Today is bloody special!
At
Ideas2IT Technologies , we have a large technology services business and 9 other products/platforms at different stages of GTM maturity.
The GTM team supporting all of this? 8 people.
That constraint forced us to go deep on AI over the last year or so. We've tried a lot, broken a lot, and thrown away a lot and I think we've finally landed on an AI-native GTM engine that actually works for us.
I laid out the structure for how I thought this should work.
Kevin Anderson and
Kavya Kotha then went and pulled off a freakin' miracle.
What's the setup? I'll focus on the two pieces doing most of the heavy lifting.
1) It all starts with context.
Every time we need to take a product/offering to market, the team gets me into a room for what we've started calling my "context rant."
These rants are really about narrative development. We unpack the market, buyers, competitors, customer problems, sales learnings, our worldview, what we believe others are getting wrong, and what we see changing before the market does. That becomes a mindmap of the narrative we want to shape, every argument, question, topic, keyword, counterargument and piece of evidence around it, and the body of work we need to create to earn the right to own that narrative.
They record the whole thing, extract what matters, and then basically ask me to go away 😂
2) Then the system.
The system now runs across
OpenKnowledge (highly recommend) as the repo, Claude (we use it extensively - code, design, workflows, agents and skills), Webflow, Search Console, Analytics,
OpenSEO and the rest of our GTM stack.
Bonus: getting the architecture right has also reduced our Claude token usage by more than half.
We have specialized agents across research, SEO/AEO, writing, editing, LinkedIn, newsletters, outreach and campaigns, supported by our own skills that encode brand, how we research, develop narratives, evaluate angles, write, edit, repurpose and publish. We also documented our best-performing assets and turned them into benchmarks, so outputs are validated against what has actually ranked, driven engagement and resonated with buyers rather than generic LLM quality scores.
Context captures what we know, skills capture how we work, benchmarks define our standard, and workflows put it all into execution.
Example: Research → opportunity → narrative/angle → writing → editing → SEO/AEO → validation → publishing.
Increasingly, this is just becoming a click-of-a-button exercise.
Now rinse and repeat for every workflow.
How much can an 8-person GTM team actually own when the operating model is designed around AI from the ground up?
Execution for 1 brand is 90% automated. By the end of Aug, we expect all 9 to be automated. Our services business is the trickiest one. Let's see how it goes.
Can you tell I'm happy? :D
#gtmengineering