← All articles

Copilot Isn't an AI Strategy. It's a Subscription.

Microsoft Copilot is useful. So is every AI feature your software vendors are adding. But your competitors can buy exactly the same licence, and a licence everyone can buy doesn't set you apart.

A glossy late-1990s style cartoon binder clip with big eyes and raised eyebrows sitting on a yellow legal pad, with an old beige monitor showing a meeting summary behind it

Key takeaways

  • Microsoft Copilot and the AI features in your CRM are genuinely useful. Use them.
  • Anything your competitors can buy with the same licence keeps you level with them. It can't put you ahead.
  • Big software vendors are adding AI to keep up, and the bill shows up in your licence: higher base prices, AI add-ons, credit packs.
  • The advantage comes from AI built on what only you have: your processes, your data and what your best people know.
  • A quick test: if your whole AI plan fits on one invoice line, it's a subscription. A strategy names a process, a number and an owner.

A version of this conversation I've had more than once: a manager at a mid-sized company tells me they're "fully AI-enabled."

I ask what they're using.

"Microsoft Copilot. It summarizes our meetings."

Every time, I have to smile. Back in 1997, a paperclip with eyebrows asked us "It looks like you're writing a letter. Would you like help?" Today the assistant is far smarter, and the question has become "It looks like you had a meeting."

I posted that exchange on LinkedIn. It was meant as a joke, but that answer is the most common AI strategy I see in small and mid-sized companies right now. And it's a subscription, which is a different thing from a strategy.

What Copilot is genuinely good at

First, to be fair: I like Copilot. Meeting summaries save time. Drafting an email in Outlook is faster. Asking Excel to explain a formula someone built in 2014 is a small daily gift. If your team lives in Microsoft 365, switching it on is a sensible move.

The same goes for the AI features in HubSpot, Salesforce and most of the other tools you already pay for. Some of them are good. A few are excellent.

The trouble starts when "we bought Copilot" becomes the answer to "what is our AI strategy?"

Is Microsoft Copilot enough for a small business?

Your competitor can buy exactly the same thing tomorrow morning. Same price, same features, same meeting summaries.

That makes Copilot a bit like electricity or email. You need it, and running without it would be a handicap. But nobody ever won a customer because they had email.

Tools everyone can buy raise the floor for everyone. If ten companies in your niche all switch on the same assistant, all ten get a little faster, and the gap between them stays exactly where it was.

There's a second, quieter limit. A general assistant knows a lot about the world and almost nothing about your business. It doesn't know which customer always pays late, why your best technician rejects a certain supplier, or which three questions a good quote has to answer in your niche. It can summarize what was said, but it can't tell which part matters.

Copilot was built for hundreds of millions of users, so being general is the whole point.

Every software vendor is bolting AI on

Look at the bigger picture and the pattern repeats. The large software companies, the ones that built their products long before today's AI models existed, are all adding AI on top of what they have.

They don't have much choice. If they don't invest, investors and customers assume they're falling behind. If they do, they spend enormous sums on AI infrastructure and still have to make those features work for every customer at once.

You can see where part of that bill ends up:

  • Microsoft raised Microsoft 365 business prices on 1 July 2026, pointing to the AI and security features it added to the base plans (heise reported on it). The full Microsoft 365 Copilot still costs extra.
  • Salesforce announced in June 2025 that list prices for its Enterprise and Unlimited editions would rise by an average of 6% from 1 August, alongside new Agentforce AI packages (Salesforce Ben).
  • At HubSpot, AI agents such as the Breeze Customer Agent and data enrichment are now metered in credits, paid on top of your seats (HubSpot investor release).

Recovering a large investment this way is normal business. Still, it's worth seeing clearly: when your vendor's AI becomes your AI strategy, you're paying for AI designed for their average customer.

And a niche champion is, by definition, not the average customer.

Where the advantage actually comes from

The companies I see getting real value from AI took one thing only they have and built AI around it. Licence count has little to do with it.

That thing is usually one of three:

  • A process nobody else runs the way you do. How you assess a damage claim, qualify a lead, price a custom order or check an incoming spec sheet.
  • Data nobody else has. Ten years of quotes, case files, test results or customer emails sitting in folders.
  • Know-how that lives in a few heads. The rules your best people apply without thinking, which nobody ever wrote down.

When AI is built on those, your competitor can't buy the same result, because they don't have your process, your history or your people. I made the same argument in why AI gives niche champions an unfair advantage, and it's why I keep saying you can use someone else's model as long as you own what the AI learns.

One example from our own work: for a truck damage assessment company we built a system around their specific claims process. The same team can now handle up to five times the volume. The value sat in their rules, which no off-the-shelf assistant knows. The full story is here.

You can have both. Keep Copilot for the email and build the edge somewhere else.

Strategy or subscription? Five questions

If you're not sure which one you have, these five questions usually settle it:

  1. Could a competitor copy our AI setup by buying the same licences? If yes, you have a subscription.
  2. Can we name the one process where AI should make the biggest difference? Not "productivity" in general. One process, with a name.
  3. Do we know what that process costs today? Hours, errors, delays, lost jobs. Without a starting number you can't tell if anything worked.
  4. Does the AI use our own data and rules, or only what's in the current document or meeting?
  5. Is someone in the company responsible for the result, not just for the licence count?

A "yes" to the first question, or a "no" to two of the others? That's normal, and it's fixable. The best first step is usually finding the one bottleneck that costs you the most and putting a number on it.

Keep the subscription. Add the strategy.

Clippy walked so Copilot could run, and I mean that kindly. The assistants got enormously better, and they're worth paying for.

But an AI plan that starts and ends with tools everyone has access to keeps you exactly where you are relative to everyone else. The advantage sits in the part nobody can sell you, which is AI shaped around your processes, your data and your people.

You have to build that part yourself.

Frequently asked questions

Often, yes. If your team works in Microsoft 365 every day, the time saved on meeting notes, emails and documents can justify the cost. Just treat it as a productivity tool for everyone, not as your AI strategy. It gives you the same benefits your competitors get.

Copilot is a general assistant built for millions of users. It works with whatever document, email or meeting is in front of it. Custom AI is built around one of your own processes and uses your own data and rules, for example how you assess claims, price orders or qualify leads. That's where results competitors can't copy come from.

No. Most companies we work with keep their standard office and CRM tools and add a custom layer for the one or two processes that matter most. The two run side by side.

Large vendors are investing heavily in AI and adding it to existing products. Part of that cost reaches customers through higher base prices, paid AI add-ons or usage-based credits. Microsoft, Salesforce and HubSpot have all moved in this direction since 2025.

Start with the process, not the tool. Find the one bottleneck that costs you the most, measure what it costs today, and test one fix on it. Then decide which tools you need.

Want to see what AI can do for you?

Tell us about your business. We'll get back to you within 24 hours.

Schedule a Strategy Call

More articles