WysLeap
WysAI intent scoring

Know which visitors are ready to convert.

WysAI gives every visitor an intent score from 0 to 1, built from five behavioral signals and fitted to your own site's conversions once there are enough of them. Spend your attention on the visitors most likely to buy.

  • No cookies
  • GDPR-ready
  • No cross-site tracking
WysAI · visitor 7f3a91c2

0.78

High intent
00.300.601

5 behavioral signals

  • Sessions4
  • Pages viewed11
  • Engaged time6m 20s
  • Scroll depth75%
  • Clicks9
Model: trained on this sitebeat default weights on held-out visitors

Illustrative example. Bars show each signal's relative contribution for this visitor.

The problem

Traffic reports treat every visitor the same

A first-time skimmer and a buyer on their fourth visit look alike in pageview totals. Intent scoring separates them.

Without intent scoring

  • Every visitor looks the same in aggregate reports
  • Follow-up effort is spread evenly, whatever the intent
  • High-intent visitors are easy to miss in the crowd
  • No single number for sales or support to act on

With WysAI

  • A 0–1 score and a band for every visitor
  • A prioritized list of who to look at first
  • Weights fitted to your site once you have enough conversions
  • A visible breakdown of which signals drive each score

How it works

Five signals, one score, a model that has to prove itself

From the first page view to a score in your dashboard, with nothing hidden about when your own model takes over.

Sessions

Return visits, same browser

Pages viewed

Automatic page views

Engaged time

Pings every 15 seconds

Scroll depth

25% steps

Clicks

Link and button clicks

  1. 01

    Signals are collected automatically

    The WysLeap tracker records page views, engaged time, scroll depth and clicks, and links sessions from the same browser with a random visitor ID. There are no page-specific rules to set up.

  2. 02

    A logistic model turns them into one score

    Each signal has a weight; the weighted sum becomes a score from 0 to 1. You see the number and its band: high (0.60 and above), medium (0.30–0.59) or low (below 0.30).

  3. 03

    It learns from your own conversions

    Every site starts with default weights. Once yours has at least 15 conversions and 30 non-converting visitors in 90 days, WysAI trains a model on your data, then retrains after every 10 new conversions.

  4. 04

    A new model has to earn its place

    Each trained model is tested on the newest 25% of visitors, held back from training, and is only used if it ranks them better (by AUC) than the default weights. Otherwise the default stays.

Scores are computed from each visitor's latest activity when you open the dashboard. Your dashboard shows how well your model ranks held-out visitors; there's no single figure for all sites, because it depends on your site's data.

Score explorer

What a score means

Drag the slider to see how WysAI bands a score and what a typical signal pattern looks like at each level.

0.75

High intent

Fixed band: 0.60 – 1.00

Typical signal pattern

  • SessionsSeveral return visits
  • Pages viewedMany pages
  • Engaged timeLong
  • Scroll depthDeep
  • ClicksFrequent

Suggested next step

Prioritize for follow-up · Review their journey in the dashboard

Illustrative example. Actual weights are learned per site.

0.000.300.601.00

Live demo

Example scores from wysleap.com

The top three visitors on our own site right now, ranked by a simplified demo heuristic with its own cut-offs (not the per-site WysAI model).

In your dashboard, scores come from your site's WysAI model and include a breakdown of the signals behind each one.

Demo · top 3 visitors by intent
WysAI model · wysleap.com trafficFull dashboard →

Use cases

Who benefits most

What teams do with a score today.

Know which shoppers are closest to buying

Prioritized follow-up

Use intent scores in your dashboard to decide which visitors and journeys deserve attention first.

Understand what drives intent

See which of the 5 behavioral signals contribute most to a visitor's score.

Plan offers around high intent

Review the pages high-intent visitors engage with and plan offers and content for them.

Compared

Traditional analytics vs WysAI

Traditional reports count what happened. Intent scoring ranks who is most likely to act.

FeatureWysAITraditional analytics
Per-visitor conversion likelihood
0–1 score
Learns from your site's conversions
Combines multiple behavioral signals
5 signals
Single metrics
Retrains as new conversions arrive
Static reports
Returning-visitor recognition
Random ID, no cookies
Usually cookies
Prioritized visitor list
Aggregate only

Behavior, not identity

Scores use sessions, pages, engaged time, scroll depth and clicks. No names or emails are collected, and IP addresses are anonymized.

Random visitor ID

Returning visitors are recognized by a random ID in localStorage, and device fingerprinting can reconnect a browser whose storage was cleared, on the same network. No cookies, no cross-device tracking unless the site turns on login linking.

Your model, your data

Weights are learned from your site's own conversions. Anonymized benchmarks shared across sites don't train or pre-score anyone else's visitors.

FAQ

Frequently asked questions

What signals does WysAI use to score a visitor?
Five behavioral signals: number of sessions, pages viewed, engaged time, scroll depth and clicks. All are collected automatically by the WysLeap tracker: page views, engaged-time pings every 15 seconds, scroll depth in 25% steps, and link and button clicks.
How is the score calculated?
WysAI uses a per-site logistic model: each signal has a weight, and the weighted sum is turned into a score from 0 to 1. Scores of 0.60 and above are high intent, 0.30–0.59 medium, and below 0.30 low.
When does WysAI train a model for my site?
Every site starts with default weights. Once your site has at least 15 conversions and 30 non-converting visitors in 90 days, WysAI trains a model on your own data, then retrains after every 10 new conversions.
How do you know a trained model is better than the default?
Each trained model is tested on the newest 25% of visitors, which are held back from training. It is only used if it ranks those visitors better (measured by AUC) than the default weights. Your dashboard shows that result for your site once a model is trained. There's no single accuracy figure for all sites, because it depends on each site's own conversions.
Are scores updated in real time?
Scores are computed from each visitor's latest activity when you open the dashboard. They are not pushed to other tools.
Does intent scoring use cookies or personal data?
No cookies and no names or emails. Returning visitors are recognized in the same browser by a random visitor ID in localStorage; clearing site data or using a private window starts a new ID, and there is no cross-device linking unless the site turns on login linking. IP addresses are anonymized.

Start scoring visitor intent.

Starts with default weights and learns once you track conversions. See which visitors are most likely to convert.

90-day free trial · No credit card required