AI Adoption in Business in 2026
Where Companies Really Are and What to Do About It
The Hype vs The Reality
If you've been reading the headlines, you'd think every business in the country has an AI-powered sales team, a fleet of automated assistants and a custom-built intelligence system running the show.
The reality is very different. A massive adoption lag - I see it a lot myself in business I speak to and the survey I carried out at the end of the year echoes this.
Almost half of all workers globally haven't used AI at all. Not once. In the UK, only around a quarter of businesses are using it in any form. And among those that are, most are doing little more than asking it to check their emails or rewrite a paragraph.
Meanwhile, the noise keeps getting louder. Agentic AI. Autonomous workflows. Custom knowledge hubs. Every conference, every LinkedIn post, every vendor pitch makes it sound like you're already five years behind.
You're probably not. And that's exactly why this matters.
This article gives you a clear, research-backed picture of where businesses actually are with AI in 2026 - and a practical framework for moving forward without the overwhelm. No jargon. No hype. Just what you need to know as a business leader to make sensible decisions about AI.
Quick Links
- The real numbers behind AI adoption
- What people are actually using AI for
- Marketing is leading the way
- Why people aren't adopting - the real barriers
- The hidden problem - people aren't talking about it
- A practical path forward - audit, strategy, training
- Don't get distracted by the shiny stuff
- What this means for your business
- How to future-proof your organisation
The Real Numbers Behind AI Adoption
Let's start with the data, because it paints a very different picture from the one most people assume.
Gallup surveyed over 20,000 workers in late 2025 and found that 49% had never used AI at work. Just 12% were daily users - up only slightly from 10% the year before. A further 26% used it a few times a week. That's progress, but it's hardly the revolution the headlines suggest Gallup Workplace AI Study.
The UK picture is even more sobering. The Office for National Statistics found that just 25% of UK businesses were using AI by late 2025. Among larger companies with 250 or more employees, that figure rose to 44% - but for SMEs, adoption remains significantly lower ONS Business Insights Survey.
At a global level, the numbers look more encouraging on the surface. McKinsey found that 88% of organisations say they use AI in some form. But only 1% describe their rollout as "mature" and just 6% qualify as high performers seeing meaningful financial returns McKinsey Global Survey: The State of AI in 2025.
We ran our own Sales, Marketing and AI Survey at the Advanced Engineering Show in late 2025 and found a similar pattern. 67% of engineering organisations were either not using AI at all or engaged in what we'd call unstructured experimentation - trying things out with no plan, no measurement and no coordination. Over 70% were not using AI regularly in any meaningful way. Nobody - not a single respondent - identified as a power user with advanced skills.
The gap between what people think is happening and what's actually happening is enormous.
FAQ
How does UK AI adoption compare to the rest of the world?
The UK sits slightly below global averages. Around 25% of UK businesses currently use AI, compared with roughly 70% globally who say they use it in some form. However, the global figure is misleading - the vast majority of those businesses are at a very early or experimental stage. The real gap isn't between countries. It's between businesses that have taken structured steps and those that haven't started at all.
What People Are Actually Using AI For
Here's the part that often surprises people. Even among those who are using AI, the level of sophistication is remarkably low.
In our engineering survey, 94% of AI users were using it to generate text - help me write this email, proof this document, draft a LinkedIn post. Half were using it for some form of marketing intelligence and 37% for basic competitive analysis. But only 14% had ever run a deep research project - the kind of structured, multi-source analysis that AI platforms like ChatGPT, Gemini and Claude are now very capable of.
This lines up with what OpenAI found when they analysed how people actually use their platform. Roughly three-quarters of all conversations are basic guidance, information-seeking and writing tasks. The dominant behaviour is still "ask a question, get an answer" rather than using AI as an integrated part of a workflow OpenAI: The State of Enterprise AI 2025 Report.
