AI adoption is no longer the question for most businesses.

The bigger question is whether AI is actually changing how work gets done.

According to McKinsey’s 2026 research on AI and operational excellence, almost 90% of organizations are experimenting with AI, yet only 7% report scaling it across the enterprise. Adoption of more advanced technologies, particularly agentic AI, is even earlier, with at most one-quarter of organizations experimenting with AI Agents in any function.

There is a significant difference between experimenting with AI and putting AI to work.

For businesses looking to create measurable value from AI, the next step is not simply adding more tools. It is identifying where work slows down, understanding how that work moves through the business, and embedding AI directly into those processes.

The AI Experiment Phase Needs to Become the Execution Phase

Businesses have spent the past few years exploring what AI can do.

Employees use AI to summarize information, draft emails, analyze documents, answer questions, and accelerate individual tasks. Those capabilities can save time, but they do not necessarily change the underlying operation.

• The quote still needs to be created.

• The RFQ still needs to be processed.

• The customer information still needs to be entered into the ERP.

• The specification still needs to be checked.

• The project still needs to move to the next stage.

• Someone still needs to follow up when information is missing.

McKinsey’s research found a clear relationship between the scale of AI deployment and productivity improvement. Organizations embedding AI across more functions reported greater productivity gains, while businesses limiting AI to a small number of use cases saw more modest results. McKinsey concludes that AI’s impact comes not simply from experimentation, but from integrating it into core operational processes.

That distinction matters. The goal should not simply be to give employees access to AI. The goal should be to determine where AI can become part of the execution layer of the business.

Operational Excellence Gives AI Somewhere to Work

One of the most important findings in McKinsey’s research is that AI and operational excellence can reinforce one another.

Companies that combine advanced technology with strong operational practices report greater productivity improvements. At the same time, organizations with higher operational-excellence maturity tend to be further along in deploying AI. McKinsey describes this as a reinforcing cycle: productivity improvements create capacity for greater AI deployment, while scaled AI can further strengthen operational performance.

That makes sense.

An AI Agent is far more useful when the business can answer questions like:

• What starts this process?

• What information is required?

• Where does that information come from?

• What rules determine what happens next?

• What system needs to be updated?

• Who needs to approve an exception?

• What does “complete” actually mean?

When those pieces are clear, AI has a framework in which it can execute.

This is why AI adoption should not be treated solely as a technology initiative. It is also an opportunity to examine the work happening between people, processes, documents, and systems.

Start With the Work, Not the Technology

McKinsey recommends starting with a small number of high-value operational priorities instead of spreading AI investments across dozens of disconnected pilots. Those priorities should connect technology investment to meaningful outcomes such as throughput, service levels, yield, or asset utilization.

For many businesses, those opportunities are already hiding inside everyday processes:

• A salesperson receives an RFQ by email and manually pulls information from attachments.

• A coordinator takes information from a signed quote and reenters it into an ERP or DMS.

• A project manager compares specifications across several documents.

• An operations employee follows up with another department because required information is missing.

• Someone downloads a spreadsheet, cleans the data, reconciles it against another system, and uploads the results somewhere else.

These processes may look different, but they share a common characteristic: people are spending valuable time coordinating information and moving work between systems.

That is where AI Agents can become tangible.

Instead of simply answering a question about the work, an Agent can participate in the process itself.

It can read incoming documents, extract information, validate it against business rules, identify missing or conflicting data, search existing records, create or update records, trigger follow-up, route exceptions, and help move the process forward.

The AI becomes part of the operation rather than another tool sitting beside it.

AI Agents Need to Fit How Your Business Actually Works

McKinsey emphasizes the importance of building an operational backbone alongside the technology. Data foundations, performance management, governance, and clearly defined processes help ensure that technology outputs translate into consistent action.

This becomes especially important when businesses move from generative AI toward AI Agents capable of executing work.

An Agent needs context.

It needs to understand:

• Your documents

• Your customers

• Your systems

• Your processes

• Your business rules

• Your approval structure

When necessary, it also needs to know when a human should step in.

At Vsimple, this is why we believe AI adoption should begin with how your business actually operates.

Rather than asking:

“Where can we add AI?”

A better question is:

“Where is work getting stuck, and what would it take for an AI Agent to help move it forward?”

That shift turns an abstract AI strategy into an operational one.

The Mid-Market Has a Different AI Challenge

McKinsey’s findings are particularly relevant for midsize organizations.

The research found that midsize companies tend to score lower on operational excellence than both smaller and larger organizations. They can lack the simplicity and speed of smaller companies while also lacking the scale, capital, and institutionalized operating models available to large enterprises. McKinsey also found that these organizations tend to lag larger businesses in the breadth and maturity of AI adoption.

That creates an interesting challenge.

But it does not mean the mid-market needs an enterprise-sized AI transformation before it can see value.

It means choosing the right starting point matters even more.

Instead of attempting to transform the entire organization at once, businesses can identify specific processes where repetitive work, clear rules, accessible data, sufficient volume, and meaningful business impact make AI especially valuable.

Think about processes such as:

• RFQ processing and quoting

• Document intake and reconciliation

• Order processing

• Specification checks

• Project coordination

• Customer and project research

• System data entry

• Missing-information follow-up

• Approval routing

Solve one real operational problem. Measure what changes. Then expand.

That creates a practical path from experimentation to adoption without requiring the business to reinvent everything overnight.

Productivity Comes From Changing the Process

Perhaps the most important takeaway from McKinsey’s research is the connection between AI, operational excellence, and productivity.

Its survey found that companies with AI embedded across multiple functions generated nearly double the profit margins of peers using AI in only a few departments. Three-year return on invested capital was also more than five times higher. Importantly, McKinsey notes that these findings demonstrate correlation rather than causation, but the patterns suggest that stronger operating systems help companies translate AI investment into measurable performance.

For leaders, that changes how AI success should be measured.

The question is not simply:

• How many employees are using AI?

It is:

• What work no longer has to be done manually because of AI?

• How much faster can an RFQ become a quote?

• How much manual data entry can be eliminated?

• How many errors can be caught before they create rework?

• How quickly can missing information be identified?

• How much more work can the same team handle?

• How much easier is it to see where work is stalled?

Those are operational questions, but they are also the questions that turn AI investment into business value.

Put AI to Work

The businesses that win with AI will not necessarily be the ones that experiment with the most tools.

They will be the ones that figure out where AI belongs inside the way their business operates.

McKinsey describes operational excellence and AI as a reinforcing performance loop. Stronger operational practices create the foundation for AI to scale, while scaled AI can strengthen productivity, decision-making, and operational performance.

That is the opportunity ahead:

• Start with a real business problem.

• Understand the process behind it.

• Give AI the context it needs.

• Let Agents take on the repetitive digital work.

• Keep people focused on the decisions, relationships, and exceptions where they add the most value.

Because adopting AI is only the beginning.

The real advantage comes when you put it to work.

Ready to put AI to work in your business?

Vsimple builds AI Agents around the way your business actually works. From reading documents and validating information to updating systems, coordinating processes, and preparing decisions, Vsimple helps turn AI from an experiment into part of everyday execution.

The Agent does the work. Your team makes the call.

Source: McKinsey & Company, Putting AI to Work: The Operational Excellence Imperative, June 19, 2026.