Predictions for every customer
The platform computes predictive profile attributes from purchase and activity history — use them in segments, journeys and emails.
What is predicted
Models retrain daily and do not need a data science team.
- RFM segments: champions, loyal, at risk, hibernating
- Churn risk and probability of purchase over a horizon from a week to three months
- Predicted customer revenue and engagement
- Cart abandonment risk

Best time and channel to send
From the history of opens and clicks the platform finds, for every customer, the hour when they usually read messages and the channel they respond to best. A campaign reaches each recipient at their best hour, a journey waits for that hour before the email, and a cascade starts with the preferred channel.
- The hour is in the customer’s time zone; without their own history, the project’s best hour is used
- best_send_hour and preferred_channel traits in segments and journey conditions
- Included in the plan together with predictions

Model quality in plain sight
Prediction accuracy is measured on a holdout sample: the report shows how much better the model is than a simple rule and whether to trust it.

Next best action for every customer
The Decision journey node picks for each participant what to do: give a discount, offer free shipping, send a selection of products or send nothing. It estimates how much revenue, purchases or returning visits each action adds for this particular customer compared with doing nothing, picks the best one and learns from the results of its decisions every night. The discount budget, offer frequency and the “no discount for those who would buy anyway” rule are node settings; some participants always get “nothing”, so the uplift has a baseline.
- Node goal: revenue, a purchase or retention within a window of up to 60 days
- Report: decisions and average outcome per action, model uplift over “nothing”
- The profile card shows why the customer got this action

- Profiles and eventsA single customer profile from website, app and CRM events.
- SegmentsAudiences by events and attributes — in real time.
- JourneysVisual scenarios: triggers, waits, branches and A/B.
- LoyaltyPoints, paying with points, tiers, friend referrals, promo codes and Wallet cards.
- Channels and campaignsEmail, web and mobile push, in-app messages, SMS, messengers and A/B-tested campaigns.
- Live chatWebsite chat, messengers and email in one inbox for your whole support team, with the customer profile alongside.
- Signup formsPopups and website forms: the contact goes straight into the profile with consent.
- Recommendations and widgetsPersonal product picks on your site and in emails.
- AnalyticsReports, funnels, retention, revenue, traffic, channel attribution, ROPO and a control group.
- AI assistantsThe in-app assistant figures things out and builds campaigns; Claude and ChatGPT over MCP.
- White-labelThe platform under your brand and domain.
Try it on your own data
14 days on the Growth plan, no card required: up to 1,000 profiles and 500 emails a day.