AI Credit: StockCake TL;DR88% of organizations use AI but only 7% have scaled it (McKinsey). Dessy Pavlova argues the gap exists because companies apply AI to fragmented operations rather than redesigning the process first. She builds systems where a single human decision triggers cascading updates across websites, scheduling, and finance, with stopgates for human oversight. NBER research found AI boosted support-agent productivity 14% on average, but the technology helped people perform work, not eliminate the need to understand it.
I have come to believe that one of the biggest misconceptions about AI is also one of the most expensive: the idea that technology can repair a business that has never streamlined how its processes actually work. AI can accelerate an operation, automate repetitive work, and surface problems at remarkable speed. Give it a fragmented business, however, and you may simply get a faster version of the fragmentation. The numbers should make business leaders pause.
McKinsey’s 2025 global survey found that 88% of respondents said their organizations were using AI in at least one business function, yet only 7% said AI had been fully scaled across the organization. Its research points to workflow redesign as one of the strongest factors associated with financial impact from generative AI. Why is the gap so large? Look at how many businesses actually operate.
A customer arrives through a website. Someone enters the information into a spreadsheet. Another person moves it into an operating platform. Someone updates the website.
Finance receives a separate piece of information. An accountant reconciles something later. Every handoff creates another opportunity for delay, duplication, or error. The business may have excellent software at every stage and still have a broken system.
I see this often because my work now sits deep in business operations, although my background is in writing and marketing. I have worked with businesses ranging from one-person organizations to multimillion-dollar operations, and the pattern is remarkably familiar. Companies accumulate tools because each one solves a particular problem. Then they use connectors such as Zapier or Make to force those tools to cooperate.
There is another possibility now. Build the system around the business itself. Imagine a company running dozens of classes. A student changes classes.
A teacher moves to another day. A human makes that decision. The information then needs to travel through the website, operating system, scheduling process, and financial records. In a fragmented setup, people may have to make every downstream change themselves.
In a properly designed AI-enabled system, the original decision can trigger the rest of the workflow automatically. The human remains the architect. AI becomes the operational engine. That is crucial because automation without oversight simply relocates risk.
I build stopgates into systems so a person can see what changed, intervene if something looks wrong, and confirm that the process has reached the intended endpoint. AI can flag an anomaly or carry information across systems. Someone still needs to understand what the business is trying to accomplish. Research supports the value of this human-machine arrangement.
A study of 5,179 customer-support agents found that access to generative AI increased productivity by 14% on average, with particularly large gains among less experienced workers. The technology helped people perform their work. It did not eliminate the need for people to understand the work. This becomes even more important as a company grows.
Growth exposes every loose connection. Before spending heavily on marketing or customer acquisition, I want a founder to map the entire journey of a customer. How does someone enter the business? Where does their information go?
Who acts on it? What reaches the customer? What reaches finance? How does the accountant know what has happened?
Where can a human error enter the chain? Getting the customer and collecting the money are only two moments in a much longer operational journey. AI gives us an extraordinary opportunity to redesign that journey. McKinsey’s latest research argues that applying AI to isolated tasks leaves substantial value on the table, while reimagined workflows can unlock much larger productivity gains.
IBM’s 2026 research also found that only 11% of surveyed technology leaders felt completely prepared for the scale of AI-agent deployment, highlighting how quickly capability can outpace organizational control. The truly revolutionary AI application for business now has nothing to do with generating another piece of content. The real power lies in the invisible architecture underneath it all, ensuring information moves flawlessly, decisions reach the right leaders instantly, financial records remain immaculate, and employees finally win back their time. I don’t want AI running a business for people.
I want it to remove the administrative friction that prevents people from running their businesses well. Build the blueprint carefully. Put humans at the right decision points. Let AI handle the legwork.
Then growth has somewhere solid to go.
