80% Have AI Agents, Only 31% Use Them
Quick answer
Gartner says 80% of enterprise apps now embed AI agents, but only 31% of firms run one in production. The gap isn't the technology — it's data quality, governance and workflow design. Here's what UK SMEs should do before switching agents on.
Gartner says 80% of enterprise apps now embed AI agents, but only 31% of firms run one in production. The gap isn't the technology — it's data quality, governance and workflow design. Here's what UK SMEs should do before switching agents on.
I see this all the time with UK SMEs. They've got Copilot sitting in their M365 license, a chatbot widget on the website, maybe a Power Automate flow with an AI step in it. So they tick the box — "yeah, we use AI agents."
But when you ask what the agent actually does for them day to day, the answer is usually "well, it's there if we need it."
That's not using AI. That's having AI sitting on the shelf gathering dust.
The numbers are proper eye-opening
Gartner forecast that by the end of 2026, roughly 80% of enterprise applications will embed at least one AI agent. Up from under 5% in 2025. That's a massive jump. The software vendors have been stuffing AI into everything — your CRM, your ERP, your project management tool, all of it.
But here's the kicker. S&P Global and McKinsey report that only 31% of enterprises actually run an AI agent in production. Meaning one in three firms has moved beyond the "it's there" stage to "it's doing real work, every day, on live data."
Embedding is easy. Operating is hard.
And Gartner expects over 40% of agentic AI projects to be cancelled by 2027. Not paused. Cancelled.
Why the gap exists
The research points to a few proper blockers, and they'll sound familiar if you've ever been through an ERP go-live:
- Data quality. 52% of firms cite it as the single biggest blocker to deploying AI agents. If your master data's a mess, your agent will confidently make wrong decisions at speed.
- Governance. Only 21% of organisations have a mature governance model for autonomous AI agents. Nobody's decided who's responsible when the agent gets it wrong.
- Unclear ROI. IBM's 2025 CEO study found only 25% of AI initiatives delivered expected returns. Hard to justify production rollout when the pilot didn't move the needle.
None of these are technology problems. They're the same foundation problems that sink ERP implementations, Power BI rollouts, and every other "digital transformation" initiative I've seen in 15 years of doing this.
What the winners do differently
The firms that are running agents in production — the 31% — share a few habits:
- They sorted their data first. Clean master records, clear ownership, one version of the truth. Boring, unglamorous, absolutely essential.
- They scoped agents narrowly. Not "an AI that runs our finance function." Something like "an agent that flags duplicate supplier invoices before they're paid." Vertical, specific, measurable.
- They kept humans in the loop. The agent suggests, a human approves. Especially for anything touching money, customers, or compliance.
BCG and Forrester clocked the median time-to-value for agent deployments at about 5.1 months. That's not bad. But it only happens when the groundwork's been laid.
What UK SMEs should actually do
If you're sitting on AI capabilities you're not using, here's where to start:
- Audit your data before you audit your AI. Are your customer records, product codes and supplier details clean and consistent? If not, fix that first. The agent is only as good as the data it reads.
- Pick one painful, repetitive, low-risk process. Something like invoice chasing, stock reorder alerts, or flagging unusual purchase patterns. Start there. Prove the value. Then expand.
- Decide who's accountable. Before you switch an agent on, name the person who owns its outputs. If nobody owns it, nobody will trust it.
The 49% gap between "has AI" and "uses AI" isn't a technology gap. It's a readiness gap. And readiness is fixable — you've just got to be honest about where you are.
References & Further Reading
- Gartner, Forecast: 40% of Enterprise Applications Will Embed AI Agents by End of 2026
- S&P Global Market Intelligence / McKinsey, The State of AI 2026
- BCG / Forrester, AI Radar 2026: Median Time-to-Value on Agent Deployments
- Gartner, Over 40% of Agentic AI Projects to Be Canceled by 2027
Not sure if your data's ready for AI?
I help UK SMEs get their data house in order before switching on AI agents. Data audits, master data cleanup, governance frameworks — the unglamorous stuff that makes the shiny stuff actually work.
Let's have a chatMatty Hatton is the founder of Digital Adaption, an ERP and data consultancy based on the Wirral. He has spent 15 years delivering ERP transformations for manufacturers, including leading the data migration on a £4.5m consolidation of four legacy systems onto a single Infor LN cloud instance for a 220-user group. He holds an MSc in Digital Transformation and IT Strategy from Manchester Metropolitan University and is Microsoft PL-200 certified.