First you will fire the overpriced experts. Then you will discover that the most expensive person is the one you no longer employ. By then, your technology provider will be revising its prices.
Imagine a director who has just discovered artificial intelligence. Yesterday, getting an expert opinion required a colleague. Today, it requires a password. The progress is obvious: a password does not argue, ask for a raise or begin every other answer with that irritating phrase, “it depends”.
Finally, someone who understands the vision. Or can at least turn it into twelve convincing bullet points.
After a few successful letters and one beautifully formatted analysis, enlightenment arrives: why, exactly, are we paying all these people?
A reasonable question. Especially if you have always measured an expert’s work by the number of pages they produce. The tool produces more. Faster. And without complaining that the input data makes no sense.
You gained access to an answer. Somewhere along the way, you also credited yourself with the expertise needed to judge it.
It is a wonderfully comfortable substitution. You can discuss a subject in a meeting that you knew nothing about yesterday. You use the right terminology. The arguments have structure. Even the conclusion sounds confident. If someone asks which assumption holds the whole thing together, you can always open another chat.
The colleague who points out a mistake becomes the problem. Not agile enough. Does not understand the future. Resistant to change. How convenient: the person capable of checking your new source of wisdom can be declared an obstacle to implementing it.
Naturally, there must be a next step. A chat assistant alone does not constitute digital transformation. You need agents, integrations and automation. A contractor builds the solution. Sales gets summaries, support gets answers, management gets a dashboard. Everything connects to everything. The presentation is excellent.
Then comes everyone’s favourite slide: staffing savings.
Out goes the person who knows why the largest customer never follows the standard ordering process. Out goes the colleague who can spot bad data before it becomes a bad decision. Out goes the engineer who understands what actually happens between the systems.
In the spreadsheet, these are three successfully eliminated cost items. In the business, they are three holes whose depth you have yet to discover.
But the reports keep coming. The indicators are green. The system works.
A bridge can remain standing for a while after you remove the engineer. That does not prove you no longer need one.
Beneath the polished dashboard lives a collection of models, connectors, permissions, external services and rules that somebody needs to understand. When everything goes to plan, that person looks like an unnecessary expense. When something goes wrong, you start asking, “Do we really have nobody who knows how to fix this?”
You did.
You presented their departure as an achievement.
Now suppose there is no catastrophe. No breach. No dramatic outage. Just an email from your provider: we are updating our plans, improving the user experience and adjusting our prices.
Your user experience is about to become considerably more expensive.
Of course, you will threaten to leave. You own a business. You know how to negotiate. Then someone will ask how long migration would take, which data you can actually move and who knows how to rebuild the processes.
Suddenly, there will be a lot less visionary thinking in the room.
The provider does not need to stop you leaving. It only needs leaving to become unaffordable.
This is where the story gets interesting. The owner who wanted less dependence on people has become more dependent on a platform. An employee’s bargaining power was an irritation. Now that power belongs to someone who does not even need to know your name.
You are a customer number. Your business continuity is a feature in their subscription plan.
You might call it technofeudalism. The quotation will probably call it “Enterprise”.
For this to happen, nobody needs to destroy open source or acquire every small developer. Alternatives can remain freely available. Without people who know how to use them, they are about as useful to you as aircraft blueprints to a passenger who has just fired the pilot.
Let us be clear: a business is not a museum of jobs. Pointless tasks do not need preserving, and poor organisation does not need defending with romantic appeals to the human touch. An expert who keeps all their knowledge in their own head creates dependency too. AI can help document processes, spread knowledge and get work done better.
Which makes it a shame if management’s biggest idea for such a powerful tool is simply: how many people can we remove?
I would expect a good director to ask a few more questions. Who will check the results? What knowledge must remain ours? What would it cost to leave? Can we keep operating when our chosen provider changes the rules?
These questions are less glamorous. They do not make for quite as pleasing a slide as lower staffing costs. But they matter a great deal to anyone who considers themselves the owner of their business.
Perhaps AI really will make you more profitable. Perhaps a smaller team will do better work. But if you also remove your ability to understand, verify and make independent decisions, you have sold part of your future bargaining power for a better result today.
The difference will become apparent at the next price increase.
Fire the expertise. Keep the subscription. Someone else will set its next price.
Then you can explain to them that they are too expensive.
Just as you once explained to your employees.