HOW IT WORKS
Deep Work. Prove the ROI. Then Run.
No internal project. We handle everything behind the scenes (data analysis, custom model training, and API integration). Everything built during your setup phase is refreshed at every run. Your custom metrics and models stay current as your product and customer base evolve.
Proven ROI. We prove the return on your own numbers before we run anything.
No new tool. Everything is delivered directly via API or webhook into the stack your team already uses (CRM, Slack, email and more).
1 | Connect to your data [ Setup phase : Foundations ]
It starts with your historical data
We combine four sources, how your customers use your product, what they pay, what they ask support for, and your sales history, into one clean dataset built on real behavioral history. Usage is where we go deepest, because that is where the earliest, truest signals live.
What you get : Data Audit & Tracking Recommendations Blueprint
For whom : RevOps
2 | Calculate metrics [ Setup phase : Representation ]
Then we turn data into critical metrics
We turn raw product events into a standardized table of 30 custom, advanced behavioral metrics, one precise, business-readable picture per account, delivered at the end of the setup phase and updated at every run.
What you get : Metrics Table
For whom : Growth, RevOps, CS and Leadership
3 | Estimate risk & opportunity [ Setup phase : Prediction ]
After that, we build your predictive revenue engine
Using your behavioral history, our machine learning model identifies the exact sequences that preceded churn, retention and upsell in your past data, the winning patterns and the losing ones. For each account, 15 days before renewal, the model scores the risk and triggers an investigation if the threshold is met.
What you get :
for every account : Churn Risk & Expansion Scores
for each alert : AI Memos
For whom : CS / AM, RevOps and Marketing
4 | Measure ROI [ Validation step before run ]
Finally, we compare our predictions against reality
Before you commit to the live run, we backtest our engine on a frozen snapshot of your historical data. By comparing our blind predictions against what actually happened with your accounts, we deliver un-biasable proof of the exact financial value we capture.
What you get : Predictability Score
For whom : RevOps and Finance
5 | Run continuously
It's ready: it runs every week
Once live, every run delivers your custom metrics table, predictive scores, and AI action memos in the exact format that fits your daily workflows via API. You also receive a impact digest, a short report on what changed, what your team did, and the impact on revenue, so you can steer.
What you get : Continuous API Outputs (Churn & Upsell Scores & AI Memos) + Impact Digest
For whom : Growth, RevOps, CS / AM and Leadership
Per-account data pushed via API / Webhook
- churn_risk
- probability score
- healthscore
- low / mid / high (certified)
- trend
- improving / stable / declining
- top_driver
- primary churn factor
- recovery_memo
- Memo AI action recommendation
- days_to_renewal
- countdown
ROADMAP
What's next
2026
Prediction Engine
Confidence intervals
Prediction Engine
Multi-horizon forecast
Growth & Activation
Upsell detection
CSM Workflow & API
Investigation on Demand
CSM Workflow & API
Slack & Teams delivery
CSM Workflow & API
Public API v1
Trust & Reliability
Drift alerting
Packaging
Model cards
2027
Prediction Engine
Cohort forecasting
Prediction Engine
Time-to-event modeling
Growth & Activation
Activation Booster
Growth & Activation
Adoption Clinic
CSM Workflow & API
In-CRM diagnostic panel
CSM Workflow & API
API v2 (write)
Trust & Reliability
SOC 2 Type II
Packaging
EU AI Act pack
2028
Prediction Engine
Real-time scoring
Growth & Activation
Plan-fit recommender
CSM Workflow & API
Conversational copilot
CSM Workflow & API
Self-serve console
Trust & Reliability
BYOC deployment
2029 · 2030
Prediction Engine
Industry-tuned models
Growth & Activation
Revenue simulator
CSM Workflow & API
White-label program
Trust & Reliability
Partner marketplace
Roadmap reviewed every quarter with subscribers
SECURITY
Data sources we process
We choose to analyse sources (Usage, billing, support, and CRM) that are traceable, objective, and non-interpretable. We are tool-agnostic, from Amplitude to Snowflake, Hubspot to Salesforce, we connect to the stack you already use.
Connected to your stack : Usage · Billing · CRM · Support
Missing yours? Just ask us.