Microsoft's 2025 Work Trend Index put a number on this. They found that 47% of employees still treat AI as a command-based tool - type a question, get a response, move on. The other half are starting to use it as more of a thinking partner, and that second group is seeing significantly better results. The split is almost exactly down the middle, but the productivity difference between the two groups is substantial Microsoft 2025 Work Trend Index. Using it as a co-worker and where prompting is a conversation rather than a highly polished prompt or two is a massive unlock for how to use and get the most of AI.
There's an interesting pattern in who uses AI and how much. Gallup found that leaders and senior managers use AI nearly three times more frequently than individual contributors - 44% compared with 23%. This makes sense when you think about it. Leadership roles naturally involve more of the tasks AI is good at - strategic thinking, synthesising information, drafting communications, delegation and planning. But it also means the people on the ground, doing the day-to-day work where AI could save the most time, are the ones using it least.
The sophistication gap is as wide as the adoption gap. Most people aren't even close to using AI in the ways it's designed to be used.
FAQ
Is AI just for writing content?
No. Text generation is where most people start, but AI platforms are now capable of deep research, data analysis, competitive intelligence, strategic planning support, image creation, workflow automation and much more. The reason most people stick to writing tasks is a lack of training and awareness - not a limitation of the technology. Even within text-based tasks, there's a massive difference between asking AI to "rewrite this paragraph" and using it to build a structured brief, draft a proposal and then critically review its own output.
Marketing Is Leading the Way
One area worth calling out is marketing, which consistently shows higher adoption rates than almost any other business function.
Research from the Social Media Examiner found that 60% of marketers now use AI every day - a significant jump from 37% in 2024. 84% increased their usage over the past year and 82% plan to increase it further Social Media Examiner AI Report 2025. This makes marketing one of the fastest-moving functions when it comes to AI adoption.
Our own survey backed this up. In engineering companies, 66% had used AI to help with marketing strategy - though primarily for text generation rather than deep research or competitive analysis.
This matters for two reasons. First, if you're a business leader looking for the best place to start with AI, sales and marketing is one of the most natural entry points. The tasks are well-suited to it - research, content creation, competitive analysis, customer profiling, messaging. Second, marketing adoption tends to lead broader business adoption. When teams see real results in one function, it builds confidence to expand into others.
If your marketing team isn't using AI yet, it's worth asking why - because your competitors' marketing teams almost certainly are. See our
AI for sales and marketing services for how we help engineering and technical companies get started. Taking the lessons from marketing and applying this to the rest of the company can be transformational. If you work in front of a screen - the applications are numerous and can come in so many different forms.
FAQ
Why is marketing adopting AI faster than other functions?
Marketing tasks are particularly well-suited to AI's current strengths - generating written content, analysing competitors, researching markets, creating imagery and building messaging frameworks. The outputs are also easy to evaluate, which means people can quickly see whether AI is helping or not. This creates a positive feedback loop where success breeds confidence, which breeds more usage. It's a natural starting point for most businesses.
A Practical Path Forward - Audit, Strategy, Training
Here's the good news. You don't need to understand agentic AI. You don't need custom data pipelines or automated systems. You don't need to hire an AI specialist or invest in expensive enterprise platforms.
What you need is a structured approach to three foundational stages that most businesses haven't done yet. Get these right and you'll see meaningful productivity gains and workflow improvements - without the complexity.
Stage 1 - Audit Where You Are
Before you do anything else, take stock. You need an honest picture of where your organisation currently sits with AI.
This means understanding who is using it and who isn't, what they're using it for, how confident they feel, what tools they have access to, and what barriers are getting in the way. It also means looking at your business workflows and identifying where AI could genuinely add value - not where it sounds impressive, but where it would make a practical difference to how your team works day-to-day.
An audit doesn't need to be complicated. It can be a structured survey, a series of conversations or a facilitated workshop. The point is to move from assumptions to facts. Most leaders are surprised by what they find - both by how little is happening in some areas and how much quiet experimentation is happening in others.
Companies with a formal marketing strategy in place were 40% more likely to have piloting or operational use of AI within their business, according to our survey. Strategy creates the structure and clarity needed to identify where AI adds value. Without it, you're guessing.
