We answer
the questions people actually ask.

An AI employee holds one job in your business: finance, compliance, communications or operations. The role works to a job description and a weekly rhythm, and commits to named outputs, the way a member of staff does. We build the role from up to 10 of the 59 skills in our library, then connect the role to Microsoft 365, Xero, Pipedrive and SharePoint. A named person at Struan signs the weekly report.

ChatGPT and Copilot wait for your prompt, then hand the work back for you to finish. An AI employee does the job without being asked. The AI employee posts to Xero, chases references and sorts the inbox inside your own systems, under the approval rules you set. A named person at Struan signs the weekly report.

Cost follows the role, so we give you a number on the call rather than from a price list. One written-up deployment gives you a benchmark: a finance assistant that replaced a part-time hire ran at roughly 74% below the cost of that hire, per the signed proposal. Bespoke work can start smaller, with a small paid proof-of-concept first. On the call we scope your role and give you a real number.

Fifteen working days, from signature to live. We move that fast because the 59 skills in our library are already written and already proven on live client work. We agree the job description, pick the skills the role needs, connect your tools, then go live. Your first weekly report arrives inside week one.

Your data stays yours. We hold and process your production data in the UK or EU, in the region you agree. Where your tools allow, your AI employee works inside your own Microsoft or Google account. We record the region, the access and the stop procedure in the data agreement, before any work begins.

An approval gate catches the mistake before the mistake reaches you. Work you want eyes on stops at a gate you set and waits for your yes. Any expense over £500, say, waits there until a person approves the spend. A named person at Struan then reads and signs every week's output. When something does slip through, we catch the work and fix the work at our cost.

An AI employee takes on the repeating work, not your people. Most deployments cover a hire the business had been putting off, or win back hours lost to admin. The managing director stops working the finance inbox. The compliance team stops chasing references. Your people spend the recovered hours on customers and decisions.

You already buy IT this way. You do not run your own server room, because a provider runs the server room for you and stands behind the result. A managed AI service does the same for the work. The AI employee holds the role. We monitor the role, maintain the role and sign off the output. You buy a finished result, with a person to call.

You take on a finance, compliance or operations role for a lot less than the equivalent hire, and you carry none of the recruitment risk. The price works because we build the role from skills that are already written and already proven, rather than starting from scratch for one client. Start with one role. Bespoke work can start with a small paid proof-of-concept.

The library holds the 59 pieces of work we build a role from. One skill reads a receipt and works out the VAT class. Another posts straight into Xero, chases five years of references, or spots a gap in an employment history. Every role uses up to 10 skills, tuned to your rules. A skill written for one client joins the library for every client after.

Yes. Your AI employee works inside the systems you already run: Microsoft 365 and SharePoint, Xero, Pipedrive and the rest. We connect the role during the build, with access limited to the role and removable at any time. Your team gets nothing new to roll out and no new login to remember.

Look for work that repeats, runs to a regular rhythm and moves a lot of documents. Look for hours your people would rather spend on judgement. The sense-check call takes fifteen minutes, tests the fit and names the first jobs worth doing. We say no plainly when the role does not fit.

No. We build your AI employee for your business alone. The AI employee works from a dedicated email address and a dedicated Microsoft Teams identity inside your organisation, and we tune every skill to your rules. What your AI employee learns in your business is never shared with another client, and never sold.

A scoped role is the job description we work to, agreed in writing before any work starts. The job description names four things: the work the role covers, the rhythm the work runs to, the outputs and where each output goes, and the points where work stops for your decision. You agree all four before you sign, and every weekly report is measured against the same four.

A desk-based role tops out at around 95% of the work, and never reaches 100%. We set that ceiling as policy. A role with site visits, physical work or in-person judgement sits well below. Some decisions stay with a person for good: financial approvals, contract signing, hiring and firing, crisis response, and the call to go live.

The work stops and waits. Any action past a threshold you set holds at the gate until you decide. You approve, amend or reject in the Microsoft Teams or Google Workspace chat you already use. The audit trail records the stop, your decision and the outcome, and the weekly report carries the same record. Email gets one more protection: a person releases every message before the message reaches a client or supplier.