What it costs
Prices on the page, like I'd tell you to.
I spend the whole of this site arguing that hiding your prices costs you customers. It would be a strange place to hide mine. Build it up below and the number moves as you go.
How big is it
What else it should do
Then, every month — optional, cancel whenever
To build it
₹47,000
Half to start, half on handover
Then per month
₹18,000
Cancel any month. Ad spend billed by the platform, not by me.
- Website build₹35,000
- AI booking assistant+₹12,000
It arrives itemised, so neither of us has to reconstruct it later.
What that number does and doesn't cover.
In every build
- Domain, hosting and SSL set up in your name
- Google Business Profile rebuilt and wired to the site
- Schema, sitemap and metadata so you are indexable from day one
- Loads in under two seconds on a phone, on 4G
- You own the domain, the site and the code
- A handover call and a written guide
Not included, and I'll say so up front
- Ad spend — that goes straight to Meta or Google, never through me
- Stock photography licences, if you want any
- Ongoing content once the retainer ends
- Anything needing a native app
These are starting numbers, not a quote. A busy salon with a rate card already written is a different job from a decor studio with four hundred photos and no prices anywhere — so the real figure comes after forty minutes on a call, and it is fixed before anything starts.
Beyond the assistant
The AI work doesn't have a price list.
A website and a booking assistant are known quantities, which is why they carry a number above. Retrieval over your own documents, or automating something your team currently does by hand, is scoped to the problem — and a figure invented on a web page would be a guess I'd have to walk back.
So it works the other way round: a call, then a written scope with a fixed price in it, before anything starts. If it isn't worth doing I'll say so on the call — that is cheaper for both of us than finding out in month two.
If you'd rather see the engineering before the conversation, the systems are written up here.
Answers out of your own documents
Contracts, manuals, policies, past quotes, years of email. Staff ask a question in plain language and get an answer with the source it came from, instead of filing a request and waiting two days.
A workflow you currently do by hand
The scoring, the form-filling, the drafting, the monthly report nobody wants to write. The rules stay in code so the numbers are reproducible; the model does the reading and the wording.
Agents with a boundary around them
Several narrow steps behind one orchestrator, each with a job it can be tested on, and a human approval gate in front of anything that writes. The model asks for data through named tools — it never gets handed the database.
Connecting it to what you already run
SQL Server, Postgres, MongoDB, a spreadsheet, your CRM. Exposed as typed, permissioned operations rather than free-form queries, so what it can and cannot reach is a fact rather than a promise.
A second opinion on an AI build
You have something half-working and want to know whether it is the retrieval, the prompt or the architecture. Usually it is the retrieval. A read of the system and a written list of what to fix first.