500 contracts a year, each typed into Word. Now: pick a client, pick a template, done.
MR3 Solutions, an outstaffing company, spent up to ten minutes on every contract and kept its business knowledge in binders. I built a system that drafts documents in under a minute — and an AI assistant that answers questions the binders used to.

This is your situation if…
- Your team produces the same kind of document again and again — contracts, offers, authorisations, invoices — by filling in a template by hand.
- Client details live in several places, so every document means cross-checking and the occasional wrong number.
- Finding out “what did we agree with this client two years ago?” means opening a folder or asking the one person who remembers.
The problem
One contract meant opening a Word template, typing in the client's details, checking them against two or three other sources, fixing the formatting and sending. Seven to ten minutes of work that needs no thinking — multiplied by more than 500 contracts a year.
The second problem was quieter. Business information lived in physical binders and scattered spreadsheets. Any specific detail — a past agreement, a promissory note, a client's history — meant physically searching for it. Two earlier attempts to automate the contracts had ended back in Word templates.
What I built
In the order it happened.
- 1Walked through every step of making a contract.
Where the data came from, who checked what, why the earlier attempts had failed. The answer: they automated the document, not the way the legal team actually worked.
- 2Client data entered once, reused everywhere.
One client record feeds every document type. No more cross-checking three sources.
- 3Templates the team maintains themselves.
A template editor with fill-in fields. When a clause changes, they change it — no developer, no waiting.
- 4Contracts and authorisations in one click.
Select a client, select a template, generate. Under sixty seconds. Promissory notes tracked in the same place.
- 5An assistant that knows the business.
An AI chat with access to the whole database. “What did we agree with client X?” gets an answer in seconds instead of a trip to the binder.

The AI assistant answering from the client database.
For technical readers: how it's built
Web app built with Lovable and Claude Code on a Supabase database; document generation and the AI assistant's data access orchestrated in n8n; template variables resolved at generation time. [confirm details]
What changed
Measured [X months] after launch.
- Duration:
- [X weeks]
- Investment:
- [price range]
- Team on my side:
- just me
- Runs in:
- MR3's own accounts [confirm]
“I'd tried to automate our contracts twice before with other consultants and both times we ended up back in Word templates. Stefan's system drafts them in under a minute and actually integrates with how our legal team already works. That last part is what the others missed.”
Petar VučkovićDirector, MR3 Solutions
Still filling in templates by hand? Let's talk.
30 minutes, free, no commitment. You talk to me from the first call to delivery. Or write to iveticpf@gmail.com.

