← Dissel AI · Perspectives

Operating model

The agentic operating model - and why most companies will get it wrong.

AI is not a feature you ship. It is a redesign of how decisions are made, who makes them, and at what cadence. A field guide for executives.

Derk Disselhoff·Founder, Dissel AI·May 2026·12 min read

For decades we lived inside the same trade-off. The business wanted to move. Tech said it would take two quarters, three vendors, and a steering committee. We called that normal. Roadmaps got shorter, slide decks got longer, and the gap between the operating model the founder had in their head and the one the company actually ran on kept widening.

Most of what we labelled "digital transformation" was just paving over that gap. New CRM on top of the old process. New dashboard on top of the old reporting. The shape of the company stayed the same because the cost of changing it was too high.

ERP
CRM
PIM
CMS
ATS
OPS
FINANCE
DOCS
Legacy stack Unified core
Live
Fifteen years of scattered systems collapse into one core - the same data, finally in one place.

The level playing field

AI changed the cost curve. Not in the way the headlines suggest - not by replacing people - but by collapsing the distance between deciding something and shipping it. A workflow that used to need a product manager, a backend team, and a six-week sprint can now be drafted on a Tuesday and run in production by Friday.

That is the part most companies miss. They are using AI to speed up the work the old operating model was designed to produce. Faster emails. Faster summaries. Faster tickets. The operating model itself - who owns which decision, on what cadence, with what hand-offs - stays untouched. Same shape, slightly faster engine. The compounding never arrives.

The winners will not be the companies that added AI to their workflow. They will be the ones who redesigned the workflow around it.

- Derk Disselhoff

The founder is the architect

Every company is shaped like its founder's mental model of how the work should flow. That model lives in their head with a level of fidelity nobody else in the building has. They know which step is fake, which approval is theatre, which hand-off only exists because someone left in 2019 and nobody re-cut the process.

For thirty years the founder could not act on that knowledge directly. They had to translate it into requirements, hand it to a build team, watch it get diluted, and accept whatever came back. The architect was never holding the tools.

Company brain
Live
Role-scoped permissions
Audit · BigQuery
DATA WAREHOUSE
ANALYTICS
CMS
CRM
COLLABORATION
DOCS
SEO
COMMERCE
One agent. One governed data layer. Every system. Every query audited.

Now they are. Agents collapse the distance between the model in the founder's head and the system the company runs on. You design the workflow, the agent runs it, and the next iteration ships the same week. The architect and the builder are finally the same person.

What an agentic operating model actually looks like

It is not a chatbot bolted onto support. It is a coherent redesign across the four layers of the business:

  1. 01Commercial - sourcing, qualifying, quoting, and renewing run on agents that pull from one unified data layer, with humans owning the handful of decisions that genuinely need judgement.
  2. 02Data - one schema, not five. Products, contacts, interactions, pipeline, and content sit in the same place, queryable by humans and agents alike. Integrations stop being a project.
  3. 03Operations - planning, fulfilment, and exception handling get rebuilt as workflows that decide and act, not as dashboards that wait for someone to notice.
  4. 04Admin - finance, HR, and compliance shrink to a thin layer of policy and review on top of agents that do the work.

Each layer reinforces the next. Better commercial data feeds better operations. Better operations data feeds better admin. Leverage compounds because the model is coherent, not because any single use case is brilliant.

Operating model · four layers, one system
01Commercial
Source · qualify · quote · renew
02Operations
Plan · fulfil · exception handling
03Data
One schema · queryable by all
04Admin
Finance · HR · compliance
Commercial, data, operations, admin - four layers on one system. Each one feeds the next.

Why most companies will get it wrong

They will treat this as an IT project. A budget will get approved, a steering group will be formed, a vendor will deliver a pilot, and the operating model around it will not move an inch. Six months in, the company will be running the same shape it ran before, with a slightly more expensive tech stack and a new committee.

The companies that win will do the opposite. The founder, the CEO, or the owner-operator picks up the pen. They redraw the model first - on paper, in a room, with someone who knows how to build. The agents come second. The org chart comes third. By the time it ships, the company is genuinely shaped differently, and the leverage shows up in the P&L, not in a slide.

Capex goes down. Revenue goes up. The reason is not the agent. The reason is that the company is finally shaped the way the founder always wanted it shaped.

- Derk Disselhoff

The invitation

If you are the person who can change the shape of your company, this is the window. Not to add AI to what you have. To rebuild what you have around it. Bring the model that is in your head. We will bring the agents, the platform, and a team that knows how to ship into production. The architect is you. The tools are finally here.

Further reading