Insights

What is an AI agent, and what should we automate first?

For operations leadsVerified August 2026

The decision an operations lead actually faces with AI agents is which process to hand one first, and how to do it without creating a mess. An agent takes a goal, works across steps and systems to carry it out, checks its own progress, and keeps going until the job is done or it hits a limit you set. A chatbot answers a prompt and stops. That shift from answering to doing is why agents have moved up the agenda, and why the first process you pick deserves care.

What an agent actually is

Picture the difference through a task. Ask a chatbot to "chase our overdue invoices" and it writes you a tidy reminder email to send yourself. Give the same goal to an agent and it pulls the overdue list, drafts a reminder tuned to each account's history, sends the routine ones, flags the awkward ones for a person, and logs what it did.

The agent has three things the chatbot lacks: it can take actions in your systems rather than only producing text; it works across multiple steps towards a goal rather than one turn at a time; and it operates with some autonomy, deciding what to do next within limits you define. None of that removes the need for oversight; if anything it raises it. The payback comes from letting the agent do the running rather than you.

Where agents help, and where they don't yet

Agents earn their place on work that is repetitive, rule-shaped, and spread across systems, the tasks where the cost is coordination rather than judgement. Chasing invoices, triaging inbound requests, reconciling records between two systems, first-pass drafting, monitoring for changes and routing them: all of these play to the strengths.

They struggle where the work turns on scarce context, high-stakes judgement, or steps a mistake makes expensive to undo. An agent that files the wrong thing is a nuisance; an agent that pays the wrong supplier is a problem. As a rule of thumb, the more consequential and irreversible the action, the more the human belongs in the loop, and the earlier you should ask whether an agent is the right tool for that step at all.

How to choose the first process

Resist the instinct to automate the most painful process first. The opening agent should be chosen to succeed and to teach, not to conquer your hardest problem. Score candidate processes against five questions.

Question You want Why it matters
Volume High and repetitive Enough repetition for the effort to pay back
Rules Clear and stable Agents follow patterns; fuzzy, shifting rules break them
Reversibility Mistakes are cheap to undo Keeps the first project low-risk while you learn
Data access The systems it needs are reachable An agent that cannot reach its data cannot act
Measurable outcome A number that moves Proof it worked, and the case for the next one

Score well on all five and you have your first agent: high-volume, rule-based, low-stakes, connected, and measurable. Inbound-request triage or invoice chasing usually fits; a nuanced pricing negotiation does not. Both of the first two are also commodity work, which is the kind you buy and configure rather than build.

Keep a human in the loop

An agent earns its place by taking the coordinating drudgery off a decision, so people can spend their judgement where it counts. Set the boundaries deliberately: what the agent may do on its own, what it must hand to a person, and what it must log. Where agents deal with customers or shape decisions about people, disclosure and oversight go beyond good practice: under the EU AI Act's transparency rules they are an obligation from 2 August 2026.

Start with one

The mistake is to plan an agent programme before you have run a single agent. Take one process that scores well on the five questions, give it clear limits and a human backstop, measure it against the number it was meant to move, and learn from what it gets wrong. One working agent, judged honestly, teaches you more about where agents fit in your business than any strategy deck, and it earns the right to the second.

First process, safe wiring, sensible oversight, a human in the loop by design. Firestarter sets the process-and-agent readiness up in its six-week accelerator.

Sources and verification. This guide describes the current capabilities and sensible adoption patterns for AI agents as understood in August 2026, a fast-moving area. Both the capabilities and the regulatory picture continue to change; validate any agent deployment against your own risk, data-protection, and EU AI Act obligations before it acts on live systems.

  • Technology lead

    Should we build or buy AI agents?

    For most mid-market companies, building AI agents from scratch is the wrong instinct. A technology lead's guide to the build-buy-configure spectrum, the hidden cost of building, and a pragmatic default.

    Read the guide

  • IT & data lead

    Is my data ready for AI?

    Whether your data is ready for AI depends entirely on what you ask AI to do with it. An IT lead's guide to the four things that actually matter, the permissions trap that catches everyone, and where to start.

    Read the guide

  • Chief financial officer

    Why do AI pilots fail to pay back?

    Most AI pilots never reach production, and fewer still return value the board can see. A CFO's guide to why AI pilots stall short of payback, and the four things the ones that pay back do differently.

    Read the guide

  • Chief financial officer

    What should AI cost my business in year one?

    The licence fee is the small part. A CFO's guide to the four real cost lines of AI in year one (tools, enablement, integration, and governance) and why the biggest one is the one most budgets forget.

    Read the guide