AI & automation

AI in your business, minus the hype.

Most businesses are being sold AI they do not need, or ignoring it entirely and hoping it goes away. There is a sensible middle: find the handful of jobs eating your week, automate those properly, and leave the rest alone.

Why it matters

The gap is opening quietly.

Nobody is going to be put out of business this year by not using AI. What is happening is slower and harder to spot: the competitor who automated their quoting now replies in four minutes instead of two days, and starts winning the jobs you used to get on price.

The businesses getting value out of this are not the ones who bought the most impressive thing. They are the ones who picked a single, genuinely annoying process and fixed it properly, then did the next one.

Admin

Is the usual first winQuoting, invoicing, chasing, scheduling, data entry. Unglamorous, endlessly repetitive, and exactly what this is good at.

Speed

Is where it shows up commerciallyReplying first wins work. Almost every automation worth doing ends up being a speed story.

People

Stay where judgement is neededThe aim is not fewer staff. It is your staff spending their day on the parts that need a human.

Straight answers

What’s worth doing, and what isn’t.

Every agency in this space will tell you everything is worth doing. Here is our actual view, which occasionally costs us work.

Usually worth it

  • Answering enquiries instantly The single highest-return thing for most businesses. Whoever replies first usually wins, and this never sleeps.
  • Quoting and estimating Where a repeatable calculation is buried in someone’s head or a spreadsheet, turning it into something instant is transformative.
  • Getting data out of documents Invoices, purchase orders, timesheets, forms. Genuinely tedious, genuinely well suited, and the error rate usually improves.
  • Drafting the repetitive writing Job adverts, product descriptions, follow-up emails, reports. A first draft in seconds that a human then makes good.
  • Making your own data answerable Being able to ask a plain question of your own records, instead of building a report every time you wonder something.

Usually isn’t

  • Anything customer-facing and unsupervised If a wrong answer costs you a customer or a claim, a person needs to be in the loop. This is where most of the horror stories come from.
  • Replacing work that is already fast If a job takes four minutes a week, automating it is a hobby, not an investment. Measure before you build.
  • Anything requiring real judgement Pricing a difficult job, handling an upset customer, deciding who to hire. Automating these produces confident nonsense.
  • Buying a platform before you have a problem Plenty of businesses are paying monthly for something nobody opens. Start with the problem, not the product.
  • Anything you cannot explain to a customer If you would be uncomfortable telling them how a decision was made, do not automate that decision.

How we work

One thing at a time, starting with the worst one.

No transformation programmes, no eighteen-month roadmaps. Find the job that wastes the most time, fix it, prove it worked, move on.

  1. 01

    A proper look at how you actually work

    Half a day with the people doing the jobs, not just the person who signs things off. The bottleneck is almost never where the org chart suggests it is.

  2. 02

    We tell you what is worth automating

    Including when the answer is nothing yet. You get the list either way, with a rough sense of what each is worth in time and money.

    Even if you take that list elsewhere
  3. 03

    Build the first one properly

    One process, done well, with a human checkpoint wherever a wrong answer would matter. Live in weeks, not quarters.

  4. 04

    Measure whether it actually helped

    Time saved, errors avoided, response times. If it did not move, we say so rather than quietly moving on to the next thing.

  5. 05

    Then the next one

    Once something is genuinely working and your team trusts it, the second is much easier. Most clients build up over a year or two.

Why us

We build this into our own work.

This is not a service line we bolted on because the word got popular. The intelligent butler capabilities we offer on our own websites are the same discipline at a smaller scale: instant lead response, review gathering, reporting that writes itself. Real things, running on real client sites.

It is also a field that changes every few months, which means most of the value we add is knowing what has actually become possible since the last time you looked, and what is still a demo that does not survive contact with your data.

What you will not get is enthusiasm as a substitute for judgement. If the honest answer is that your money is better spent on a better website, or on hiring someone, we will say that.

Questions

The things people are too polite to ask.

Is this going to replace my staff?

That is not what we sell and it is rarely what happens. In practice the work that gets automated is the work nobody wanted: the retyping, the chasing, the copying between systems. Your people end up doing the parts that need a person. If your goal genuinely is headcount reduction, we are probably not the right people, because we would design it differently.

What about our data? Is it safe?

A fair question and one you should ask everybody. It depends entirely on what gets built and where it runs. Some things can stay wholly within your own systems; others involve a third party, in which case you should know exactly which one and what they do with what you send. We will tell you plainly, in writing, before anything is built. If you are in a regulated sector, say so at the start because it changes the options.

How much does it cost?

It depends far too much on the process to put a number on this page. A single well-defined automation is usually a modest project. A proper look across the business costs more and often saves more. What we can promise is that you get a real figure before anything starts, and that we will tell you the expected payback rather than just the price.

We tried something like this and it did not work. Why would this?

Usually one of three reasons: it was built for a process nobody had actually mapped, it had no human checkpoint so people stopped trusting it, or it solved a problem that was not costing anything. All three are avoidable, and all three come from starting with the technology instead of the job.

Do we need to be technical to work with you?

No. If anything it helps that you are not, because it forces us to explain things in terms of what changes for the business rather than how it works underneath. You need to know your own processes. That is the expertise that matters here.

Does this mean our website is built by AI?

No. Those are two entirely separate things and we keep them that way. Your website is designed and built by people, deliberately, because that is what makes it good. This service is about putting sensible automation into how your business runs. Different job, different conversation.

What’s the job you keep putting off?

Tell us the thing that eats your week. We’ll tell you honestly whether it is worth automating, and roughly what that would take.