How to Measure AI ROI in Manufacturing: Start With the Workflow

Ask a manufacturing leader whether AI is helping and you’ll usually get a yes. Ask what it’s returning and the room goes quiet. That gap is the real problem with AI ROI in manufacturing: most companies picked up the tool before they decided what job it was supposed to do.

During a recent Moore on Manufacturing episode, Mike Sibley of James Moore shared practical insights on how manufacturers can tell whether AI and automation are paying off. The discussion highlighted the importance of tying every AI investment to a specific workflow before trying to measure it.

The Cost Side Is Moving Too

Return is only half the equation, and the cost half isn’t standing still. AI pricing increasingly runs on credits and usage, and as Mike pointed out, that cost is going up. Companies that roll out AI without usage limits can blow past their credits before anyone notices.

So the first step isn’t measuring output. It’s knowing what you’re spending. That means setting limits, monitoring how employees use the tools and understanding how the pricing actually works. “That can get complicated,” Mike said. It’s also a big reason so many companies are still trying to figure out how to measure ROI at all.

Tie AI to a Workflow or You Can’t Measure It

Here’s where most AI efforts go sideways. Someone uses it to draft emails faster, everyone agrees it’s handy and nobody can put a number on it.

Mike’s advice is direct: “Build it into your workflows.” Finance is the easiest place to start. Producing the monthly reporting deck, financial reports and KPIs is repeatable work with a known time cost. Automate it and you can see exactly what came back. Accounts payable is another good candidate, since getting invoices into the system faster is easy to track. Engineers are building it into their own processes too.

The payoff goes beyond hours saved. When the finance team stops assembling reports by hand, Mike noted, “that gives them more time to actually be more proactive” and support other parts of the business.

Before any rollout, answer three questions. What workflow are we automating? What is it helping? How will we measure against it? If the first answer is fuzzy, the last one will be too.

Three Returns Worth Tracking

Capacity You Can See

The clearest signal is people getting more done in the same hours. Faster access to information counts here, with one condition: people still own the decisions. As Mike put it, “we don’t let the AI make the decisions for us.”

Hires You Didn’t Need to Make

Mike called this one of the easiest ways to show value. If AI lets you avoid a hire you otherwise would have needed, that’s a number any leadership team understands right away.

Soft Costs

Some returns never show up as a line item. Mike describes these as soft costs, the opportunity cost of time. When your team has room for work they couldn’t get to before, that has value even if it never lands on an invoice.

Count the Review Time

One thing that quietly eats into return is checking the work. AI can hallucinate or simply miss what you asked for, which is why, as Mike noted, even Anthropic tells users they still have to review the output. That review isn’t optional, so it belongs inside the workflow you’re measuring.

Start With One Workflow

Mike sees adoption lagging most among smaller manufacturers, roughly those under $50 million, because they don’t have the time or resources to put into it. That’s exactly why the starting point should be narrow. Pick one repeatable workflow. Know what it costs today. Then track capacity, avoided hires and review time once AI is running.

If your finance team still builds the monthly reporting deck by hand, how many hours a month is that costing you?

Watch the full episode to hear Mike’s take on tariffs, cash flow and the risk and opportunity planning he recommends for the next 12 months.

All content provided in this article is for informational purposes only. Matters discussed in this article are subject to change. For up-to-date information on this subject please contact a James Moore professional. James Moore will not be held responsible for any claim, loss, damage or inconvenience caused as a result of any information within these pages or any information accessed through this site.