← Back to the Agent Playbook
Interactive walkthrough · Play R-05

Watch the Context Foundation upgrade every HubSpot agent

Click through how JayOh builds the context every HubSpot agent reads: ICP, personas, offers, competitors, voice and proof, extracted from your own material, scored for completeness and kept current.

7 steps~3 minutesUse ← → keys

Click the numbered orange boxes in the guide to jump to each part of the screen. Illustrative mockup with fictional data.

How the Context Foundation play works

Context Foundation structures your ICP, personas, offers, competitors, brand voice, proof points and playbooks into HubSpot's context layer, so every agent writes and answers like someone who knows your business.

The problem

Generic context gets generic results. Most HubSpot portals feed agents almost nothing about who they sell to, how they win or what proof they have, so drafts sound like a template and answers miss the point.

How the play runs

  1. Gather playbooks, positioning, win/loss notes, call transcripts and case studies
  2. Extract and structure them into HubSpot's context layer, section by section, with your team's approval
  3. Score completeness and close the gaps, starting with personas, competitors and proof
  4. Set an owner, a refresh cadence and stale flags

What JayOh builds

A structured context pack, a completeness score, persona and competitor sections built from real calls, proof points tagged by segment, and a governance cadence.

What we measure

  • Context completeness score
  • Blind-rated draft quality before and after
  • Rep edits per draft
  • Output quality across every agent

Guardrail

Stale context is worse than none. Assign one owner, review monthly and update immediately on launches and pricing changes.
HubSpot featureContext layer for agents
TeamRevOps, Marketing, Sales
Timeline2-week quick win
FeedsEvery HubSpot agent

Context Foundation FAQ

Why do HubSpot AI agents give generic output?

Usually because they have little context about your business. Agents read the context you give them, such as ICP, personas, offers and proof, before writing or answering. Thin context means generic results.

What goes into an agent's business context?

Company and product descriptions, ICP rules and disqualifiers, buyer personas and pains, offers and pricing, competitors and how you win, brand voice, proof points and sales playbooks.

Where does the content come from?

Mostly from material you already have: positioning decks, win/loss notes, call transcripts, pricing pages and case studies. We extract and structure it, then your team approves each section.

How do you keep context from going stale?

Each section gets an owner, a monthly review and stale flags for pricing, competitors and launches, so agents don't repeat outdated information.

Related plays

Want every agent to sound like your best rep?

We score your current context, build the foundation from your own material and show the before-and-after drafts in two weeks.