Apply
Agent strategy is a portfolio decision, not a technology preference.
The question is not whether the company should use agents. It is which workflows are expensive enough, feasible enough, and valuable enough to deserve one.
How I see it
Agentic AI strategy
Agentic AI strategy is how you decide which expensive workflow deserves an agent, and which ideas you will not fund.
Most teams write a vision of an agentic enterprise, then fund the loudest demo. The expensive work is still gathering context, interpreting documents, updating systems, and chasing the next step.
They do that because they treat strategy as a technology preference. I have watched a capable agent pursue a goal the business did not actually want. A high-value idea with no owner and no system of record is not a first project. A better engine on the wrong objective just arrives at the wrong result faster.
You do not need a multi-year program. You need a short ranked set: do first, investigate, defer, do not fund. And a first win a COO can see in a quarter.
Common mistakes
What teams usually get wrong.
Strategy as a list of tools
Model vendors and platforms are inputs. They are not a strategy for where the company will create leverage.
Funding the loudest use case
A polished demo from one function can beat a quieter workflow that is worth five times more.
Waiting for a complete operating model
You do not need a new operating model to run one production agent. You need a workflow, a baseline, and controls.
A useful diagnostic
Five questions before you fund the work.
Do you have more than five AI ideas and no ranking method?
You need prioritization before architecture.Can you name the economic surface area of the top workflow?
If not, the strategy is still a conversation, not a plan.Is there a do-not-fund list?
A strategy that funds everything is a wish list.Does the first project have an executive owner?
Orphan projects do not become operating leverage.Would a no-agent path still be considered?
If every answer is an agent, the strategy is ideological.
Economic model
Priority scorebusiness value × feasibility × time-to-value ÷ (risk × implementation cost)
Use this to compare opportunities, not to pretend precision. The ranking is the product.
Three credible paths
How far should you go?
Do not force one solution. Choose the path the economics, the risk, and the organization can support.
Improve the work you already have
Keep the process mostly intact and use AI on the bottlenecks that create delay, rework, or follow-up.
The workflow is already sound and a few steps create most of the friction.
Gains are usually incremental. The operating economics do not change much.
Redesign the workflow around AI
Question every handoff, queue, and duplicate step, then rebuild the process around what AI can now do.
The process grew over years and coordination now costs more than the work itself.
Requires process change, clearer ownership, and a willingness to retire old steps.
Put an agent on a high-value outcome
Give an agent responsibility for one valuable result across systems, with humans at the control points that require judgment, authority, or risk acceptance.
The workflow is high-value, variable, multi-step, and worth engineering for production.
Needs stronger architecture, evaluation, controls, and monitoring. This is not a prompt project.
When this is the wrong next step
Do not fund an agent here.
- The company has not used basic AI tools yet and needs adoption before agents.
- There is no executive who can choose one workflow and protect it from a program.
- The real problem is org design, pricing, or demand, not workflow cost.
A useful next step
Bring one workflow. Get guided into production.
We guide the implementation, go deep on the technical path, and stay hands-on through operations — or tell you when a simpler answer is better.
Discuss an AI opportunity