usersRequires purchaseRequires purchase
Prompt response time~30 secondsReported as slow~3 to 4 minutes
Production readinessYesYesDemo-ready
Code qualityOptimized HTML/CSS/JSVariableQA issues reported

Feature comparison

FeatureAB Tasty EviOptimizely OpalKameleoon PBX
Test ideationEvi IdeasTest ideationLimited
Visual editorEvi ContentVariation editor (Google Gemini)Prompt-based generation
Hypothesis creationEvi HypothesizeNatural language interfaceNot mentioned
Results analysisEvi AnalysisResults summarizationNot mentioned
Revenue projectionsEvi Explore (RevenueIQ)No equivalentNo equivalent
Feedback analysisEvi Feedback (NPS/CSAT)Not mentionedNot mentioned
AI-based targetingEmotionsAI/WandzNo equivalentNo equivalent
Contentsquare integrationNot nativeNot nativeNative
Figma integrationNot mentionedNot mentionedBeta feature
SPA/Dynamic JS supportYesYesYes

Article

5min read

Search Reimagined: AB Tasty Site Search

Your search bar is the fastest path to conversion. Yet for too long, site search solutions have been frustrating: complex, rigid, and disconnected from what e-commerce and product teams actually need to do their jobs.

Even customers using the most expensive search solutions told us they were dissatisfied. The tools were too complex, built for developers instead of business teams, and required painful manual work to sync marketing campaigns across their site.

We heard you. Loud and clear.

So we rebuilt Site Search from the ground up. The revamped AB Tasty Site Search is now available—designed to drive revenue, empower your teams, and integrate seamlessly into your workflow.

Semantic Search That Understands Intent

The biggest transformation? We’ve eliminated the endless synonym nightmare.

Traditional search engines rely on keyword matches. If a customer searches for “running shoes” but your product catalog says “athletic footwear,” you lose the sale. Fixing this meant maintaining massive synonym lists that were tedious to build and impossible to keep current.

AB Tasty Search uses a hybrid engine that blends keyword matching with semantic AI. The system understands the meaning behind queries, not just the exact words typed.

The impact:

  • Zero “No Results” dead ends: Semantic AI delivers relevant results even when users misspell, use natural language, or describe products contextually
  • No manual synonym maintenance: The system handles variations automatically
  • Real revenue gains: Early clients report a +11.6% increase in transactions and a +6.9% lift in add-to-cart rates
  • Understanding of user needs: We leverage AI to understand user queries without requiring a complex setup.

Built for Business Teams, Not Just Developers

We know the pain of being bottlenecked by engineering resources. Need to spotlight a product for a campaign? Waiting days (or weeks) for dev time kills momentum.

The new AB Tasty Search puts control where it belongs: in the hands of merchandisers and marketing teams.

We provide intuitive low-code/no-code tools that enable rapid iteration without developer dependency. This agility is delivered through two key mechanisms:

  1. Widget-based deployment: You can deploy the search functionality quickly using a simple widget that adds an overlay when a user clicks the search bar. This widget-based deployment allows you to challenge your existing solution in days, not months, enabling faster time to value.
  2. Intuitive merchandising controls: Merchandisers can optimize ranking, spotlight key products, and tailor results to user behavior. You can easily boost or bury products based on any catalog attribute, set query-specific redirections, and make real-time product ranking adjustments.

The result? True autonomy and agility for non-technical teams.

The Strategic Advantage: A Unified Platform for Discovery

Here’s where it gets really powerful: the revamped Search isn’t a standalone tool. It’s part of your complete optimization platform.

Search, Recommendations, and Merchandising now share the same rules, the same interface, and the same data. That marketing campaign you’re launching? Apply your merchandising logic once, and it works consistently across product discovery, search results, and recommendation widgets.

Why buy Search alongside Recommendations & Merchandising?

  • Consistency across channels: Same rules across Search, Merchandising, and Recommendations – no more manual syncing.
  • Personalization: Leverage your unified customer data to tailor search results by user type and behavior.
  • Simplified implementation, stack, and cost: One vendor, one dashboard, one connection to your catalog, and one source of truth for all your optimization activities.

This unified approach allows you to deploy product discovery as a global strategy, challenge it with unified KPIs and a dashboard, and optimize holistically.

What’s Next? Conversational Discovery


The search bar is evolving, and so is AB Tasty. We are already looking beyond traditional search to deliver a truly multi-modal experience that blends conversational, discovery, and keyword search into one seamless journey.

And with our approach to personalization, every search result can be personalized in real time based on user behavior, preferences, and context—delivering the right products to the right people at exactly the right moment. Read more about AdaptiveCX here.

Our upcoming Shopping Assistant will bring an AI-powered chat interface to your site through the same easy widget deployment. Natural language conversations will guide shoppers to the right products, increasing conversion while reducing returns.

The revamped AB Tasty Search module is generally available, ready to replace basic native search solutions (like Shopify or Salesforce) or replace your complex, developer-dependent tools.

If you are looking to eliminate “no results” experiences, empower your merchandising teams, and unify your product discovery strategy under one intelligent platform, the time to explore the new AB Tasty Search is now.

Talk to your AB Tasty representative to start testing the new Search module today. The best way to know if it works for you? Test it.