AI moves $4.7 trillion of profit. The question is your sector.

Bain & Company published research on 8 September 2026 putting $4.7 trillion of profit in play between 2025 and 2035. Bain describes the effect on corporate profit pools as more than triple the internet's, in half the time, and says AI will structurally touch 71% of sectors against 41% for the internet (Bain & Company, 2026).

Two things in that are worth your attention, and neither is the headline number.

The list of affected sectors got longer

The internet reshaped retail, media, travel and financial services. It largely left healthcare, manufacturing and professional services alone. Those three are inside the 71% this time.

That matters for Ireland, because that is where a large share of mid-sized Irish firms actually sit. Engineering and manufacturing, clinics and care providers, accountants, surveyors, solicitors, recruiters, consultants. Businesses that came through the last twenty years with their model intact are in scope now.

Productivity is the small part

Bain splits the $4.7 trillion three ways: productivity, innovation, and shifts in market share. Productivity is the piece everyone talks about, and Bain's own framing is that most of the opportunity and most of the disruption sits beyond it.

That is the uncomfortable bit. Doing the same work with fewer hours is the easy story to tell a board. The larger number is in what your competitors start offering, and in the customers who move because of it. You cannot get to that number by trimming admin time.

The figure is also about a fifth of projected global profits in 2035. Not a rounding error, and not evenly spread.

Four groups, not one trend

Bain clusters industries into four patterns rather than one:

• Technology foundation. The firms selling the compute, models and tooling everything else runs on.
• Rewired. Sectors where the work itself gets rebuilt around AI rather than assisted by it.
• Augmentation. Sectors where people stay central and get faster, with judgement still human.
• Revolution. Sectors where the product, the pricing or the customer relationship changes shape.

Most Irish SMEs are in augmentation, and quietly hoping that is the whole answer. Some are in revolution and have not noticed, usually because the change is arriving through a client's procurement process rather than through a competitor's launch.

The practical value of the four groups is that they are four different questions. An augmentation business asks where its expensive people lose hours. A revolution business asks what a client will expect to buy in three years. Answering the wrong one is how a firm ends up with a chatbot and no advantage.

What conviction actually looks like

Bain's line is that the more conviction you have about your own sector, the faster you can build a lead. Conviction is the word to sit with, because conviction is not a slide.

In practice, in a business of 10 to 250 people, it looks like this:

1. Name the two decisions that make you money. Quoting accurately. Deciding what to take on. Scheduling the right person. Whatever it is in your business, it is rarely more than two or three.
2. Find where those decisions wait. They almost always wait on information somebody has to go and find, or on one person's availability.
3. Build one narrow thing against one of them. Not a platform. One step, using the systems you already pay for, with a named person reviewing the output before anything moves.
4. Read your own numbers. After a month you know whether your sector rewards this and where. That is conviction, and it cost you weeks rather than a year.

The firms that will take share are not the ones with the best AI strategy document. They are the ones who learned something specific about their own economics early, and then kept going.

Why we take this approach

At Artellis I build one workflow at a time for Irish and UK companies, and the reason is the argument above. A single working step tells you more about your own position than any amount of forecasting, and it leaves you with something in use rather than something in a folder. Where that fits, how we work and the one workflow approach set out what a first build involves. If you are still deciding what counts as a workflow at all, start here.

Bain's number is not a prediction about your company. It is a statement that the profit in your sector is going to move. Whether it moves toward you depends on how quickly you find out where.

If you want to work out which of the four groups you are in, and what a first build would be, twenty minutes is enough to get to a straight answer.