AI Skill Source Integrations
HubSpotHubSpot

Turn HubSpot customer context into AI Agent Skills

Develop AI Agent Skills from the reasoning inside contact histories, deal notes, and customer conversations. Preserve who needs what, which commitments are real, and what an experienced account owner would ask next.

Carry the judgment behind the next conversation

Source study / One deal, three perspectives

Discovery note / May 6

A clear need, with a test still to define

“Our dispatchers lose time handing jobs between regions. A pilot would need to show that those handoffs improve.”

Maya can explain the operational problem and help define a useful pilot. The team still needs to agree how it will recognize improvement.

Pilot success criteria / To confirm

Follow-up email / May 9

A budget conversation with conditions

“Send the pilot scope and total cost. I can review funding once we have those, but I have not approved it yet.”

Alex’s willingness to review funding is conditional. Preserve the missing scope and cost so a promising reply is not mistaken for budget approval.

Budget approval / Pending review

Technical review note / May 10

A start date depends on another review

“We need to review the data flow before scheduling the pilot. Our review slot is still being confirmed.”

The proposed start date depends on a review that has not been scheduled. Keep that dependency attached to the next conversation.

Data review / Not yet scheduled

An illustrative relationship map built from fictional CRM records. Select a contact to examine the context they add to the same deal.

The opportunity

A deal stage is a summary The expertise is in the conversation

Your experienced account owners know which reply changes the picture. They recognize an unresolved buying condition, connect an objection to the right person, and ask the question that makes the next step useful.

Develop that judgment into AI Agent Skills. Give an agent the reasoning to prepare a better conversation, surface a missing commitment, and carry the customer’s context into a handoff.

Customer needWhat must improve?
Decision authorityWho can commit?
ReadinessWhat is still open?

01 / From activity to a useful question

Teach an agent what to ask next

Retain the source, the expert’s interpretation, and the follow-up it calls for. That connection is the starting point for a skill your team can examine and test.

Reasoning study / Regional rollout

1 / Source

Discovery note
“A pilot would need to show that those handoffs improve.”

2 / Interpretation

Define what a useful pilot proves

The problem is specific. The success measure is still open. An expert asks how the customer will judge the outcome before discussing a wider rollout.

3 / Ask next

What would Maya need to see to call the pilot successful?

Confirm with the operations lead

Keep the limit attached

Do not invent a target from the existence of a business problem.

1 / Source

Finance follow-up
“I can review funding once we have those, but I have not approved it yet.”

2 / Interpretation

Retain the condition behind the reply

Funding review depends on scope and total cost. An expert keeps those prerequisites visible and confirms who has authority for the proposed purchase.

3 / Ask next

What scope and cost does Alex need, and who can approve this purchase?

Confirm with the finance director

Keep the limit attached

An open conversation about funding does not establish approval.

1 / Source

Technical review note
“Our review slot is still being confirmed.”

2 / Interpretation

Separate the hoped-for date from a commitment

A review must happen before scheduling the pilot. An expert asks who owns that review and what it could change about the proposed start.

3 / Ask next

Who will confirm the review slot, and how does it affect the pilot date?

Confirm with the IT lead

Keep the limit attached

Leave the start date unresolved until the responsible owner confirms it.

Candidate questions for expert review. The account owner confirms the interpretation and decides what to do next.

02 / Make the judgment reusable

Give the next account owner a better starting point

A useful skill explains how to read a customer situation. Keep the evidence, unresolved questions, and ownership together so another person or agent can continue with context.

Explore skill validation
Skill development brief

Read the buying conditions

Recognize
A positive reply that still depends on evidence, approval, or another person’s action.
Investigate
Connect each condition to its source, the customer it affects, and the person who can resolve it.
Test
Compare the reasoning with a different account. Ask an expert where it applies and which signals would change the next question.

Source context → Expert review → Candidate AI Agent Skill

Begin with a customer story

Which conversation changed how your team understood a deal?

Bring the contact history, the turning point, and the account owner who can explain it. Explore how that experience could become an AI Agent Skill for your next customer conversation.

Explore your HubSpot sources