The Coordination Tax | Eliran Keren

Seven AI tools. Focus efficiency at 60%.

Pick one: the tools are failing, or something else is happening.

Here’s what ActivTrak found after tracking 10,584 workers across 180 days before and after AI adoption: time spent on every single job responsibility went up. Email: +104%. Messaging: +145%. Business management tools: +94%. No single activity category showed time savings. Not one.

Deep focus sessions fell 9%. The share of work time spent in uninterrupted concentration — the thing AI was supposed to protect — dropped to 60%. A three-year low.

Organizations aren’t failing to use AI. They’re using it everywhere. The problem isn’t the tools.

What the productivity report can’t explain

There’s a number buried in the Boston Consulting Group 2026 research on AI brain fry. Workers using three or fewer AI tools report efficiency gains. Workers using four or more tools show productivity decline.

Read that again: more tools means less output.

That’s the opposite of what the productivity story predicts. You install more capability, you get less work done. Not because the tools don’t function. Because something is happening in the space between them.

The average organization now runs seven AI platforms. Up from two in 2023. Each one is useful. Each one has its own interface, its own context, its own demand on your attention.

None of them know what the others are doing.

The person in the middle

When your Slack AI processes a thread, your email AI drafts a reply, your project AI logs the task, and your meeting AI captures the action item — four systems just touched one event. Each one processed a fragment. None of them handed off.

You did.

You are the coordination layer. Not by choice. By necessity. Every context switch between tools — checking what one said to inform another, re-explaining the situation to a system that has no idea what came before — is coordination work. Hidden work. Work that doesn’t appear in any productivity metric.

But it shows up as a 13-minute average focus session.

It takes 23 minutes and 15 seconds to refocus after an interruption. When your day involves seven tools, each demanding entry, each requiring translation into the next — you never actually refocus. You accumulate what I’d call a coordination tax: the cognitive cost of being the middleware in your own system.

The wrong prescription

When this data surfaces in a leadership meeting, the instinct is to cut the tool count. Run an audit. Train people to use fewer tools better.

That’s solving for the symptom.

The problem isn’t that you have too many tools. The problem is that no layer exists above them. Nothing that knows you committed to something last Wednesday. Nothing that knows the Slack thread, the project board, the email chain, and the meeting notes are all about the same thing. Nothing that can tell you: here’s what actually matters today, and here’s why.

The human fills that gap. By default. At the cost of 9% of their deep focus time. At the cost of doubled email volume. At the cost of a 23-minute refocus window they never actually get back.

What the data is actually describing

This isn’t a productivity paradox. It’s a coordination architecture failure.

Agentic AI will not make this better on its own. 40% of enterprise applications will embed task-specific agents by end of 2026. More capable tools, deployed into the same uncoordinated stack, produce the same result: more cognitive load on the person managing the system, not less.

The fix isn’t fewer tools. The fix is a coordination layer above the tools — something that takes the routing work off the human, tracks what’s active, surfaces what matters, and closes loops the human would otherwise carry in their head.

That layer doesn’t exist at scale yet. The personal stack is still missing its infrastructure.

The productivity data has been trying to say that for three years.

The question is whether we name it right.

Eliran Keren — Founder of Deeplica, building the coordination layer for the person at the center.

Sources: ActivTrak — AI Productivity Report 2026 · Workplace AI: Focus Time Hit a Three-Year Low · ‘AI Brain Fry’ Is Real — BCG 2026 · AI Workplace Productivity Digest, April 10