01 — Services
AI Automation & Agents
Automation that takes repetitive, manual work off your team, built as narrow jobs with a clear start and end. Every run is logged and reported, and a person approves what matters.
We start with an audit, not a build. Our published proof is in automation: a nightly ERP-to-shop sync that runs unattended and emails a report every morning. The agent pipelines we build for clients are the ones we first run on our own work, every week.
What you get
- 01
Automation audit — a written map of what to automate first, what to leave alone, and why. Fixed price, agreed up front.
- 02
Agent pipelines with human approval — agents draft, prepare and sort; a person approves before anything is published or sent.
- 03
Shop, ERP and feed automation — product data kept in step on a schedule, without manual exports.
- 04
Enquiry assistants — on your site, WhatsApp or Telegram; they answer common questions, qualify the request and hand over to a person. AI disclosure is built in.
- 05
Monitoring, logs and a report after every run — you see what ran, what changed and what failed, without asking.
- 06
Handover, documentation and support — written so your team knows what runs, when, and what to do if it stops.
Why it works
Narrow jobs, not a robot employee
We don't promise an AI that replaces a person. We automate one clearly defined task at a time — a sync, a draft, a first reply — because narrow jobs can be tested, measured and trusted.
A person approves what matters
Agents prepare; people decide. Anything that reaches a customer or the public waits for a named person's approval, and an assistant always says it is an AI.
Proof after every run
Logs, reports and tests instead of promises. It is the method behind our Semplo Home integration: every nightly run is recorded and reported, and every fix was tested before release, so nobody has to check by hand that it worked.
Related work
Questions
Straight answers
What is an AI agent, and how is it different from ordinary automation?
Ordinary automation follows fixed rules: when this happens, do that. It is the right tool when the steps never change, such as a nightly product sync. An AI agent uses a language model for the steps that need reading, writing or judgement — drafting a post, understanding an enquiry, summarising a document. Most real systems mix the two, and we use plain automation wherever it is enough, because it is cheaper and more predictable.
What happens when it gets something wrong?
It will, sometimes. Language models make mistakes, so we design for that from the start. Anything that reaches a customer or the public waits for a person's approval. Every run is logged, so you can see what was done and why, and a failed step stops and is reported instead of passing bad data on.
Who sees our data?
We agree that with you before anything is built. The audit lists which data each automation touches and which outside services process it — usually the AI model provider and the systems you already use. We send only the fields a job needs, and access stays limited to the people working on your project.
Does a chatbot have to say it is an AI?
Yes. The EU AI Act's transparency rules have applied since 2 August 2026: people must be told when they are talking to an AI system, unless that is already obvious. Every assistant we build says so at the start of the conversation. This is a plain-language summary, not legal advice.
What does it cost to run?
Running costs are usage-based: AI model providers charge by the amount of text processed, so the cost follows the volume of work. We estimate it for your case in the audit, before you commit to a build.
How do we start?
With the audit. We look at how the work is done today and give you a written map of what is worth automating first, what is not, and what each step would involve. The audit has a fixed price, agreed up front.
