AI · ASSISTED WORK

Use AI for repeat work, with person still in control.

Draft, sort, search, and prepare reports from trusted information. Human checks result before it leaves system.

HUMAN-APPROVED AI

Assistant prepares.
Person decides.

Enquiry arrivesDraft reply
Trusted filesSource shown
Human reviewSend or edit
CONTROL STAYS VISIBLE
DraftSortSearchSummariseReview

Good AI job is narrow and repeated.

Assistant should remove preparation time, not hide decision.

01

Draft repeated replies

Prepare useful response from approved facts and current enquiry.

02

Sort incoming enquiries

Group, label, and route work so right person sees it sooner.

03

Search trusted documents

Find answer and show source instead of guessing from open internet.

Useful because control stays visible.

Keep model inside clear boundaries and make review easy.

01

Use approved information

Decide which pages, files, and rules assistant may use.

02

Keep task narrow

One repeat job. One clear output. No vague “run business” promise.

03

Show source and draft

Person can see what answer used and change it before action.

04

Measure saved work

Check fewer clicks, less typing, faster response, or fewer missed items.

Fits where people already repeat same work.

Website enquiries, support, trusted knowledge, and daily operations.

Website enquiries prepared from services, pricing rules, and next steps.

Customer support drafts with human approval before sending.

Internal knowledge search with exact source attached.

Daily summaries prepared from clean operational data.

AI belongs here when it prepares work for review.

Enquiry Inbox is the primary AI example. It is not repeated as a SITE, SYSTEM, or APP example.

AI · OUR OWN ASSISTED-WORK DEMO

Enquiry Inbox

WhatsApp, Instagram, and phone enquiries enter one queue. AI prepares a reply. A person approves, edits, or rejects it.

Fictional bakery demo opening on who still owes you, with one-tap payment reminders and weekly order slots

Questions worth asking before AI touches customers.

Most of them are about control. That is the right thing to worry about.

01

What does it cost?

The Social Inbox package is ₹4,999 to set up and ₹1,199 a month, for messages arriving on their own from Instagram and your Facebook page. Adding WhatsApp is ₹6,999 to set up and ₹1,799 a month, because WhatsApp is a separate setup that Meta has to approve and charges for. Either way it covers up to 100 prepared replies a month. Larger AI work sits inside a business system and starts at ₹40,000. Messaging and AI provider charges are separate, agreed in writing first, with a spending limit set before any real customer message is handled.

02

Will a machine be replying to my customers?

No. It writes a draft and stops. A person on your side reads it, edits it if needed, and presses send. Nothing reaches a customer without someone choosing to send it.

03

What if it writes something wrong?

Then you change it or throw it away, and nobody outside your business ever saw it. That is exactly why approval comes before sending, not after.

04

Where does it get its answers from?

Only from what you give it: your prices, timings, services, and the way you already answer people. It is not left to invent from general knowledge.

05

Do I need WhatsApp connected for this?

Only if you want messages to arrive on their own. It works from the start by pasting a message in by hand, which is often enough to prove it is worth connecting anything.

06

Are there running costs beyond your fee?

There can be. Messaging providers and the AI itself charge for use. Those are passed on as they are, agreed in writing first, with a spending limit set before a single real customer message is handled.

Show what team keeps typing, sorting, or searching.

Send one repeated example. We map smallest safe assistant.