What deal risk reveals about the AI-native GTM system

Most people will watch this video and see AI automating Deal Risk.
But it's missing the underlying shift that defines AI-native GTM 👇

→ a "Deal Agent" reviews your CRM, calls, emails...
→ identifies what's changed
→ explains risk
→ recommends what should happen next

It's useful. Faster. Easy to understand.
But I think it's the least interesting part.

I’ve found myself using this example in almost every conversation with forward-thinking VPs of RevOps recently.

Not because they're looking for another bolt-on AI workflow.
Because it reveals a deep problem they've been experiencing for years
- but haven't been able to describe.

Every commercial team is trying to answer the same questions:

→ Which deals are healthy?
→ Which are drifting?
→ Where should we focus next?

Today, answering those questions means reconstructing what’s happening from CRM records, call transcripts, emails, meetings and activity.

→ rep prepares for a call
→ manager reviews the pipeline
→ RevOps builds another report

Everyone is trying to answer the same question:

→ what’s actually happening?

Adding AI makes that reconstruction faster.

It can summarise the call, analyse the CRM, assess deal risk and recommend next steps. The work becomes faster. The organisation still has to recreate its understanding every time it wants an answer.

That’s the hidden work most GTM teams have quietly accepted for years.
And I think that’s the transition we’re living through.

Most companies are adding AI to existing work.

The more fundamental shift is building an operating layer that maintain understanding continuously.

If a system is going to do that, it needs more than an agent.
It needs to know what’s happening.

💡 That’s Live State

A continuously maintained understanding of the opportunity, not just a collection of records.

It needs to know how your company thinks.

💡 That’s Memory

Your qualification standards.
Your commercial judgement.
Your management principles.
Codified once.
Applied consistently.

It needs to know how to apply that knowledge.

💡 That’s Skills

→ When is a deal genuinely at risk?
→ When should a manager step in?
→ When should forecast confidence change?

Only then does the Agent become interesting.

It combines live state, memory and skills to understand what’s happening, decide what matters and coordinate the next action where people already work.

And every decision leaves behind something valuable:

→ evidence
→ reasoning
→ action
→ outcome

Now the system can learn.

That’s why I don’t think Deal Risk is really the story.
It’s simply one capability that reveals where GTM is heading:

storing information → maintaining understanding
automating tasks → operating capabilities
individual judgement → organisational memory
periodic reviews → continuous awareness

We’ve spent the last 20 years building commercial software that stores information.

I think we’ll spend the next 20 building systems that continuously understand, decide, act and learn.

Watch the original video on LinkedIn

Originally posted on LinkedIn