AI prompts that read from HubSpot — and write back, with approval.
Stop pasting deal context into ChatGPT. Stop wondering which AI use closed which deal. CommonPlace pulls HubSpot data into every prompt run, writes updates back through approval workflows, and attributes outcomes — closed-won, support tickets, NPS — to the specific prompts that touched them.
Most AI use is invisible to your CRM. That's the problem.
Today's AI workflow is reps tabbing between HubSpot and ChatGPT, pasting deal context one way and drafted emails the other. CommonPlace closes the loop: HubSpot context flows IN, drafts flow BACK, and the attribution model tells you which prompts move revenue.
- ✗Reps paste deal context into ChatGPT — IT has no visibility
- ✗Drafts go back into HubSpot manually, with copy-paste drift
- ✗No one can tell finance which AI use closed which deal
- ✗Sensitive deal data leaks across consumer AI accounts
- ✓Prompts pull from HubSpot at run time — read-scoped per prompt
- ✓Write-backs queue for approval, or auto-apply on trusted prompts
- ✓Beacon attributes closed-won to the prompts that touched the deal
- ✓All data flows through your org's LLM keys, your org's audit log
One round trip — from deal context to deal record.
Read
Prompt declares which HubSpot fields it needs — deal stage, contact name, last touchpoint. Conduit fetches them with scoped permissions.
Generate
Your team's LLM (Claude, GPT, Gemini — your choice, your keys) drafts the output: follow-up email, deal update, QBR brief.
Write back
The draft goes back to HubSpot as a deal-record update, a new note, or a queued task — with optional approval before it lands.
Attribute
When that deal closes, Beacon weights this prompt's contribution against other executions in the prior 30 days. Time-decay credit.
Tie every closed-won deal back to the prompts that touched it.
When a HubSpot deal closes, Beacon walks back through the executions in the prior 30 days that read or wrote that deal record. Each prompt earns a credit-weighted share, decayed by recency. Your QBR finally has a number for "AI's revenue contribution."
ATTRIBUTED VALUE: $58,800 · MODEL: time_decay_v1 · HALF-LIFE: 7 DAYS
Built so revenue can finally trust AI's contribution.
Per-prompt field scoping
Every prompt declares which HubSpot properties it reads — and which (if any) it writes. The connector enforces this at run time. No prompt can see fields it didn't ask for.
Optional approval queues
Mark a write-binding requires_approval: true. Drafts queue in the approvals inbox; a library manager reviews and approves before the change lands in HubSpot.
Encrypted OAuth tokens
HubSpot OAuth tokens are encrypted at rest, never returned in API responses, and rotated independently. One connector per org; revoke any time.
Works alongside Slack + Teams
Run a HubSpot-aware prompt from Slack /prompt or Teams @CommonPlace — the connector context flows in regardless of the runtime. Same library, same attribution.
HubSpot-specific questions
Give revenue the AI attribution it's been asking for.
Connect HubSpot in 60 seconds. Define what counts as an outcome. Wait one nightly attribution sweep. Watch which prompts move money.
hello@getcommonplace.work