AI agents in plain English, and where they actually help

There’s a lot of breathless talk about AI at the moment, and not much of it tells you anything useful. So here’s the plain version, with no jargon and no hype: what an AI agent actually is, where it earns its keep for a normal business, and where it doesn’t.

We build these now, alongside the sites and apps we’ve always made. But we build them the same way we build everything else: only when they solve a real problem, and only as much as the job needs.

What an agent actually is

A chatbot answers a question. An agent does a job.

That’s the difference worth holding onto. The old website chatbot was a glorified search box: you typed a question, it matched some keywords, and it spat back a canned answer or told you it didn’t understand. An AI agent is different in kind. It can read your actual content and work out what someone’s really asking even when they phrase it oddly. Then it does something about it: books the appointment, checks the order, raises the ticket, or hands a tricky one to a human.

Think of it less as a smarter FAQ and more as a junior member of staff who works every hour of the day, never forgets your policies, and only does the specific tasks you’ve trained it on.

Where it genuinely helps

Agents earn their place when there’s repetitive, predictable work that still needs a bit of judgement. A few cases where we’ve seen them pay off:

• Answering the same questions, endlessly. Opening hours, do you cover my area, what’s your returns policy, can I bring my dog. The questions that eat your evening one message at a time.
• Catching enquiries you’d otherwise lose. Someone lands on your site at 11pm with a question. A human won’t reply until morning. By morning they’ve booked someone else. An agent answers there and then, and captures the lead while they’re still interested.
• Qualifying before it reaches you. Instead of a flat contact form, an agent can ask the two or three questions that tell you whether an enquiry is worth your time, and pass on a tidy summary rather than a one-line “interested, call me”.
• Internal donkey work. Pulling the right answer out of a pile of documents, drafting the first version of a reply, summarising a long thread. The stuff that isn’t hard, just slow.

The thread running through all of these is the same: the agent handles the volume, and a human handles the exceptions.

Where it’s the wrong tool

We’re as wary of AI hype as we are of the hype around websites versus apps, because pretending AI fixes everything is how people waste money.

An agent is the wrong tool when the cost of being wrong is high and rare. Anything involving real money moving, legal or medical advice, or a decision you’d want to stand behind personally, should go to a person. An agent is also pointless if you don’t actually have the volume to justify it. If you get three enquiries a week, you don’t need an agent to answer them, you need a phone.

And an agent is only ever as good as what it’s working from. Point it at a thin, out-of-date website and it’ll give thin, out-of-date answers, confidently. The groundwork still matters.

It still has to be built properly

This is the part the hype skips. An agent dropped onto a site without care will happily make things up, leak information it shouldn’t, or send people in circles. The value is entirely in the boring details: what it’s allowed to say, where it gets its facts, when it hands off to a human, and how you keep an eye on what it’s telling people.

That’s the same discipline we bring to building by hand. An agent is software, and software you can’t see into or change is a liability, not an asset. We build them so you can read what they’re doing, adjust what they say, and switch them off if you ever want to.

The honest version

An AI agent isn’t magic and it isn’t a member of staff. It’s a tool that’s very good at a narrow set of repetitive, language-heavy jobs, and useless or dangerous outside them. Used well, it quietly handles the stuff that was eating your time and losing you enquiries. Used badly, it’s an expensive way to annoy your customers.

If you’ve got a job in mind and you’re not sure which it is, tell us what you’re trying to do and we’ll give you a straight answer, including “you don’t need one” if that’s the truth.