AI adoption

Two free hours of diagnosis to find the one or two workflows worth doing; knowledge base, reception and automation added one at a time.

Highlights

Three entry points: cover, upgrade, keep

AI does not have to be a big investment — pick the one or two workflows that hurt most.

Cover the gap

Data entry, late-night queries, report reconciliation — the dull, repetitive work goes to AI, not to overtime.

Upgrade the team

AI takes the first line and the repetition; exceptions and judgement stay with people. The same team, more output.

Keep the knowledge

Documents, SOPs and quote logic go into a knowledge base — it stays when a veteran retires, and new staff can ask.

Process

From diagnosis to go-live

Four steps, each with a deliverable. You decide the next step only after the last one lands.

  1. Free diagnosis (half day)

    On site or by video: we watch how the work flows and name two or three processes worth changing.

  2. Written scope

    What we do, which data is used, who owns it, how it is accepted.

  3. Pilot, 4–6 weeks

    One workflow goes live first; we measure time saved and errors made.

  4. Go-live and maintenance

    Expand only once it is steady; from then on it sits inside your IT support, same contact.

What it covers

Six things AI can take over

No one needs all six. The diagnosis picks the one or two that hurt most; the rest wait.

  • Knowledge base

    Documents, SOPs, contract terms and quote logic answered in one question — with sources you can open.

  • Reception

    Prices, stock and opening hours answered first; negotiation or complaints go to a person with the conversation summary.

  • Documents and reports

    Invoices, delivery notes and reconciliations read and checked by AI instead of retyped by hand.

  • Quotes and enquiries

    A first draft built from your own quote logic, for a colleague to check once.

  • Meetings and notes

    Recordings turned into text with actions and owners pulled out — notes exist before people leave the room.

  • Legacy data

    Messy exports from an old system turned into a knowledge base you can ask questions of.

Boundaries

Three rules, agreed before we start

AI fails in two ways: confidently inventing an answer, and confidential data going where it should not. Both are handled from day one.

  • If it is not there, it says so

    Answers come only from your documents. When the answer is missing, it says so instead of inventing something plausible.

  • Data stays where you choose

    Your server or your cloud; which documents may enter is written down before work starts.

  • Who sees what

    Permissions by role — finance, HR and management see different things, and changes are logged.

Pricing

AI packages: start with a free diagnosis

The first two hours are free, then MOP 500 an hour. Delivery is quoted on scope and starts with a pilot.

AI adoption

Starts with two hours of diagnosis: find the one or two workflows worth doing, then talk projects and price.

AI project

quoted per scope

knowledge base / reception / document and report automation · 4–6 week pilot first

No big-bang replacement: one workflow first, expanded only once it measures useful.

  • Company knowledge base: documents, SOPs, quote logic, answers with sources
  • AI reception: repeated questions handled first, complex ones passed to a person with a summary
  • Invoice, report and reconciliation automation
  • Data stays where you choose (your server or your cloud)
  • Permissions by role, written down
  • A four to six week pilot before deciding next steps

See all prices

A good fit when

Documents are scattered and new staff keep asking the same questions

Repeated enquiries eat the team day

There is old-system data you want to query again

You want to try AI without a big commitment

Book a free
IT health check

A short call: is the IT reliable, how are backups and permissions handled, and how fast would you want someone to answer in an emergency?