Nottingham's AI Moment: Atos's Digital Agentic Centre and What It Means for SMEs

What did Atos actually announce for Nottingham?

According to the East Midlands Combined County Authority's announcement on 20 October 2025:

• £10m investment by Atos, the IT services company, in a Digital Agentic Centre in Nottingham focused on AI and automation
• A second hub in Birmingham as part of the same investment
• 50 graduate and apprentice jobs across the Midlands in 2026
• Support from the region's mayor, Claire Ward, who said: "Supporting the development of digital skills and employment is a cornerstone of my newly unveiled East Midlands Growth Plan"

What the announcement didn't give, and we couldn't confirm: the centre's exact location, its opening date, how many roles are in Nottingham rather than Birmingham, and which clients it will serve. Atos is also the partner behind the British Chambers of Commerce's 2026 survey of AI in the workplace, quoted below.

Fifty roles won't transform Nottingham's job market by themselves. The signal is more useful than the number: a large IT services firm chose Nottingham for agentic AI work.

What does "agentic AI" mean, in plain English?

An ordinary AI chatbot answers a question. An AI agent is given a goal, access to specific tools and data, and permission to work through the steps itself.

A concrete example from our own work: an agent that opens a purchase-order email, reads the attached PDF, matches each line to the right product in a catalogue, drafts a quote and then stops for a person to approve it. No single step is clever. Doing all of them reliably, with variation in every order, is what makes it an agent.

Two cautions are worth knowing:

• The label is being stretched. Gartner forecast in June 2025 that more than 40% of agentic AI projects will be cancelled by the end of 2027, citing cost, unclear value and weak risk controls, and estimated that only around 130 of the thousands of vendors claiming agentic products are genuine. It called the practice "agent washing".
• Responsibility doesn't move to the software. The ICO's January 2026 report on agentic AI states that "AI agency does not remove human, and therefore organisational, responsibility".

Why Nottingham?

Nottingham is the region's largest concentration of business and digital work, which makes it a logical base:

• 9,940 registered enterprises in 2026, up 3.8% on 2025, including 570 in information and communication, the most of the four East Midlands areas we compared (ONS via Nomis, 2026)
• £11.5bn of economic output (GVA) in 2022, and £19.2bn across the wider conurbation, nearly 15% of the East Midlands total (Nottingham City Council)
• Headquarters for Boots, Experian, Capital One and E.ON (Invest in Nottingham)
• Two universities, the University of Nottingham and Nottingham Trent

There may be more to come. An East Midlands AI Growth Zone bid was described as "under active consideration" in the House of Commons on 18 March 2026, with Nottinghamshire benefits mentioned, but nothing has been confirmed.

Where are Nottingham's small businesses with AI right now?

The national data is useful because it shows how far apart the headlines and the reality are. The surveys disagree on the numbers because they measure different things:

Whichever figure you prefer, three points hold across all of them:

• Most use is shallow. The most common use is generating text (18% of firms with 10+ staff), and only 10% of adopters describe their use as "extensive" (ONS, 2026).
• Bespoke systems are rare. A University of Essex working paper estimated about one in ten firms have implemented a bespoke AI system (ISER, 2026).
• The barriers are knowledge and confidence. The ONS lists lack of expertise, cost and difficulty identifying use cases; Simply Business found security and privacy (44%), no clear use (39%) and accuracy (36%) as the top worries.

That gap is the real story for Nottingham. The big firms are building agentic centres. Most small firms haven't yet automated a single process.

What should a Nottingham SME do about it?

Not panic, and not buy a platform. Our suggestion, based on how our own AI project went:

1. 1Pick one process that costs hours every week. Order entry, quote preparation, invoice matching, booking admin, chasing paperwork.
2. 2Measure it now. How long does it take, how often, and what goes wrong? Without a baseline, you can't tell if the automation worked.
3. 3Check it's suitable. Good candidates are frequent, rule-heavy and tedious. Poor candidates are rare, high-stakes or different every time.
4. 4Keep a person in the loop. Let the system prepare the work and a human approve it. The ICO's draft guidance on automated decisions says token human review isn't enough; the reviewer must understand the output and be able to override it.
5. 5Own what's built. Make sure you keep the accounts, workflows and data if you change supplier.
6. 6Check funding before you start. See below.

RAND's research on why AI projects fail put "misunderstanding the problem" first among five root causes (RAND, 2024). Steps one to three exist to avoid it.

Is there funding for AI projects in Nottingham?

Possibly, though nothing is guaranteed and amounts are still emerging:

• EMCCA Business Investment Fund. Announced 7 September 2026 and launching late 2026, it includes a Business Growth and Productivity Grant covering technology and innovation for SMEs in Nottingham, Nottinghamshire, Derby and Derbyshire, applied for through the East Midlands Growth Hub. Around 500 businesses are expected to be supported; grant sizes aren't published yet.
• Local Growth Fund. EMCCA's £107m fund, approved in June 2026, includes £12.3m for innovation and cluster development and £2.4m for local business support (EMCCA, June 2026).
• Made Smarter East Midlands. Matched grants of up to £20,000 for manufacturers adopting digital tools, listed to March 2026 with an extension expected. Our Made Smarter guide explains eligibility.

What does this look like for a small firm in practice?

Our published AI project is with Webro, a Midlands manufacturer and cable supplier. It came to us through a referral from our East Midlands Chamber Business Advisor.

Webro's customers send purchase orders as PDFs with hundreds of lines, using their own part numbers rather than Webro's. Matching them by hand could take up to 48 hours for a large order. Another team had taken the project on with a three-month timeline and handed it back undelivered.

We proposed three weeks after reviewing the brief, and delivered an AI agent that receives the email, extracts the PDF, reads each line, matches parts using the descriptions and prepares a quote for the team to review and send. For the complex orders it was designed for, the work now takes around two minutes. The director's reaction:

That is what an agentic system looks like at SME scale: one process, measured before and after, with a person making the final decision.