The problem in this industry
In SaaS, the product itself is the richest source of intent data — who’s using which feature, how many seats are active, whether an account is trending up or going quiet. Yet most lifecycle messaging ignores it, firing the same day-3 email and the same “upgrade to Pro” banner at everyone regardless of whether they’ve hit their activation moment or are three seats over their license. The result is nudges that land at the wrong time: an upgrade prompt to an account that hasn’t seen value yet, or a renewal reminder to a champion who was ready to expand. The three moments that matter — activation (does the account reach first value?), expansion (are they outgrowing their plan?), and retention (is a healthy account about to lapse?) — each need a different play, and each is legible in usage data if you act on it. The failure mode is over-messaging across surfaces: an in-product banner, a lifecycle email, and a CSM reaching out all in the same week, none aware of the others. KaireonAI turns product-usage signals into next-best-actions. It computes account-level health and usage ratios, gates offers on where an account actually is in its lifecycle, coordinates in-product, email, and CSM outreach under one contact budget, and surfaces the single most valuable play per account.What you build in KaireonAI
You turn product-usage signals and lifecycle plays into a small set of platform building blocks. Here is how a typical product-led or sales-assisted setup maps on:| Platform concept | How you use it in SaaS / B2B |
|---|---|
| Offers (plays) | Upgrade to Pro, add seats, switch to annual, feature-trial unlock, onboarding-webinar invite, usage-based upsell, renewal reminder, win-back for a lapsing account |
| Channels | In-product banner, email, in-app message, CSM (human) outreach task |
| Decisioning gates (Eligibility → Fit → Match) | Eligibility: attribute_condition on plan_tier, and a seat gate — pitch “add seats” only when seats_used / seats_licensed >= 0.9. Fit: propensity_threshold in fit mode on account health. Match: boost the play that matches the account’s lifecycle stage |
| Contact policies | customer_total_cap so in-product + email + CSM share one account-level touch budget; frequency_cap per surface; cooldown on a declined upgrade; do_not_contact for accounts that opted out of marketing |
| Scoring approach | A gradient_boosted model thrives on rich usage features (logins, feature adoption, seat trends); logistic_regression gives a clean, explainable expansion-propensity baseline; an online_learner adapts continuously as product usage streams in |
A worked example
Northwind is a 40-seat Team-plan account, 8 months in, withseats_used at 38, weekly active users trending up 20% this month, and a strong account_health score. The primary contact, Jordan, hasn’t been pitched anything in two weeks. Renewal is 90 days out.
Inventory: 8 plays in play
Upgrade to Pro, add seats, switch to annual, feature-trial unlock, onboarding webinar, usage-based upsell, renewal reminder, win-back.
Eligibility & fit gates → 4 remain
Onboarding webinar drops (8 months in, well past onboarding). Win-back drops (the account is healthy and growing, not lapsing). The renewal reminder is held for the 90-day window. Four survive — and the seat gate opens “add seats,” since utilization is at 38/40.
Contact policies suppress → 3 remain
A
customer_total_cap keeps the account’s total touches modest so in-product, email, and CSM don’t pile on. Lower-priority plays wait.Scoring + expansion-aware ranking
The gradient-boosted expansion model scores the survivors on usage momentum and health. Ranking blends PRIE; add seats ranks first — high propensity (utilization is at the ceiling and users are growing), strong impact (seat expansion is durable revenue), and timely emphasis.
Delivered: 'add seats' as a CSM play
Because the account is high-value and healthy, the top play routes to Jordan’s CSM as a warm-outreach task rather than an impersonal banner — with the decision trace showing the usage evidence behind the recommendation.
Measuring success
- Business Dashboard — acceptance and revenue per play (expansion, upgrade, renewal), and the funnel from configured to delivered.
- Model Health Dashboard — expansion-model AUC and feature importance; watch for drift as your product and usage patterns evolve.
- Expansion uplift — an experiment with a holdout of eligible accounts that get no expansion play proves the program drives incremental expansion, not expansion that would have happened organically.
- Contact-budget adherence — the suppression rate confirms your account-level
customer_total_capkeeps in-product, email, and CSM coordinated.
Where the agentic layer helps
Decisioning Autopilot watches expansion and churn models for drift and your lifecycle experiments for winners, then proposes retrains and challenger promotions — reviewable in the inbox or auto-applied through the audited path. As your product ships features and usage patterns shift, this keeps the models current without a standing manual retrain cadence. Decision Sentinel catches the accidental self-inflicted outage: a seat-gate threshold typo or an over-tight cap that silently stops in-product nudges from rendering. It watches the decision stream for empty-decision and suppression spikes and alerts (with optional auto-pause) before a whole cohort stops seeing plays. The governed AI assistant lets a lifecycle PM change a gate or cap in plain language, routed through approvals, so the growth team moves fast without editing production directly.Try it
Onboarding & Activation Tutorial
Build a staged activation sequence that stops the moment an account reaches first value.
Churn Prevention Tutorial
The retention half of the lifecycle — risk scoring and well-timed saves.
Computed Values
Derive seat-utilization and health signals inside the decision with the formula engine.
Open the Playground
Register and build a SaaS lifecycle flow end to end.