AI adoption for regulated firms

AI your firm can actually put in front of client data.

Operisys helps regulated and trust-dependent firms — accountants, solicitors, recruiters, brokers, healthcare practices — automate the repetitive work without putting confidential client information somewhere it shouldn't be. Built and documented by one person who does both sides: the automation and the data protection.

£1,450 · two days

The problem

The professional work is strong. The operation around it is running on inboxes and memory.

Most professional firms don't lose clients on the quality of the work. They lose them in the gaps: slow first response, documents that arrive late, and no clear view of where each job actually sits.

Failure 01

Enquiries go cold

New enquiries arrive by phone, email, web form and WhatsApp — and sit unanswered while a competing firm replies first.

Failure 02

Documents chased by hand

Staff request the same ID, statement or signed letter three times, job by job, message by message.

Failure 03

Nobody can see the job

“Where is this right now?” takes three systems and a colleague to answer, and deadlines live in people's heads.

Failure 04

Follow-up depends on memory

Status updates and reminders happen when someone remembers — not when the client or the deadline needs them.

Failure 05

Reporting is rebuilt monthly

Partners reconstruct workload, revenue and bottlenecks by hand in spreadsheets instead of seeing them live.

Failure 06

AI is used without control

Staff quietly paste client detail into chatbots with no boundaries, no review, and no audit trail — a real compliance risk.

The engagement£1,450 · two days
Start here

AI Readiness and First Automation.

Two days, one fixed price, and something working at the end of it — not a report recommending that you buy something.

See what the two days include
Day one

I map how work actually moves through your firm and identify the three highest-volume repetitive tasks worth automating.

Day two

I build one of them, working, in your tools.

You also get a written note covering what data the automation touches, where it goes, the lawful basis, and how long it's kept — so it stands up if a client or a regulator asks.

  • No ongoing commitment
  • No platform to buy
  • Nothing worth automating? I'll tell you on day one and invoice half
Output

What day two actually produces.

Illustrative

Enquiry intake

Enquiries arriving by web form, email and WhatsApp land in one place, deduplicated, tagged by service line, with an acknowledgement out inside two minutes. The partner sees an unanswered-enquiry count each morning instead of finding out a week later.

Illustrative

Document chasing

The system knows which documents each open job is missing, asks the client for them, chases on a schedule, and stops chasing the moment the file arrives. Nobody asks for the same statement three times.

Illustrative

The data note

A two-page written record for the automation: what personal data it touches, which systems it moves through, the lawful basis relied on, retention period, and who reviews what. Handed over on day two, ready to show a client or the ICO.

These are illustrations, not client work. Operisys is new, and named client examples will appear here as engagements complete.

See what the two days include
Safe AI use

AI that a regulated practice can actually stand behind.

An impressive demo is not the goal — a system you can trust under pressure is. That is why the automation here is deliberately bounded.

Boundary 01

AI suggests. People decide.

The automation drafts, classifies, and summarises. Anything touching advice, merits, or a client outcome is routed to a person — never answered by a model.

Boundary 02

Every action is reviewable.

Material outbound messages and actions can be held for human approval, with changes and decisions recorded against the relevant job.

Boundary 03

Clear boundaries, on purpose.

What AI may and may not do in your firm is agreed in writing before anything goes live — so it holds up under an audit, not just in a demo.

Firms that want ongoing help after the two days usually move to governance support, further automation, or a managed arrangement. That conversation happens after something is working, not before. See all services

Who you deal with

One person, doing both sides of the job.

Ashkan Gholizadeh, Founder, Operisys
Ashkan GholizadehLinkedIn

Operisys is run by Ashkan — a software engineer with a law background. An LLB, an LLM in international commercial law, and years building systems for environments where the data can't leak and the process has to hold up afterwards. You deal with the person doing the work. Where a job needs a specialist, I bring one in and say so.

Meet Operisys

Software engineering

I build the automation myself — intake, integrations, reporting — rather than stopping at a recommendation.

Law background

An LLB and an LLM in international commercial law. Risk and careful language are part of the design from the start, not a review at the end.

Data protection

Every automation ships with a written note: what data it touches, where it goes, lawful basis, retention.

Systems that hold up

Years building for environments where the data can't leak and the process has to stand up afterwards.

Questions

The things worth asking before you book.

Is £1,450 the whole cost?
Yes, for the two days. There's no platform to buy and no retainer. If you want further work afterwards it's quoted separately, and you'll have seen the standard of the work before deciding.
What if you find nothing worth automating?
I'll tell you at the end of day one and invoice half. That's happened in consultancy far less often than it should — most firms have at least one job that shouldn't be done by a person.
Who sees our client data?
Only me, and only what's needed to build the thing. Every engagement ships with a written note recording exactly what data the automation touches, where it goes, and how long it's kept. Client data is never used to train anything.
We already use ChatGPT. Isn't this the same?
Staff pasting client detail into a chatbot is precisely the problem. The difference is boundaries, an audit trail, and a written record you can show a client or a regulator. The tool isn't the risk; the absence of a documented process is.
Do you need access to our systems?
Usually read access to one or two, agreed in writing beforehand, and revoked at the end unless you want the automation to keep running. You can watch every step if you'd prefer.
What happens if it breaks after you've gone?
You own the automation and the documentation. It runs in your tools, on your accounts. If you want me to maintain it, that's a separate arrangement — not a condition of the two days.

Find the repetitive work worth automating first.

Twenty minutes is usually enough to tell whether there is anything here worth doing. If there isn't, I'll say so.