AI & Automation · Evidence
Case studies with dates, not adjectives
Anyone can say “AI transforms your business”. We’d rather show you the systems running our own business right now — what each one does, when it shipped, and where a human stays in charge. Every date below is real. Where a number wasn’t measured, it isn’t here.
- ● Our own systems, dated
- ● No invented numbers
- ● A person approves the big stuff
/01 — WHY WE CASE-STUDY OURSELVES FIRST
Why we case-study ourselves first
We looked at every AI company serving Bournemouth, Poole and Dorset and could not find one published case study a reader could actually verify. We’re not going to add to that pile. Our first case studies are the systems we run our own company on — because you can see them working, we can show the real dates, and nobody had to approve a press release. Client case studies will join this page as projects launch, with the client’s permission and their numbers, not ours.
One more honest note: some of what follows is deterministic automation rather than AI — rules, not guesses. We label which is which, because knowing when not to use AI is most of the job.
/02 — CASE STUDY: THE ENQUIRY THAT USED TO VANISH
Case study: the enquiry that used to vanish
Type: deterministic automation, live since 1 August 2026.
The problem.
Someone picks a service on our booking page, chooses a time, types their name and phone number — then never enters the emailed confirmation code. Before this system, that person left no trace at all. The warmest lead a website makes, gone silently.
What we built.
The booking system now parks those details safely the moment they’re typed. A scheduled job sweeps every fifteen minutes; anything older than twenty-five minutes triggers one alert to our team chat — name, service, the slot they wanted — so a human can ring back while the enquiry is still warm. It reports each person exactly once, never texts anyone, and clears itself after 24 hours.
Where the human stays in charge.
The system never contacts the customer. It tells us, and a person decides.
Verifiable.
Try it: start a booking at book-service and abandon it. We’ll be the ones who call you.
/03 — CASE STUDY: THE REVIEW REQUEST THAT SENDS ITSELF
Case study: the review request that sends itself
Type: deterministic automation, live since 25 July 2026.
The problem.
The best moment to ask for a Google review is shortly after a job is finished — and it’s exactly the moment a busy technician forgets.
What we built.
When a job is marked complete, a review invitation queues automatically and sends from our own mail server about half an hour later — inside sociable hours only. The system tracks what it has asked so nobody is asked twice, and a heartbeat file proves every run happened, because “it ran and found nothing” and “it never ran” should never look the same.
Where the human stays in charge.
A person marks the job complete; the automation only handles the remembering.
Verifiable.
The reviews it gathers are on our Google profile — and on our reviews page, which only ever shows words verified against Google’s own records.
/04 — CASE STUDY: LIVE DATA FROM A CAMPER VAN, ON A PUBLIC WEBPAGE
Case study: live data from a camper van, on a public webpage
Type: systems integration with live data, live since 2 July 2026.
The problem.
Off-grid electrical systems — boats, vans, cabins — generate rich live data that usually stays locked inside an installer’s app.
What we built.
Our own camper van’s Victron energy system streams its live state — battery, solar, consumption — through a secure server-side connection onto public dashboard pages, with the access token held server-side where a visitor can never reach it, and a cache layer so a burst of visitors costs the van nothing. Victron accepted 365 Techies as a Recommended Software Integrator on the strength of this work.
Where the human stays in charge.
The integration is read-only by design — the public page can see the van; nothing on the internet can touch it.
Verifiable.
The dashboard is live on this site right now — watch the van think.
/05 — WHERE AI ITSELF EARNS ITS PLACE
Where AI itself earns its place
The systems above are mostly rules — and proud of it. AI enters where judgement or language is the job:
AI voice agents
answering calls in natural language, taking bookings and messages, handing anything unusual to a person. From £95/month.
Drafting and summarising
enquiry summaries, reply drafts, report wording — always with a person approving before anything consequential happens.
Our own engineering
much of the software described on this page was built with AI-assisted engineering, reviewed line by line by a person. We sell what we practise.
That split — rules where rules win, AI where AI wins, humans in charge of both — is the whole philosophy. It’s also why nothing on this page needed a disclaimer.
// SERVICE & BUILD
Monthly service — confirmed with your quote
One-off design and build — quoted by complexity. Every system on this page started exactly the way yours would: a problem described in plain English, then a scoped build.
// GOOD QUESTIONS
Frequently asked
Are these case studies real?
Yes — they are the systems running 365 Techies itself, with real ship dates from our own version-control history. You can test two of them from this website today: abandon a booking and see who calls you, or watch the live van dashboard.
Why are there no customer case studies yet?
Because we publish evidence, not anecdotes. Client AI projects will appear here once they’re launched and working, with the client’s permission and their measured results — not before. In a market where nobody publishes verifiable proof, we’d rather be slow and checkable than fast and vague.
Would my system be identical to these?
No — and be wary of anyone who says otherwise. These show the shape of our work: understand the workflow, automate the mechanical part, add AI only where it earns its place, and keep a person in charge of anything consequential. Your build is scoped around your workflow and quoted individually.
Got a job these remind you of?
Describe the repetitive work in plain English — we’ll tell you honestly whether it wants rules, AI, or a person, and quote the build around your answer.
01202 775566 · help@365techies.co.uk · MON–FRI 9AM–5PM