AI in Manufacturing Starts With People

Interview by Nishkam Batta, Editor-in-Chief, HonestAI Magazine

Manufacturers are being presented with more AI tools, platforms, and use cases than ever. But having access to better technology does not automatically make an AI implementation successful.

For Doug Hardesty, Principal and Founder of Hardesty Operating Group, successful AI adoption begins somewhere much more familiar to manufacturers: people, processes, data, and clearly defined problems.

Hardesty brings 30 years of manufacturing experience, spanning work with large global organizations as well as smaller companies. Today, his work includes helping companies implement systems, solve operational problems, and think strategically about AI adoption.

In a recent conversation with Nishkam Batta, Editor-in-Chief of HonestAI Magazine, Hardesty discussed why manufacturers should resist treating AI as simply another technology implementation and why the companies that build the right foundations today may be better positioned to capture AI's value tomorrow.

The People on the Floor Come Before the Technology

When manufacturers think about AI readiness, data is usually one of the first topics to come up.

Hardesty believes the conversation needs to begin even earlier.

Manufacturing has already been through decades of technological change, from automation and robotics to more recent digital initiatives. That experience can make manufacturers more familiar with technological transformation than organizations in some other industries.

But familiarity with change does not eliminate the need for employee buy-in.

Hardesty argued that manufacturers need to build a coalition around AI before becoming consumed by the technology itself. That means understanding how employees work, involving the people closest to the problems, and finding internal influencers who can help others see the value of change.

Those influencers may not always sit in leadership positions. They may simply be the people colleagues trust and listen to every day.

Fix the Process Before You Automate It

Once manufacturers establish internal support, the next question becomes: where should AI actually be applied?

Hardesty's answer is practical. Look for use cases capable of delivering visible benefits.

That could mean opportunities on the factory floor, in logistics, warehousing, procurement, or elsewhere in the operation.

But a quick win should create value for more than the company's financial statements. Employees using the system should also see an improvement in their daily work.

There is another important warning: manufacturers should not use AI to simply automate an inefficient process.

Before implementing technology, teams need to ask whether the existing workflow is actually the workflow they want to carry forward.

AI may make a process faster. But if the underlying process is poorly designed, manufacturers risk making the wrong process faster instead of fixing it.

Smaller Manufacturers Don't Need to Start With Massive Systems

The AI conversation can sometimes make implementation sound like something reserved for manufacturers with large IT departments, sophisticated data teams, and major technology budgets.

Hardesty sees a different path for smaller manufacturers.

Strategic thinking, he argued, can come from anywhere in an organization.

Employees working with a process every day often already know where the problems are. Even in a small company, talking to the people doing the work can reveal where time is being lost and where improvements could create value.

And building an AI foundation does not necessarily mean immediately implementing a large enterprise platform.

For some manufacturers, the starting point could be structured spreadsheets and better discipline around what information is being captured.

The important question is not simply, "What AI can we implement today?"

It is also: What data will we wish we had collected a year or two from now?

Data Collection Is Becoming Part of Manufacturing Infrastructure

For manufacturers building new operations, Hardesty believes data collection should be considered from the beginning.

Equipment, processes, and systems should be designed with the understanding that operational data will eventually be needed for analysis, problem-solving, and AI.

For established plants, the starting point may look very different.

Information may still exist on paper, inside spreadsheets, across machines, or within the experience of individual employees.

That does not mean manufacturers should wait for perfect data. The first step is to begin collecting what is available and bringing it together.

Machine data can provide one starting point. From there, manufacturers can expand into planning, logistics, warehousing, and supply-chain information.

The objective is not to build everything at once. It is to start somewhere useful and keep improving the data foundation over time.

Start With the Problems You Already Understand

One of the strongest themes from the conversation was the importance of focus. Manufacturers do not need to pursue every possible AI opportunity simultaneously.

If a company has a recurring operational problem and already has years of information around that problem, that may be a logical place to begin.

Problem-solving notes, machine information, spreadsheets, historical decisions, and other operational records may contain useful knowledge even when the information is not organized into a sophisticated data architecture.

Hardesty's advice is straightforward: start using what you have, learn from it, and continue building.

This also means aligning AI investments with a company's core competencies.

Rather than attempting to become world-class in every business function at once, smaller manufacturers can identify the one or two capabilities that matter most to their growth and build stronger systems around them.

AI Readiness Should Match the Company's Stage of Growth

A startup, a growing manufacturer, and a mature global operation should not approach AI in exactly the same way.

Hardesty described AI readiness through the lens of company maturity.

A founder-stage company may be focused on winning its first contracts, shipping orders, or raising capital. A company entering rapid growth may be implementing its first formal systems. A larger organization may already have significant infrastructure but face a different problem: connecting data and processes across functions.

The AI strategy should reflect that reality.

The goal is to build enough structure today that the organization does not reach its next stage of growth only to discover that the information it needs was never captured.

For smaller manufacturers in particular, this can make AI preparation less about buying technology today and more about developing data discipline for tomorrow.

The Next AI Breakthrough May Be Better Human-Machine Collaboration

Looking toward 2027, Hardesty expects the bigger change to come not only from what AI can technically do, but from how manufacturers learn to adopt and implement it.

He believes the industry is still working out which use cases create meaningful value and how AI should fit into existing organizations.

One area he sees as particularly promising is deeper collaboration between workers and machines.

Traditional industrial automation can perform a task repeatedly, but when something goes wrong, diagnosing the problem may still require experienced employees to investigate equipment, alarms, and process conditions.

AI could change that interaction.

Instead of equipment merely generating an alarm, Hardesty envisions systems capable of giving workers more context about what is happening and helping them reach root causes faster.

For manufacturing teams, that could turn the relationship with equipment from one-way monitoring into something more collaborative.

AI in Manufacturing Is Not Just a Technology Decision

The conversation ultimately returned to where it began: people.

Manufacturers can invest in sophisticated AI systems, but technology alone does not create adoption. Nor does it guarantee that the right business problem is being solved.

Successful implementation requires manufacturers to understand their processes, involve the people closest to those processes, collect useful data, identify achievable use cases, and build capabilities appropriate for their stage of growth.

AI may eventually transform how factories make decisions, diagnose problems, and coordinate operations.

But the foundation is much less futuristic.

It starts with knowing what problem matters, collecting the information needed to understand it, and bringing people along as the organization changes.

And that may be the more useful way for manufacturers to think about AI, not as technology replacing the organization they already have, but as a new capability that works alongside their people to help the organization operate better.

About Doug Hardesty

About the Interviewer