The question worth asking isn't "where should we use AI." It's "what should the operating model look like, with AI as one of its components."
As fractional operations lead for a US-based AI startup at MVP stage, the temptation was to let the product's own AI capability define how the company operated internally too. That's a mistake. The company still needed SOPs, workflows and execution frameworks that had nothing to do with the product's AI, just to get an MVP out the door reliably.
Build the operating foundation first
Before optimising anything with AI, we established the operational backbone: documented SOPs, a recruited and structured remote team with clear ownership, and a centralised project management system for the product development cycle. AI adoption came after that foundation existed, not instead of it. The same sequence, foundation before automation, is what drove the results in building an AI operating system for an e-commerce agency, where a 60-80% cut in manual processing only happened after ClickUp was established as the operating backbone.
Where AI actually belongs
Across engagements, I use the same toolstack, Claude, ChatGPT and Higgsfield for AI capability, n8n, Zapier and Make for automation, wired into whatever work operating system the business already runs on, ClickUp, Notion or Asana. The tools aren't the point. Deciding what stays human, what becomes a repeatable process, and where AI creates real leverage, is.
Business first. Technology second. AI where it matters.
A framework for the decision
For any piece of work, ask three questions: does this require human judgment that can't be codified, is this already a repeatable process that just needs to run faster, or is this pure execution that AI or automation can fully own. Most businesses skip straight to automating without answering the first two questions honestly. This is essentially the Engineer and Navigate stages of the ORQENA framework applied specifically to AI adoption decisions.
What this looked like in practice
At the AI startup, that meant AI was central to the product but nowhere near the operating model until the basics, team structure, SOPs, execution cadence, were solid enough to support faster iteration. Get that sequence backwards and AI just accelerates a company that doesn't know what it's accelerating toward.
Two businesses, one operating principle
The AI startup and the e-commerce agency looked nothing alike on the surface, one was building an AI product, the other was a marketing agency using AI as a delivery tool. But the operating model question was identical in both: what belongs to a human, what belongs to a documented process, and only then, what belongs to AI. That question, not the specific tools, is what ORQENA's capabilities are actually built around.
What an AI operating model is
An AI operating model is not a stack of models and not a list of tools. It is the decision, inside the operating model, about what stays with a person, what becomes a repeatable process, and what AI is allowed to run. For a growing business, that decision comes after the operating foundation exists, not instead of it.
What AI transformation consulting means here
AI transformation consulting, here, means changing how the business operates so AI has somewhere real to sit. It is not a tool rollout, not a hiring plan, and not a tour of model types. The sequence is the same as the rest of ORQENA: observe how work actually moves, engineer the operating system, then apply AI where it creates speed, consistency or capacity.
What a buyer should expect
A useful engagement starts with the constraint, not the model. We look at the workflow, the owner, and the decision that keeps returning to the founder. Only then do we decide what to automate, what to augment, and what to leave human. If the founder is still the system, that is the problem to fix first — the same one in the founder bottleneck.
What is an AI operating model?
An AI operating model is not a stack of models and not a list of tools. It is the decision, inside the operating model, about what stays with a person, what becomes a repeatable process, and what AI is allowed to run. For a growing business, that decision comes after the operating foundation exists, not instead of it.
What does an AI transformation consultant do?
They do not start by selecting a model. They decide where AI belongs in how the business already runs, then build that into the operating system: process, ownership, and the few places a tool creates leverage.
How is an AI target operating model different from buying tools?
A target operating model says how the business should work once AI is part of it. Tools come after that picture exists. Another assistant, before the workflow is clear, only speeds up a process that was never designed.