AI AgentResults & management

Discover Victor

Q: Which exports can Victor read?

Kameleoon, AB Tasty, Dynamic Yield, Optimizely and VWO exports, as well as GA4 or Piano for segments.

Q: What happens if a test isn't reliable?

It is classified as non-interpretable. Victor checks the SRM, duration and power before any verdict, and does not read a result that doesn't hold up.

Q: Is the business impact quantified?

Only for a winning test, and as a range. A neutral test is a result in its own right, documented as such.

Q: Who decides what happens next with a test?

The Welyft expert decides whether to stop, deploy, iterate or abandon. Victor prepares the report and its evidence.

Q: Do agents have access to our data?

Only to the sources you open, with a scope and access defined by contract. Agents work on your tools and your exports; no personal data is copied into a deliverable.

CRO cycle · 05–07 · Run, analyse, learn & share

Victor monitors every live test, reads all segments, checks statistical reliability and explains the reasoning behind each result. The end-of-test report is ready for presentation.

Checks the launch of each test
Checks SRM, duration and power
Classifies the result according to the Welyft decision tree
Delivers the report and one-pagers

Book a demo See the whole fleet

Victor

Résultats & pilotage

Driven by a Welyft CRO expert

Real-world use cases

What Victor can do for you

Scenario 1 · Classifying a test
Victor
Live
You : Classify this test from the Kameleoon export and prepare the one-pager.
09:41 · Seen
VictorThe pre-registered primary KPI decides, the guardrails are checked. The test is classified according to the Welyft decision tree and the one-pager is ready, with the segments that explain the result.
Scenario 2 · Launch control
Victor
Live
You : Checks the test launch: traffic allocation, tracking by variant, JS errors.
11:08 · Seen
VictorTraffic allocation is compliant, each variant reports its tracking hits and no JS error is detected. I'll check back on the SRM once the first weekly cycle is complete.
Scenario 3 · Reliability of a result
Victor
Live
You : Is the result reliable? Check the SRM and the power.
14:26 · Seen
VictorI check the sample ratio, the duration in complete cycles and the power reached. If any of them fails, the test is classified as non-interpretable rather than read incorrectly.

Illustrative examples: each mission starts from your data and your tools.

Deployment

Agents trained like experts, updated like products.

Each agent is a versioned plugin on GitHub. Three ways to make the most of it.

During a mission

During the engagement

Our consultants call on the agents at every stage of the CRO cycle. You receive the deliverables, without installing anything.

Via the Welyft platform

Remote, at Welyft

The fleet runs on our platform, connected to your tools with your consent. An expert validates every deliverable.

Internalized within your organization

ConstructionWelyft builds the agent using its method, test learnings and CRO best practices.

PersonalizationEach agent is tailored to your method, brand guidelines and vocabulary.

GitHub MarketplaceYour custom agents install as plugins from a repository dedicated to your organization.

UpdatesEvery Welyft agent update is versioned, reviewed and delivered to you.

Your toolsConnected to your stack via MCP.

GuardrailsA fixed permissions matrix that decides and validates.

Within your organization

We build your agents using our method, customize them, then a GitHub plugin marketplace delivers their updates.

Let's talk about your environment Train your teams: AI for CRO

Watch Victor read a test and deliver his end-of-test report.

Victor, Welyft's results & monitoring agent, oversees your live tests, verifies their statistical reliability and explains the why behind each result.

Book a demo

Its skills

Monitor, analyze, report back

Skills · Monitoring

Live
Controlled launch
Traffic allocation, tracking by variant and JS errors are verified as soon as the test goes live.
/victor:monitoring
Live
SRM alert
A sample imbalance triggers an alert (p < 0.01) before any reading of the result.
/victor:monitoring
In progress
Daily review
Automatic review of each ongoing test is under development; the Value Hub's daily brief already covers the program.
/victor:monitoring

Skills · Analysis

Live
Every segment read
Device, source, new and returning: Victor reads every segment from your exports and your analytics.
/victor:analysis
Live
Statistical reliability
SRM, duration, power and peeking are checked before any verdict.
/victor:analysis
Live
The why explained
Cross-analyses and position against Welyft's history explain what moved the KPI.
/victor:analysis

Skills · Reporting

Live
Welyft decision tree
Winner, loser, neutral, stopped or non-interpretable: every test receives a clear classification.
/victor:reporting
Live
One-pagers and deck
Brief, summary and production version, then the Welyft-branded report deck.
/victor:reporting
Live
Learnings database
Every documented test joins the learnings database, which feeds Amir's hypotheses.
/victor:reporting

Human first

Victor prepares, the Welyft expert decides.

Deliverable: Rapport de fin de test
The expert decides: arrêter, déployer, itérer ou abandonner

Its safeguards

The pre-registered primary KPI decides; a degraded guardrail causes the test to lose.
Zero invented figures; a neutral test is a result.
Business impact given as a range, only for a winning test.
In the event of an incident, a pause is recommended and traffic is returned to control; the test becomes non-interpretable.

Under the hood

What Victor relies on

Trained on

Nos rapports de fin de test

Data used

Kameleoon, AB Tasty, Dynamic Yield, Optimizely or VWO exports
GA4 or Piano
Pre-registered test brief
A/B testing tool (AB Tasty connector)

Specialties

/victor:welyft-rapport-test
Alertes test live · coming soon

It's live : /victor:welyft-rapport-test

In progress : The automatic daily review of each ongoing test has yet to be built.

Our offers

Offers where Victor is involved

FAQ

Frequently asked questions about Victor

Une question sur votre contexte ? Nos experts vous répondent.

Posez votre question

Which exports can Victor read?

Kameleoon, AB Tasty, Dynamic Yield, Optimizely and VWO exports, as well as GA4 or Piano for segments.

What happens if a test isn't reliable?

It is classified as non-interpretable. Victor checks the SRM, duration and power before any verdict, and does not read a result that doesn't hold up.

Is the business impact quantified?

Only for a winning test, and as a range. A neutral test is a result in its own right, documented as such.

Who decides what happens next with a test?

The Welyft expert decides whether to stop, deploy, iterate or abandon. Victor prepares the report and its evidence.

Do agents have access to our data?

Only to the sources you open, with a scope and access defined by contract. Agents work on your tools and your exports; no personal data is copied into a deliverable.

In a few seconds

Victor, from data to deliverable

An end-of-test report, on brand

Results & management

Verdict, one-pager and learning

Rapport de fin de test

Verdict vérifié, prêt pour la restitution.

Test verdict

Winner

SRM
Test duration
Power

verdicts, from winner to non-interpretable

one-pagers: brief, high-level, prod

Learning sheet

Each test enriches the learnings database

CSV from the value dashboard

Dataviz to the 2026 style guide

Presentation deck

Illustrative demonstration: the examples are not client results.