Stage 2 - Build an AI Strategy
An audit tells you where you are. A strategy tells you where you're going and how to get there.
This doesn't mean a 50-page document. It means answering a few practical questions: what are the three to five workflows where AI could make the biggest impact? What tools and platforms do we need? What are the rules - what's acceptable to use AI for and what isn't? Who is responsible for driving this forward? How will we measure whether it's working?
The companies that are seeing real results aren't the ones with the most advanced AI tools. They're the ones with a clear plan for how to use them. BCG found that just 5% of companies have achieved genuine AI maturity, while 46% remain at an early experimental stage - but the mature companies are seeing 1.7 times greater revenue growth BCG: The Widening AI Value Gap.
The difference isn't the technology. It's the strategy behind it.
Stage 3 - Personalised, Role-Specific Training
This is where most companies fall down completely. They give people access to AI tools - maybe a ChatGPT licence, maybe a Copilot subscription - and then expect them to figure it out on their own. Some run a generic webinar or share a few links to articles. Then they wonder why nothing changes.
Effective AI training needs to be personalised to what each person actually does. A sales director needs to learn different things from an operations manager. A marketing coordinator will use AI differently from a finance lead. Generic "introduction to AI" sessions might raise awareness, but they rarely change behaviour. Yes a basic foundation course is great, but taking the application to a more personalised level thereafter is a real game changer.
Training should be hands-on and practical. Show people how to use AI for their specific tasks - their proposals, their reports, their research, their customer communications. Give them frameworks for briefing AI properly, not just typing questions. Teach them how to review and refine AI outputs so the quality is consistent. And importantly, teach them what AI isn't good at - so they develop judgement alongside skill.
Remember the stat: 30% of the workforce has had zero AI training and 61% have spent fewer than five hours on it. The bar is incredibly low. Even a focused half-day of practical, role-specific training puts your team ahead of the vast majority.
And the effect of training goes beyond skills. Our survey showed that regular AI users were more than
twice as likely to be excited about AIrather than cautious. Experience doesn't just build capability - it builds confidence and shifts sentiment. Training is the single fastest way to move people from fear to enthusiasm.
Depending on your company, a pilot training and use case approach with the employees who are keen may just show all the other sceptics in the company the benefits. Not all employees are equal. Gaining an insight into the attitudes beforehand around AI is also key, as transforming your business with AI is as much a cultural, change management process as it is a systems and technology one!
FAQ
Do I need to invest in expensive AI tools to get started?
Not necessarily. ChatGPT, Claude and Gemini all have free tiers or relatively low-cost subscriptions. The bigger investment is time - time to audit your current position, build a strategy and train your people properly. A £20 per month AI subscription in the hands of someone who's been trained to use it well will outperform a £500 enterprise licence used by someone who's still typing "rewrite this email" into a chat box. Also a piloting licences with a few key individuals is a great idea - depending on the size of your business.
What's the difference between unstructured experimentation and a proper AI pilot?
Unstructured experimentation is when individuals try AI on their own, with no coordination, no measurement and no shared learning. A proper pilot picks a specific workflow or task, defines what success looks like, trains the people involved, runs it for a set period and measures the results. One produces random anecdotes. The other produces data you can act on. In our survey, 29% of engineering companies were stuck in unstructured experimentation - lots of tinkering, no results.
Don't Get Distracted by the Shiny Stuff
This is the part where I want to be direct with you.
There is an enormous amount of noise right now about agentic AI, autonomous systems and fully automated workflows. The vendor pitches are slick. The case studies are compelling. The LinkedIn posts make it sound like everyone is building AI agents that run entire departments.
The research tells a different story. McKinsey found that 62% of organisations are experimenting with AI agents, but fewer than 10% have scaled them in any function. Only 23% are actively building towards agentic systems, and fewer than 5% of enterprise applications include agentic features in 2025 McKinsey Global Survey: The State of AI in 2025.
OpenAI's own data shows that among enterprise customers, custom GPT and Projects usage grew 19 times year-to-date - which sounds impressive until you realise that 20% of enterprise messages now flow through these custom setups. That means 80% of enterprise AI usage is still basic chat OpenAI: The State of Enterprise AI 2025 Report.
