The Goats of Growth podcast

How a CMO Rebuilt Marketing Around a Shared AI Operating System — Steve Wooledge, CMO at Collate

0:00
47:33
Rewind 15 seconds
Fast Forward 15 seconds

Steve Wooledge didn't build a bigger marketing team. He built a shared AI operating system, and Collate no longer has a social media hire or a customer case study writer. The contractors who covered that work came in house.

Steve is CMO at Collate, the semantic intelligence platform built on open metadata. He has spent 20 years in data and analytics marketing at Alation, Alteryx, Hazelcast and Teradata. Collate is the commercial backer of the Open Metadata Project, and he explains why a context layer matters when AI agents query company data.

His central argument: AI is only as trustworthy as the context it can see. He says Claude Opus pointed at a data warehouse gets roughly 60% of answers right, and Collate's benchmarks with context reach 94 to 95%. The same idea runs his own marketing org. The team keeps its skills, messaging and memory in one GitHub repo and updates it through pull requests with peer review, about 50 in the last six months.

The concrete number: Steve puts AI at roughly 20% of his marketing budget saved, and thinks 30 to 33% is reachable. He's clear that's a gut estimate, not a measured one. We also get into where he reinvests it (organic over paid search and social), how his SDRs use Common Room signals, and the 7:30 a.m. email his AI second brain sends him from every call.

WHAT WE COVER

  • Why AI answers on structured data run about 60% accurate without context
  • How Collate's context layer gets benchmarks to 94-95%
  • A team GTM operating system in GitHub: skills, messaging, pull requests
  • Cursor for non-coders, and why he keeps his "brain" local
  • The roles Collate didn't hire after bringing content in house
  • The 20% budget saving, and why he pushes it toward 30%
  • How he evaluates AI tools: referrals, not demos
  • Fathom call summaries feeding a daily action-item email
  • Rapid Fire 5, Jessica Storry's bonus question, and Steve's mulligan question for the next guest

CHAPTERS

0:00 Intro and Common Room

1:26 How Collate uses Common Room signals

4:16 What Collate does: a context layer for AI

7:11 From AI noob to the "AI CMO"

10:22 Cursor, explained for non-coders

13:41 Whisperflow and Friday team AI sessions

16:14 Claude Code from your phone

17:41 What AI did to marketing velocity and hiring

20:03 Quantifying the savings: 20% to 30%

24:19 Where the savings go: organic over paid

26:05 Sales cycles and deal velocity

26:58 How Steve evaluates AI tools

29:13 Why human content matters more now

30:09 The Fathom second brain and 7:30 a.m. email

32:50 The shared GTM OS: pull requests and peer review

35:32 Rapid Fire 5

42:36 Bonus question: starting over, not in tech

44:08 Steve's question for the next guest

45:47 Where to find Steve

Connect with Steve Wooledge on LinkedIn, or learn more about Collate at getcollate.io.

This episode is sponsored by Common Room (https://www.commonroom.io/?utm_source=goat-podcast&utm_medium=partner), the AI GTM Intelligence Platform trusted by teams like Okta, Anthropic, and Fireworks AI.

New conversations with go-to-market leaders every week. Subscribe to The GOATS of Growth on Apple, Spotify, or YouTube.

More episodes from "The Goats of Growth"