By now, AI has worked its way into nearly every business in some form. McKinsey’s 2025 Global Survey on AI found 88% of organisations use it in at least one part of the business, up from 78% the year before.
What the headline number leaves out: roughly two-thirds haven’t gone past piloting it.
Using AI and scaling it are two different projects. AI-enabled business scaling means building what worked in a small test into how the business actually runs, not leaving it sitting on the side as a tool nobody’s redesigned anything around.
It’s a meaningful gap.
Only 39% of businesses report any earnings impact from AI at all, and most of that group puts the figure under 5%. Most of the value is still sitting untapped on the other side of that gap.
Table of Contents
What Counts as AI-Enabled Business Scaling
A chatbot answering customer emails is a small test.
Redesigning the whole support workflow around that chatbot, retraining the team, and changing how performance gets measured is scaling.
EY’s research calls this the difference between “bolt-on” AI and “built-in” AI. Bolt-on means applying AI to a process that stays exactly as it was. Built-in means the process itself gets rebuilt to fit what the technology is now capable of.
Bolt-on delivers a modest efficiency bump and not much else. The real payoff sits with built-in AI, and that’s precisely where most businesses give up.
Swapping in a new subscription is quick. Rebuilding the process around it takes real time, which is exactly why so few see it through.
Money is moving toward this shift regardless of whether individual businesses are ready for it. Gartner projects worldwide AI spending will reach US$2.5 trillion in 2026.
At that scale, sitting stuck between using AI and scaling it isn’t a neutral place to be. It’s an expensive one.
What You Really Get With AI-Enabled Business Scaling
AI at scale doesn’t pay off all at once. The return builds across three distinct stages, and the gap between them is bigger than most businesses expect.
Stage 1: Bolting AI Onto Existing Work
Deploying AI tools into existing workflows, the simplest stage, delivers workforce productivity gains of 10 to 15%. Useful, but not the number that changes a business.
Stage 2: Scaling Into the Functions That Actually Matter
Businesses reshaping core functions around AI, not just the support functions sitting at the edges, pull 72% of their total AI-driven value from that core work alone.
Stage 3: Where AI-Enabled Business Scaling Has the Biggest Impact
Only 46% of AI-mature companies have gone on to invent new offers or business models around what AI now makes possible, and that’s where the biggest gains tend to sit.
Why AI-Enabled Business Scaling Plans Come Apart
AWS’s research on what scaling AI actually requires lands on three things: people, technology, and process. Most of the friction shows up in the first and third, not the second. Getting the technology working is rarely what stalls a business.
Three Ways a Scaling Plan Falls Over
What stalls it usually comes down to a few things:
- A model gets handed from one team to another with no documentation.
- A pilot never had input from the people who’d actually use it day to day.
- Or there’s no standard process for retraining and monitoring once the model goes live and the data underneath it starts to shift.
What AI Scaling High Performers Do Differently
McKinsey’s data backs this up from a different angle entirely. Roughly 6% of respondents fall into what McKinsey calls AI high performers.
These businesses are three times more likely than everyone else to have genuinely redesigned their workflows around AI, rather than dropping it unchanged into the old ones.
Senior leaders at these businesses stay closely involved too: high performers are three times more likely to have leadership actively engaged in the programme, rather than checking in now and then while one team runs the whole thing.
They also tend to back it with real money. More than a third of high performers put over 20% of their digital spend behind AI, and about three-quarters of them are already scaling it, compared with roughly a third of other businesses.
The Main Barrier to AI-Enabled Business Scaling
There’s a cultural layer underneath all of this too. AWS’s research points to resistance from within the business itself as one of the recurring blockers.
- Teams often don’t fully understand what the AI can and can’t do.
- Workflows that were never built with AI in mind resist being changed.
- And there’s the everyday friction of asking people to work differently after years of doing something one way.
None of that shows up on a project plan, which is exactly why it’s so often the thing that quietly stalls a scaling programme that looked fine on paper.
The tools matter less than which team gets to redesign the workflow, and whether leadership actually shows up for it.
The Major Issue Behind AI-Enabled Business Scaling
The effort behind a successful scaling programme doesn’t go where most people expect. It roughly breaks down like this:
- About 70% into people and process
- 20% into technology and data
- Only 10% into the algorithms themselves
Most businesses assume AI scaling is mainly a technical exercise, so a split like that tends to catch them off guard.
Who’s Missing From Your AI Scaling Team
The people side is also where the current bottleneck sits. In McKinsey’s survey, software engineers and data engineers are the roles businesses report as most in demand for their AI programmes.
Local hiring doesn’t make it any easier either, given how thin the market already runs on AI-specific skills.
Businesses weighing whether to build that capacity in-house or bring in outside engineering support are really deciding how fast they can move from a small test to a properly scaled programme.
A hybrid setup that pairs in-house strategy with specialist AI delivery capacity is how plenty of businesses close that gap without waiting out a slow local hiring market for roles that are scarce everywhere right now.
Scaling AI Step by Step
None of this needs to happen in one leap either. Forvis Mazars’ work with businesses going through digital transformation points to an incremental approach as the more reliable path:
- Pick a piece of the workflow.
- Get a measurable return on that piece.
- Then move to the next one, rather than attempting a full rebuild in one go.
Businesses that try to redesign everything at once tend to lose momentum halfway through. The ones that pick a scope they can actually resource properly, staff and all, are the ones still running the programme twelve months later.
Making AI-Enabled Business Scaling Happen
The businesses actually scaling AI aren’t the ones with the most advanced tools. They treated it as a redesign of how work actually gets done.
They put the right people around it. And leadership stayed genuinely involved, instead of handing the whole thing to one team and walking away.
That’s a harder project than switching on a new subscription, but it’s also the only version that shows up as real return once the numbers come in.
Contact Outsourced Staff if the people side of that equation, not the tooling, is what’s holding your AI programme at the testing stage.
FAQs
What does AI-enabled business scaling actually mean?
AI-enabled business scaling is about weaving AI into how the business actually operates, rather than parking a tool beside the process that was already there. A pilot that stays a pilot, however good the results, isn’t scaling.
Scaling means the workflow, the team, and how performance gets measured all change around what the technology can now do.
How long does AI-enabled business scaling actually take?
There’s no fixed timeline, and businesses that expect one tend to get frustrated early. What consistently speeds it up is redesigning the workflow rather than leaving it as is, and having leadership genuinely engaged rather than leaving the programme to one team.
What roles does a business need in place before scaling AI?
Software engineers and data engineers are consistently the hardest roles to fill and the most in demand for AI programmes, according to McKinsey’s research. Beyond the technical roles, someone senior enough to own the redesign of the actual workflow matters just as much.
Does AI-enabled business scaling mean cutting staff?
Scaling AI doesn’t necessarily mean cutting staff. McKinsey’s survey found expectations vary widely: some businesses expect workforce reductions from AI, others expect growth, and a large share expect little change at all.
What’s consistent among high performers is a shift in what people spend their time on, not simply fewer people doing it.