YOUR M.O. ISN'T YOUR O.M.
Yeah, it matters
Most people have a way of working. A rhythm. A set of instincts built up over years of figuring out what works, and what absolutely blew up in your face. That's your M.O., right? Your modus operandi. And it's valuable.
Howwwever, your M.O. only lives in your head and it's unique to you. Let’s be clear, individuality is important, and everyone needs to bring their unique talents and flair to situations and teams. Diversity (despite any political nonsense to the contrary) is crucial. They cannot replace you.
But when you're working as part of something bigger, towards shared goals, you'll need something else *cue drumroll * You need an O.M. An operating model.
And today, with AI-driven workflows and automation becoming an integral part of how functions operate, the gap between having one and not having one has never been more consequential.
What even is an operating model?
An operating model is the documented, repeatable infrastructure that governs how stuff gets done inside an organisation. It covers how decisions are made, how work flows from brief to delivery, how you connect to wider product or business outcomes, and how quality and consistency are maintained at scale. So how your work is connected outside of yourself.
It's not a strategy doc. It's not a process map gathering digital dust in Onedrive. It's the connective tissue between what you're trying to achieve and how the day-to-day actually runs.
If Susan is the beating heart 😍, Roger is the appendix (seriously, what does he even do?) and Trevor is the left lung - the operating model is the fascia. The thing that holds the whole structure together and lets everything move properly.
Why does this matter?
The more senior you are, the less your value lies in execution and the more it lies in enabling execution consistently; across teams, tools, time zones, and stakeholder expectations.
Without an operating model, everything runs on individual responses to the same stimuli. Decisions get made based on who asks, who asked, and how at the end of her tether Caroline is today - not what's been broadly agreed. Processes live in people's heads. Priorities shift because there's no clear framework for what should take precedence. I was recently asked why we even need to change things "because I know the priority." But the thing is, on a day-to-day basis, that doesn't live in everyone's head rent free. That’s when your deliverables slide.
Add AI into that environment and say goodbye to any hopes of efficiency, and hello to absolute utter chaos. Teams start adopting tools independently, so context ends up in twenty different places. There's no governance over what's AI-generated versus human-reviewed. Automations get built on top of undocumented processes, which means they automate the inconsistency rather than solve it. Shudder.
An operating model creates the conditions for your function to work without you having to be present for every decision. In an AI-driven environment, it also creates the conditions for your tooling to actually work. This isn't an operational nice-to-have anymore, kids. It's the foundation everything else sits on.
How to know you're missing one
A few diagnostic signs:
You make the same decision repeatedly with no documented rationale for why. Different stakeholders have different expectations of what your function does and who it serves. Onboarding a new team member means weeks of you talking them through everything from scratch. Work arrives without a consistent brief and leaves without a consistent quality check. You're being asked the same questions over and over (fuuun) by different people, in different channels, at different times. Sometimes the same though (soooooo fuuuuun).
….You spend more time managing ambiguity than doing the good stuff.
If you're still the answer to questions that could be systematised, you haven't finished the job.
How to approach building one
Like everything you do, start with purpose.
Before you document a single workflow, get clear on what your function actually exists to do inside the business. Not in the abstract; specifically. What outcomes does it drive? Who are the stakeholders it serves? What decisions does it own, and which ones does it inform?
From there, you can build out the four components that any operating model needs: a governance layer (so, who decides what), a delivery layer (how work moves from le brief to le output - and connects to les outcomes), a quality layer (how standards are set and maintained), and a measurement layer (how you know it's working).
You don't have to build it all at once. Start with whichever layer is creating the most friction right now, document it properly, test it with the people who touch it, and iterate. Build modular so it’s easily updatable.
What this could look like
Once your operating model exists - once the knowledge is documented, the processes are defined, and the decision-making is clear and consistent, you can start layering intelligent systems on top of it. Agentic workflows. Automated routing. AI that can answer questions, triage requests, and surface the right information to the right people without you in the loop for every touchpoint.
I've done this. I built a knowledge base that captured everything my function knew about a new service. Every question, most edge cases, every process; and then used an AI agent to make that knowledge accessible to the teams who needed it. GTM teams could get accurate answers instantly rather than routing through me. My active hours on that project dropped significantly. The teams got what they needed faster. And the knowledge base keeps updating as new questions surface, so the system gets smarter over time.
When I demonstrated this to the team recently, setting up the agent itself took about thirty seconds. Talk about anticlimactic…
What I will die on a hill about: the tool is the easy bit. The expertise is in building a knowledge base that's structured well enough for AI to synthesise accurately, and knowing how to prompt so it retrieves the right thing in the right way.
The agent didn't create clarity. It inherited it. The knowledge base had to be comprehensive and well-structured before the agent could do anything useful with it.
Didn’t get that? The automation was the last step, not the first.
If you think simply adopting AI tools will create order from chaos. You’re wrong, Josh. AI scales what already exists. If what exists is a stinking pile of … sweet nothings, that is what you'll scale.
The goal isn't a perfectly pretty and well formated document. It's a living system.
One that other people can use without you having to explain it every time, it’s accessible and there in the moment of need, not when you’re free to reply. One that new tools and automations can plug into without breaking. One that doesn't leave when you do, affords you to completely unwind on that 2 week Europe trip… and when you’re back, you can move on to the good stuff.
That, my friends, is when you know it's working.
