Count the software tools your business uses to run its day.

Most mid-sized businesses are somewhere between eight and fifteen. A CRM. An accounting platform. A project management tool. A marketing automation system. An analytics dashboard. Spreadsheets — always spreadsheets. Communication tools. Document storage. Scheduling. Inventory. The list goes on.

Now count how many of those tools talk to each other without a human in the middle.

For most businesses, that number is much smaller. And the gap between those two numbers represents something important — a category of work that almost no one accounts for in their planning, their hiring, or their cost modelling.

It’s the cognitive work between platforms. The invisible layer.

It looks like this. Your sales team closes a deal. Someone checks the CRM, then updates a spreadsheet, then sends an email to operations, then creates a job in the project management system, then attaches the relevant documents. The information exists — it was captured in the CRM when the deal was logged. But it takes a person thirty minutes to move it through the chain. That person is not doing strategic work. They're being a data courier.

Or this. Your marketing team publishes content. Someone manually checks the analytics after a few days, copies the numbers into a report, reformats them into something the leadership team can read, and distributes it. The raw data was sitting in the analytics platform the whole time. But extracting it, synthesising it, and making it legible required a human's time and attention.

Every business has dozens of these workflows. The glue work. The translation work. The "I need to check X before I can update Y" work. The formatting, the copying, the reformatting, the filing.

This work is invisible in the sense that it doesn't appear on any job description. Nobody was hired to be the glue between your platforms. But a substantial portion of your team's cognitive bandwidth is absorbed by it — estimates I've seen put it at 20 to 40 percent of knowledge worker time, depending on how integrated the tech stack is.

Here is why this matters right now.

For the last decade, the answer to this problem was integration. Build an API connection. Use a middleware tool. Get your CRM talking to your accounting software. These solutions helped, but they required significant technical effort to implement and maintain, and they only worked for structured, predictable data flows.

The messier, more judgement-heavy connections — the ones that required a human to look at context, make a decision, and take an action — remained manual. Because software could automate the movement of data, but it couldn't automate the thinking required to know what to do with it.

That boundary is shifting.

AI — specifically the kind that can read context, apply reasoning, and take conditional actions — is capable of handling the cognitive work that previously required human presence. Not just moving data, but interpreting it. Not just triggering a workflow, but deciding which workflow to trigger based on what the data actually means.

This is not a theoretical capability. It is available today, in production, in businesses that have made the investment to deploy it properly.

What changes when you remove the cognitive glue work from the human workload?

The first change is capacity. People who were spending a third of their time on coordination and translation tasks suddenly have that time back. They can do more of the strategic work — the work that actually requires human judgement, creativity, and relationship management.

The second change is speed. Decisions that previously waited for a person to notice, process, and act on information can now happen continuously. Your CRM updates the moment a deal closes, because the system sees it and acts. Your reporting refreshes daily, because no human has to trigger the process.

The third change — and this is the one that matters most for business structure — is that the case for certain roles changes. Not because those people are being replaced, but because the composition of what they do changes. Less glue work, more judgement work. Which is a better use of capable people, and usually a more satisfying job to do.

The platforms in your business are probably not the bottleneck. The invisible work happening between them almost certainly is.

That's the work AI is about to claim.