We connect your first three channels and drop your knowledge base into the tenant. Classify and Retrieve come online immediately — your inbound is tagged, and the system starts citing the pages your senior reps already rely on. Suggest runs in shadow mode, drafting replies nobody ships yet, so you can watch its voice. By end of day, your team has one inbox instead of five.
Your corpus has grown. Suggest moves from shadow to default — every thread lands with a grounded first draft. Act is wired: quotes, refunds, and scheduling flow through your ERP and billing systems without tab-switching. The model has watched a month of your senior reps’ edits — it sharpens on those signals. Score shows early patterns: which intents get clean first drafts, which still need the room.
ClassifyRetrieve· FULL CORPUSSuggestActLearnScore
[EXHIBIT B] · GROUNDED DRAFT LOOP
[DAY 90]CRUISING
The room hears one voice.
The team has shipped thousands of drafts. Autonomous send rules, scoped by intent and account, handle the clean cases — order confirmations, standard quotes, thank-you replies. Your senior reps are where they should be: the ten threads a week that actually need them. Executive dashboards land in the drafting surface itself, not a separate BI tool.
ClassifyRetrieveSuggestAct· AUTOLearnScore· EXEC
[EXHIBIT C] · FULL MODULE RING
[04]GATESControls you hold, not us
[GATE α]
Autonomous send
OFF by default. Enable per intent, per account.
[GATE β]
New channels
Each new channel requires explicit admin approval.
[GATE γ]
Knowledge updates
Every corpus change is versioned and auditable.
[GATE δ]
Model swap
Choose which model serves drafts for each intent.
[ACCESS]REQUEST INVITATION
We're onboarding in cohorts. Tell us your shape.
Who you sell to, how many channels you run, and one thing your best rep knows that the rest of the team doesn't. We'll reply with a walkthrough grounded in your actual docs — not a generic demo.
CLASSIFIED · PRE-LAUNCH · BY INVITATION
[RESPONSE SLA] 1 business day · [REGION] EU/US · [CAPACITY] Q1 2026