For manufacturing CEOs and operating leaders · November 10, 2026
We already use AI. What should a company AI plan add?
A company AI plan earns its place when it helps leaders make a better operating decision across the business. Start by identifying what existing programs already cover. If the next investment adds no useful capability, the plan should say so.
By Dan Caulfield · CEO, Smokejumper
Respect the work underway
Begin with the current plan.
Your team may already have AI in engineering, finance, customer service or the ERP roadmap. The people responsible for those programs should help define the next decision. Ask what is working, what is already funded and which improvements are expected without another project.
A company-level assessment should make those commitments visible. It should also identify whether a useful priority falls between departments, competes for the same employees or depends on a decision no one has been asked to own.
The answer may be to support an existing program, change its sequence, test one bounded addition or defer new work. A longer list of AI ideas is not enough to justify another investment.
An incremental-value test
Compare the next step with a credible alternative.
Choose one operating outcome and describe two paths: what the company expects from its current plan, and what would change under realistic performance and failure assumptions. Give both paths the same time horizon and constraints.
For example, a team might propose help with a planning exception that requires inputs from sales, production and purchasing. Before building anything, check whether the existing system rollout already addresses it. Then ask the employees who resolve those exceptions what information is missing and which decisions require their judgment.
This is an illustrative planning situation, not a client example or a claim that AI will improve the result. If process clarification or an existing system feature is the better next step, choose it.
Bring this to the leadership meeting
Write a decision card for one priority.
Use these six prompts to make the proposed investment reviewable. Unanswered questions belong on the card.
Destination: What should work differently for the company? Name the operating outcome and the decision it supports.
Current coverage: Which existing team, system or funded program addresses this outcome? What is expected if we add nothing?
Employee evidence: Who does the work today? Review ordinary cases and exceptions with them. Record the baseline, the missing information and the work a change would add.
Accountability and boundaries: Who owns the operating result and can accept or reject a recommendation? Identify approved information, access limits, dependencies and the specialists responsible for unresolved questions.
Smallest useful evaluation: What limited comparison would help make the decision? Include employee time, specialist effort, review and support in the cost. Agree the evidence and acceptance criteria before evaluating.
Next decision: Set a review date and choose among continuing, changing, joining an existing program, deferring or stopping. Name who decides and who carries the next phase.
Keep the card short enough to discuss. Its purpose is to expose a consequential choice, with evidence someone can check.
From plan to execution
Give the work people, time and a release decision.
If an evaluation is justified, protect time for the employees who know the work and bring in the specialists the problem needs. Define what they will develop together, how people will learn to use it and who decides whether it is ready for live use.
That is the purpose of Smokejumper’s joint virtual R&D Lab model: work alongside company employees on a phased plan with accountable owners and measures. An evaluation can also show that a proposed capability should not proceed.
For related decisions, see ERP and AI sequencing, cross-department handoffs and the company plan.
A useful first conversation
What would the next investment add?
If you lead an established defense supplier or regulated manufacturer, bring one operating priority and the programs already addressing it. A free 20-minute introduction with Dan can help determine whether a company-level assessment and phased plan would be useful.
Schedule a free 20-minute introduction with Dan →
For the introduction, you call Dan at 760.803.7619.