Agentic AI, custom knowledge bases, integrated data systems and automation - these are all coming. They matter. They will reshape how businesses operate. But jumping to those before your team can confidently use a basic AI chat is like installing a CNC machine before your operators can read engineering drawings. The sequence matters.
The companies that will genuinely benefit from advanced AI are the ones that first built the foundation - access, skills, confidence and a strategy for how AI fits into their operations. Without that, advanced tools just become expensive experiments that nobody uses properly.
As for AI app -there are hundreds - all customising and tailoring how AI is used. Many of these functions can be done using a handful of tools, frontier models like chat GPT, Gemini, Claude can do more than people think. Starting with these and leveraging them fully first is the way to go and where your workflows and needs start developing and your adoption is maturing, the individual apps may be the way to go. Either way the frontier models can give you guidance on this.
"Waiting for the next update is not a strategy, it's falling behind."
FAQ
Should my company be looking at agentic AI?
Eventually, yes - but probably not right now. Agentic AI involves AI systems that can take actions, make decisions and complete multi-step tasks on their own. It's powerful, but it requires a level of organisational readiness that most SMEs haven't built yet. If your team isn't using basic AI tools confidently and regularly, agentic systems will add complexity without delivering value. Master the fundamentals first, then scale up when your people and processes are ready.
What This Means for Your Business
If you're reading this as an MD, commercial director or operations director of an SME, here's what all of this data actually means for you.
You're probably not as far behind as you think. The majority of businesses - including most of your competitors - are either not using AI or using it in the most basic way possible. The headlines are misleading. The reality is that most organisations are still at the starting line.
But that doesn't mean you can afford to wait. The gap between those who are taking structured steps and those who aren't is widening fast. BCG found that the top 5% of AI-mature companies are pulling away from everyone else, and the advantage compounds over time.
The opportunity right now isn't about buying the best tool or building the most advanced system. It's about doing three things that most of your competitors haven't done: auditing where your organisation is, building a clear strategy for adoption, and training your people to use AI with confidence and purpose.
These three stages alone - without touching data integration, automation or agentic systems - can produce meaningful productivity gains and fundamentally change how your team works. Faster proposals, better research, sharper messaging, more consistent output, less time on repetitive tasks.
The problem isn't the technology. The problem is fear of the unknown, a lack of investigation and a lack of coordinated action. The businesses that face it, take stock and start building from the ground up will be the ones with the advantage - not in five years, but right now.
How to Future-Proof Your Organisation
The pace of AI development isn't slowing down. New models are released every few months. Capabilities that seemed futuristic 12 months ago are now standard features. The tools will keep getting better, faster and more accessible.
But tools without skills are just expensive subscriptions. And skills without strategy are just random productivity boosts that never compound into something bigger.
Future-proofing your organisation doesn't mean chasing every new development. It means building the capability to absorb and use whatever comes next. That means a team that's comfortable with AI, confident in how to evaluate and use new tools, and working within a clear framework that connects AI usage to business goals.
Start with the audit. Build the strategy. Invest in proper, personalised training. Then iterate - review what's working, expand what's delivering results, and keep developing your team's skills as the technology evolves.
The companies that win in the next three to five years won't be the ones with the fanciest AI systems. They'll be the ones that built the skills, the confidence and the structure to use whatever comes next - and started that process now.
Your technical operations run with precision because you engineered them that way. Your approach to AI deserves the same discipline. Strategy first. Skills second. Tools third. Everything else follows.
Ready to take the first step? Book a
free strategy audit and let's work out exactly where your business is with AI - and build a practical plan to move forward. Or explore our
AI for sales and marketing services to see how we help engineering and technical SMEs adopt AI with purpose.
Written by Stefan Buss, founder of Sales & Marketing Engineers. With a background in industrial engineering and over two decades helping technical companies systemise their growth, Stefan brings an engineer's precision to sales, marketing and AI adoption strategy.
