Login
English

Select your language

English
Français
Deutsch
Platform

PROMPT-BASED EXPERIMENTATION

Optimize any website by chatting with PBX, Kameleoon’s AI. Learn more

SOLUTIONS
Experimentation
Feature Management
KEY Features & add-ons
Mobile App Testing
Recommendations & Search
Personalization
specialities
AI in Experimentation
Single Page Application
Data Security & Privacy
Data Accuracy
Integrations & APIPartners ProgramSupportProduct Roadmap
Solutions
for all teams
Marketing
Product
Engineering
Data Scientists
For INDUSTRIES
Healthcare
Travel & Tourism
Financial Services
Media & Entertainment
E-commerce
B2B
Automotive

Samsung gains autonomy and speed in its most advanced tests with PBX

read the success story
PlansResourcesCustomers
Book a demo
Book a demo
All customer stories
AS Watson

How AS Watson connects customer insights to AI-powered experimentation

Industry

E-commerce & Retail

key features

PBX

250+

experiments a year targeted by a team of 2 CRO specialists

30min

from insight to launch-ready experiment

+1.3%

checkout starts

In this story


Focus on continuous innovation
PBX accelerating testing
PBX as an ideation co-pilot
Conclusion
‍
request a demo
request a demo

At a glance:

AS Watson is the world's largest international health and beauty retailer, with more than 17,000 stores across 31 markets. Its portfolio includes Superdrug in the UK and Kruidvat in the Benelux. As digital experiences become increasingly important to its omnichannel strategy, experimentation has become a key driver of customer journey optimisation.

With Kameleoon PBX, AS Watson is accelerating experiment creation, reducing manual work and helping its two-person CRO team scale experimentation without increasing reliance on development resources.Key takeaways:

  • From customer insight to launch-ready experiment in 30 minutes
  • More time for strategic optimisation, with AI handling repetitive tasks
  • +1.3% checkout starts
  • +1.0% conversion rate

The challenge: Plenty of insights, limited resources

At Kruidvat, one of AS Watson's leading brands in Belgium and the Netherlands, a central team of just two CRO specialists manages more than 100 experiments a year across three e-commerce websites and two mobile apps, attracting millions of visits every week.

The team relies on a technology stack that includes Contentsquare, GA4, Figma, Jira and Kameleoon.

Optimisation opportunities come from multiple teams, including merchandising, marketing, UX, analytics and CRO. While this creates a wealth of ideas, it also puts significant pressure on a small team responsible for turning them into live experiments.

The challenge wasn't a lack of ideas. It was turning those ideas into experiments quickly enough.

Customer insights lived in Contentsquare. Business knowledge was spread across documents and teams. Design, analytics and development happened in separate tools.

Each experiment required manually bringing this information together before the team could even start building a variation.

{{quote-1}}

Rather than growing the team, AS Watson wanted to remove the friction slowing experimentation down.

The solution: Connecting customer insights to experimentation

To address this challenge, AS Watson began connecting its experimentation workflow with AI, in two stages.

The first step was making AI more context-aware. Behavioural data from Contentsquare helps identify friction points in the customer journey. By combining these insights with business rules, analytics, technical knowledge and learnings from previous experiments, the team can develop stronger, better-documented hypotheses.

The second step was automating specific tasks in the experimentation lifecycle. Rather than attempting to automate the entire CRO programme, AS Watson identified repetitive, time-consuming tasks that AI agents could take on.

This is where Kameleoon comes in, with two Prompt-Based Experimentation (PBX) agents:

  • PBX Build : Using a prompt prepared by the team or its internal AI agents, PBX generates the front-end code needed for an experiment, removing the need to manually develop every variation.
  • PBX Configure : The team can set up and launch experiments faster, including audiences, goals and performance metrics, while retaining final approval.

Connected to business context and analytics data through AS Watson's wider workflow, these agents allow the CRO team to spend more time prioritising opportunities, improving hypotheses and interpreting results, rather than handling technical execution.

From customer insight to launch-ready experiment in 30 minutes

One example comes from the Kruidvat.nl shopping cart page.

Using Contentsquare, the team identified a friction point in the checkout journey. For customers with larger baskets, the main checkout button was pushed below the fold, forcing shoppers to scroll down before they could continue.

The hypothesis was simple: keeping the checkout button visible with a sticky CTA could make it easier for customers to move to the next step and improve conversion.

The team used its AI workflow to turn this opportunity into a PBX-ready prompt. PBX Build quickly created the variation, while PBX Configure helped set up and launch the experiment.

After a final quality check by a CRO specialist, the experiment was ready to go live.

It took approximately 30 minutes to move from identifying the opportunity to having a launch-ready experiment.

For a team of just two specialists, this is more than a time saving. It fundamentally changes how experiments can be created and launched every day.

