Everyone wants AI agents right now.
Very few companies actually understand how work flows through their business.
I wrote about:
tribal knowledge
invisible operational labor
why AI amplifies chaos when systems are unclear
the future of operators in AI-first companies
and why “ask Sam” is not a scalable infrastructure plan
This one is for the founders AND operators watching org charts change in real time.
This is an excellent framing on why AI adoption is stalling in practice. The invisible labour list especially: the ops person fixing CRM data, the CS lead intercepting escalations, the manager translating chaos into work. Naming that as the actual operating system, not a support function, is the bit most AI discourse skips entirely. Really glad someone said it this clearly.
It sparked something I keep seeing from the other side. I build operational systems for scaling businesses, and the pattern you're describing is exactly what walks through the door. The common instinct is "we need to document everything first." But what I've found across 250+ engagements is that traditional documentation doesn't hold. You write the SOP, it's stale in a fortnight, everyone goes back to asking Sam.
The thing that actually works is embedding the knowledge into the tools. Build the workflow so that using it IS learning the process. The database captures the context. The status field makes the decision visible. The system becomes the SOP. Then AI has something structural to run on, not just tribal memory.
I wrote about the same diagnosis recently: The Knowledge Your Business Runs On Isn't Written Down (https://substack.com/home/post/p-195332534), your piece is the "why it matters now" that mine was missing.
"AI is about to expose every startup operating on tribal knowledge."
That line hits 🔥
AI isn't replacing operators.
It's exposing how much of a company's success was quietly being held together by operators all along.
The future isn't bigger teams.
It's smaller teams, AI leverage, and experienced people brought in to solve specific problems, build the right systems, and move on when the work is done.
At Fract75, we believe this is exactly where the market is heading.
Not more headcount.
Not more consultants.
Not more management layers.
At Fract75, we think the future belongs to a few founders obsessed with the problem, AI handling the repetitive work, and experienced operators brought in where judgment, execution, and clarity create the most leverage.
The companies that win won't have the most AI agents.
They'll have the clearest and most adaptable operating systems.
Great piece, Katie! Too many are trying to automate a process without understanding the complex web of dependencies, flows and nuance that need to be mapped out. That stage is often just.. missing.
Great practical take about missing decision context and the importance of tribal knowledge.
That's right 🔍👀
This is an excellent framing on why AI adoption is stalling in practice. The invisible labour list especially: the ops person fixing CRM data, the CS lead intercepting escalations, the manager translating chaos into work. Naming that as the actual operating system, not a support function, is the bit most AI discourse skips entirely. Really glad someone said it this clearly.
It sparked something I keep seeing from the other side. I build operational systems for scaling businesses, and the pattern you're describing is exactly what walks through the door. The common instinct is "we need to document everything first." But what I've found across 250+ engagements is that traditional documentation doesn't hold. You write the SOP, it's stale in a fortnight, everyone goes back to asking Sam.
The thing that actually works is embedding the knowledge into the tools. Build the workflow so that using it IS learning the process. The database captures the context. The status field makes the decision visible. The system becomes the SOP. Then AI has something structural to run on, not just tribal memory.
I wrote about the same diagnosis recently: The Knowledge Your Business Runs On Isn't Written Down (https://substack.com/home/post/p-195332534), your piece is the "why it matters now" that mine was missing.
Great post, Katie.
"AI is about to expose every startup operating on tribal knowledge."
That line hits 🔥
AI isn't replacing operators.
It's exposing how much of a company's success was quietly being held together by operators all along.
The future isn't bigger teams.
It's smaller teams, AI leverage, and experienced people brought in to solve specific problems, build the right systems, and move on when the work is done.
At Fract75, we believe this is exactly where the market is heading.
Not more headcount.
Not more consultants.
Not more management layers.
At Fract75, we think the future belongs to a few founders obsessed with the problem, AI handling the repetitive work, and experienced operators brought in where judgment, execution, and clarity create the most leverage.
The companies that win won't have the most AI agents.
They'll have the clearest and most adaptable operating systems.
Great piece, Katie! Too many are trying to automate a process without understanding the complex web of dependencies, flows and nuance that need to be mapped out. That stage is often just.. missing.
Written like someone who’s seen things!
I find it fascinating that so many smart people build a company but don’t realize the fragility of the system running it.
What have you found works for getting founders and executives accept the state of unclear manual work and systematize operations?