AI for business: what can AI mean for your organisation?

What AI can mean for your business

AI for business sounds big, but the value is usually in small, repeated tasks. A quote that’s ready in fifteen minutes instead of an hour. A customer query that lands with the right colleague straight away. A contract that’s already been checked for unusual clauses. A few minutes at a time, but a thousand times a month.

What AI is genuinely good at in 2026 is working with language and unstructured information: summarising text, searching documents, drafting copy, pulling data out of emails and PDFs, and spotting patterns in large volumes of data. What it can’t do is carry responsibility. An AI application makes mistakes, and the question is always whether you catch those mistakes in time and what they cost.

That’s why I start with the work, not the technology. The question isn’t “what can we do with AI?”, but “which work is costing us too much time, money or quality right now, and does AI help with that?”. That distinction sounds small. In practice, it decides whether you end up with a working application after a year, or a pile of pilot projects. Why adoption starts with the work, not the tool, is a theme I come back to across this site.

Across sectors, the first wins tend to come from similar places: marketing and sales content, and admin or back-office work. The businesses that hold back most often cite a lack of experience with AI, rather than cost or technology, as the main reason. That’s exactly where the right approach makes the difference.

Applications by department

The opportunities sit in different places per department. A few recognisable examples, all based on work that comes up often and has a similar shape:

• Customer service: categorising and routing queries, suggesting draft replies that a staff member checks, and answering frequently asked questions from your own knowledge base.
• Sales: drafting quotes and proposals based on earlier documents, summarising call notes, and preparing customer information ahead of a meeting.
• Marketing: producing variants of copy, rewriting content for other channels, and analysing customer feedback for recurring themes.
• Finance and admin: reading invoices and receipts, flagging anomalies, and adding plain-language commentary to reports.
• HR: drafting job adverts, answering internal questions about policies, and keeping onboarding materials up to date. Be cautious with anything that touches decisions about people.
• Operations and IT: summarising incident reports, making documentation searchable, and supporting planning with forecasts.
• Management: pulling information from different sources together and interpreting reports faster. This is where AI overlaps with business intelligence.

Many of these applications come down to automating recurring steps in a process. How you decide which processes are suitable for that, and how to do it without automating a process you’d be better off dropping, is covered on the AI automation page.

AI by industry

The departments look similar across sectors; the rules of the game don’t. In financial services, oversight and explainability set the boundaries. In healthcare, privacy and patient safety weigh heavily. In government, transparency and equal treatment play a bigger role. In education, it’s also about what students are and aren’t allowed to use AI for.

For two groups in particular, the questions are similar enough that they deserve their own treatment:

• Small and medium-sized businesses. Less volume, no in-house data team and a tighter budget, but short lines of communication. Which applications work there, and how to start without a big budget, is covered on AI for small business.
• Professional services firms. Accountants, lawyers, advisers and agencies work almost entirely with documents and knowledge. That makes AI especially useful there, but confidentiality and professional rules need extra care.

Opportunities and risks

AI offers three kinds of opportunity. Time: less manual work on recurring tasks. Quality: fewer errors and more consistent output, provided there’s a check step. And new possibilities: information that used to be out of reach, such as thousands of documents or customer messages, suddenly becomes usable.

Not every opportunity is equally valuable. The applications that succeed tend to share a few traits: high volume, low variation, data that already exists, and an error that’s visible and affordable.

The risks are rarely technical in nature:

• No owner. A pilot that works but isn’t picked up by anyone in the business evaporates. That’s the most common reason AI pilots stall.
• No upkeep. AI applications need ongoing maintenance. Whoever doesn’t plan for that sees quality drift downward without noticing.
• Privacy and confidentiality. Putting customer or staff data into a tool that isn’t set up for it is a data protection risk and a trust risk.
• Regulation. The AI Act sets requirements that depend on an application’s risk level, and its reach isn’t limited to EU-based organisations: under Article 2 of the Act, it also applies to UK organisations that place an AI system on the EU market, or whose AI system’s output is used in the EU, regardless of where they’re established. Under the Digital Omnibus on AI, which took effect on 27 July 2026, the rules for the high-risk areas in Annex III, such as recruitment and credit scoring, have been delayed until 2 December 2027, and for AI embedded in regulated products until 2 August 2028. The UK has no AI Act of its own: it regulates AI through existing sector regulators instead, applying shared principles rather than one cross-cutting AI law, as set out in the UK government’s AI regulation: a pro-innovation approach. A UK-only organisation with no EU market presence and no EU users of its AI output sits outside the Act’s scope; everyone else should know which category their applications would fall into. A delay isn’t a cancellation. How to arrange that without creating a paper exercise is covered under AI policy and governance.
• Shadow use. If the organisation doesn’t offer an approved tool, staff use their own. Then you carry the risk without any oversight or control.

Getting started with AI

You don’t need to start with a fifty-page strategy. You do need a few deliberate choices, in this order:

A few practical points to go with it:

1. Start with using, not building. A lot of value sits in tools and AI features you already have or can easily buy. Custom development only makes sense once standard tools clearly fall short.
2. Set a few ground rules. Which tools are allowed, which data can’t go into them, and who decides. One page is enough to start.
3. Invest in skills. Staff who know what AI can and can’t do get more out of it and make fewer mistakes. A workshop or training session built around their own tasks works better than a generic demo.
4. Look beyond the licence. Integration, upkeep and your own people’s time are often bigger than the cost of the tool itself.

Anyone who wants more clarity before committing budget can start with an assessment. More on that in the next section.

How I can help

I’m an independent AI advisor. I advise, build strategy and roadmaps, select vendors, test business cases, set up governance, and guide implementations from your side of the table. I don’t build or sell software myself, and I don’t earn anything from the tools or parties I advise on.

Concretely, I can help in these ways:

• Knowing where you stand. With an AI Readiness Assessment I map strategy, data, technology, people and governance in 2-3 weeks, with a 90-day plan as the result. Fixed price: €7,000.
• A shared picture at the top. An Executive Briefing of one half-day gives your leadership team or board a realistic starting point.
• Strategy and roadmap. From scattered ideas to choices, priorities and a plan. Everything about this form of advice is covered under AI consulting.
• From plan to working application. I guide the rollout, steer the delivery partner and make sure it lands in the day-to-day work. See AI implementation.
• Choosing and testing. A Vendor Review, an AI Due Diligence, or a second opinion on a proposal that’s already on the table.

The building itself is done by your own team or a party you contract directly. Where that’s useful, I introduce partners from my partner network, with no margin on their rate. What the different forms cost is set out clearly on the pricing page.

If you have a specific question, book a thirty-minute call. By the end you’ll know what I’d do, even if that means you don’t need me.