{{quote-2}}

The results: Measurable impact on conversion

The sticky checkout button experiment reached statistical significance within a few weeks, delivering positive results across two key metrics.

  • +1.3% Checkout starts (p < 0.01)
  • +1.0%Conversion rate (p < 0.01)

These results show how a relatively simple optimisation, identified through behavioural data, can deliver measurable business impact when quickly turned into an experiment.

It's this faster journey from insight to action that makes the difference for AS Watson. Contentsquare's behavioural data informs the hypothesis, while PBX accelerates test creation and configuration, reducing manual work between each step.

‍

Scaling experimentation with AI, without growing the team

For AS Watson, the biggest benefit goes beyond generating code faster.

With just two CRO specialists, every hour saved on building or configuring an experiment can be reinvested in experimentation strategy.

With PBX Build and Configure, the team can launch experiments faster, test more technically ambitious ideas with less dependency on development resources, reduce manual tasks and dedicate more time to identifying the right opportunities.

Alongside Kameleoon, AS Watson is experimenting with its own AI agents to handle other tasks in the CRO lifecycle. One agent can prepare prompts from experiment briefs, while another supports results analysis and documentation of learnings.

These complementary initiatives are helping AS Watson gradually extend automation beyond variation creation.

Today, almost every experiment incorporates AI at some point in the workflow.

Key learnings

  • Context makes all the difference. AI delivers better results when it understands behavioural data, business rules and company knowledge. This helps teams develop more relevant hypotheses.
  • Keep humans in control. AI accelerates execution, but experimentation specialists remain responsible for prioritisation, quality and final approval.
  • Automate tasks, not decisions. For a small CRO team, the real value comes from delegating repetitive work to AI and protecting time for decisions that require human expertise.
  • Think beyond individual tasks. Automating just one step can simply move the bottleneck elsewhere. Connecting the different stages of experimentation is what creates the greatest efficiency gains.
  • Start small and improve continuously. Rather than waiting for a perfect AI roadmap, AS Watson has evolved its workflow incrementally, refining its automation with each experiment.

What's next: Towards a more connected experimentation lifecycle

AS Watson's vision goes beyond individual AI features.

By combining Contentsquare's customer intelligence with Kameleoon's PBX, the retailer is gradually building an ecosystem where insights, execution and learnings can continuously inform one another.

Today, AI agents handle specific tasks within the experimentation lifecycle, while CRO specialists remain involved at key checkpoints.

Looking ahead, the ambition is to go further: enabling every experiment to inform the next, with stronger hypotheses, faster execution and a more systematic approach to capturing and reusing learnings.

{{cta-block}}

AI Predictive Targeting

View More

Prompt-Based Experimentation

View More

Deliver outstanding customer experiences with Contentsquare x Kameleoon integration

View More

"We weren't short on data. We were drowning in it."

Peter Harsevoord
CRO lead at Kruidvat.nl

"We're just two CRO specialists managing more than 100 experiments a year. Before PBX, turning an idea into a test involved a lot of manual work and coordination. Today, with PBX, we can build and prepare an experiment in minutes, without compromising the quality of our tests."

Peter Harsevoord
CRO lead at Kruidvat.nl
USE CASE #1
USE CASE #2
USE CASE #3
USE CASE #4
USE CASE #5
Team organization
Book a demo
Book a demo
integrations

Power experiments with user insights

Connect Mixpanel to Kameleoon for real-time insights and cohort-based experimentation.

learn more
learn more

Your entire test workflow, powered by PBX agents.

Bring the power of connected agents to your experimentation programme, without the complexity of building and maintaining your own AI infrastructure.

Book a PBX demo
Book a PBX demo
Experiment your way

Get the key to staying ahead in the world of experimentation.

[Placeholder text - Hubspot will create the error message]
Thanks for submitting the form.

Newsletter

Platform
ExperimentationFeature ManagementPBX Free-TrialMobile App TestingProduct Reco & MerchData AccuracyData Privacy & SecuritySingle Page ApplicationAI PersonalizationIntegrations
guides
A/B testingPrompt-Based ExperimentationFeature FlaggingPersonalizationFeature ExperimentationAI for A/B testingClient-Side vs Server-Side
plans
PricingMTU vs MAU
Industries
HealthcareFinancial ServicesE-commerceAutomotiveTravel & TourismMedia & EntertainmentB2B & SaaS
TEAMS
MarketingProductDevelopers
Resources
Customers StoriesAcademyDev DocsProduct RoadmapCalculatorWho’s Who
compare us
OptimizelyWingify
partners
Our Partner EcosystemBecome a PartnerIntegrations DirectoryPartners Directory
company
About UsCareersContact UsSupport
legal
Terms of use and ServicePrivacy PolicyLegal Notice & CSUPCI DSS
© Kameleoon — 2026 All rights Reserved
Legal Notice & CSUPrivacy policyPCI DSSPlatform Status