INSIGHTSSTATEMENT

A tool only works if it fits what your team already does

People are rebuilding their TVs to do what they want. The same holds at work: a tool that fits what your team already uses is a tool that gets used.

By

Monday, 8.30
RotaQuotesOrders
What do I need to know today?
Two appointments, one quote waiting for approval and your order is on its way.Rota · Quotes · Orders

THE PROBLEM

Yet another new system to learn

A new tool is meant to make the work easier. Yet you often notice the opposite. Your team gets one more portal, with its own login and its own way of working. After a few weeks everyone goes back to what they were used to.Meanwhile the information is still spread out. A colleague opens the planning first, then the customer system and then a supplier's portal. That happens again every day, by hand.

WHAT IS BEING SAID

People are rebuilding their TV around how they watch

Something striking is happening with smart TVs this month. Many owners find the home screen too full, with ads, sponsored rows and apps they never open. They do not buy a new TV. They ask an AI assistant to clean up the TV they have.On 17 October a developer showed a TV that Claude Code had cleaned up over Wi-Fi. According to the project page of CleanYourTV, that post was viewed 7.4 million times. Within four days more than 200 people had sent in details of more than 170 TV models.The AI assistant switches off the ads, the sending of usage data and the preinstalled apps. Nothing is removed, so everything can be put back. Then a calm home screen with only the owner's own apps takes its place. Before the assistant switches anything off, it asks which apps the owner uses.A similar project for Fire TV, dumbtv, puts it at the top of its page: your TV should do what you paid for. The apps and the HDMI inputs stay. Only what sat in between goes.This is not research. These are the makers' own pages, with reports from users. The makers name the limits too. On a Fire TV, updates are blocked while the filter runs. Changing the wrong part of a TV can make it unusable. And according to dumbtv, Samsung and LG give owners no such access.

At work, AI is moving the same way: to where people already are. In November 2024 Anthropic introduced MCP. It is an open standard that lets an AI assistant connect to the systems where the information is kept. Anthropic's aim: one way of connecting, instead of a separate connection for every source.In July 2025 came a directory of connections for Claude, with tools such as Notion, Canva and Stripe. Anthropic puts the difference like this: without a connection you explain everything to the AI each time, with one the AI looks it up itself. Connections to online services are only available on the paid plans there.

AJEKO'S VIEW

A tool has to fit what people already do

The TV owners show what holds at work too. Nobody asked for a new TV or a new screen to learn. They kept their own apps and took away what stood between them and their goal. And that goal was simply watching TV.Tools and automation in a business work the same way. A new portal asks your team to work differently. If that does not fit how they work, it is not used in practice. An improvement that fits what your team already uses is something nobody has to learn. That is why it gets used.Many colleagues already have an AI assistant open all day, such as ChatGPT, Claude or Copilot. If you connect the portals that are now checked by hand to it, a colleague asks the question where they already work. The portals still exist. Nobody has to go past them any more.

What does your team have to learn?A new portalA new login and a new screenIn what your team already usesNothing new
Where does a colleague look something up?A new portalIn each system separatelyIn what your team already usesIn the chat that is already open
What happens after a few weeks?A new portalEveryone goes back to the old wayIn what your team already usesIt is part of the everyday work

HOW TO APPROACH IT

Start from what your team already uses

That fits the principle Your way. The goal is to work smarter, and how you get there fits your business and the systems you already have.

  1. 1

    Look at where your team works now

    Which apps are open all day? Think of e-mail, WhatsApp or an AI assistant. That is where the improvement belongs.

  2. 2

    Find what is checked by hand

    Which portals does a colleague open every day to look something up or retype it? That work can move to the place from the first step.

  3. 3

    Bring the information to that place

    Connect those systems to your team's AI assistant, for example through MCP. That way a colleague gets the answer in the chat that is already open.

  4. 4

    Let your team decide what stays

    As with the TV, the people who work with it choose what they use. What nobody misses can go.

What that looks like in practice is shown by the AI WhatsApp summaries. Recruiters at a staffing agency keep in touch with candidates on WhatsApp, on their own phones. The usual solution was a shared inbox, but then recruiters have to move to a different program. That is why the system was connected to the number they already had. Recruiters keep the same phone, the same app and the same number. AI saves a summary of every conversation in the candidate's file.The AI phone call summaries work the same way. After every phone call the summary goes into the file and by e-mail to the colleague who called. Nobody has to do anything for it.

This approach does not always fit. Sometimes what your team is used to simply no longer works. With the margin dashboard, the margin was in a spreadsheet, and a colleague only saw their own figures once someone had worked them out. There a separate dashboard was the better choice. Even then it starts from how the work runs today.And as with the TV, a connection needs maintenance. If a portal changes, the connection has to change with it. You also decide what the AI can and cannot access.

SOURCES

Curious what this means for your business?