AI for businesses that aren't “AI companies”
Most businesses don't need to become an “AI company” to get real value from AI — they need their existing data joined up enough that AI has something coherent to work with. That's a data and systems project, not a research project.
What does “AI for businesses that aren't AI companies” actually mean?
It means using AI as a practical tool inside an ordinary business — a manufacturer, a distributor, a professional services firm — rather than building AI products or hiring a data science team. Most UK mid-market businesses fall into this category, and most of the AI commentary aimed at “AI companies” simply doesn't apply to them.
Where does AI actually help a business like this?
Wherever there's structured, joined-up data and a routine task built on top of it — surfacing patterns in operational data, accelerating suitable routine work, automating tasks where the process is well-defined enough to support it. AI supports people making decisions; it doesn't replace their judgement, and outputs that could materially affect operational, legal or financial decisions should be reviewed by an appropriate person before being relied upon.
What has to be true before AI is useful, not just novel?
The data has to actually be joined up. AI applied to fragmented data — scattered across an ERP, a CRM and a dozen spreadsheets that don't talk to each other — mostly produces fragmented, unreliable results. The unglamorous work of connecting your systems is what makes the AI step afterward actually useful, rather than a demo that doesn't survive contact with the real business.
Where should a non-AI business actually start?
Not with AI. With finding out where your data and systems actually stand — the free AI Exposure snapshot measures that directly, across data, integration, traceability, containment and leadership visibility, and names your single biggest constraint. For most businesses, it isn't AI at all.
RELATED READING