How a small Indian manufacturer can automate enquiry-to-quotation with AI agents

The enquiry-to-quotation bottleneck

Most small manufacturers in India run the same loop: a WhatsApp message or email arrives with a vague request, "need 500 pcs of M8 bolts, SS304, delivered to Pune." Someone reads it, checks stock in a spreadsheet or ERP, asks the shop floor about capacity, calculates material cost plus conversion, adds freight and GST, types a quote in Word or Excel, emails it back, then follows up three days later. Multiply this by twenty enquiries a day and the owner or sales person spends half their time on paperwork instead of production.

What changes when the loop runs on your own machine

adris.tech is a free Windows desktop app that brings a team of AI agents and the tools they need into one window. It runs locally, uses your own AI keys (Groq, NVIDIA, Gemini, OpenAI, Anthropic, or a local model), and costs nothing. No cloud subscription, no data leaving your premises. You install it, add your keys, and the agents start working.

For enquiry-to-quotation, three modules matter most: Krew (specialist agents), Automation (visual workflows), and Vault (DNS-level blocking so the machine stays clean).

Step 1: Capture every enquiry in one place

Create a Mail Watcher agent in Krew. Point it at your sales inbox (IMAP) or the WhatsApp Business API webhook. The agent reads each message, extracts: part number or drawing reference, quantity, material grade, delivery pincode, and any special terms (packing, certification, payment). It writes a structured row into a local SQLite table, no cloud database needed.

Step 2: Enrich with internal data

Add a Stock Checker agent. Give it read access to your ERP export (CSV, Excel, or direct DB if you allow). The agent matches the enquiry line items against current stock, open purchase orders, and work-in-progress. It tags each line: in stock, can make in 3 days, needs raw material, out of capacity.

Step 3: Price it

A Pricing Engine agent takes the enriched row. It pulls your master price list, applies your markup rules (material + conversion + overhead + freight slab for the pincode + GST), and outputs a line-item quote. If a line is needs raw material, the agent fetches the latest vendor quote from your approved supplier sheet and adds lead time.

Step 4: Generate the quotation document

A Document Builder agent uses a Word template stored locally. It fills in: buyer name, GSTIN, line items, totals, terms, validity, and your bank details. Saves the file as PDF in a dated folder on your machine. No manual copy-paste.

Step 5: Send and track

An Outreach agent emails the PDF from your Outlook or Gmail account (SMTP) with a standard covering note. It logs the sent date, quote number, and expiry in the same SQLite table. A Follow-up agent wakes up on day 3 and day 7: if the quote status is still open, it sends a polite nudge and notifies you on the desktop.

Wiring it together in Automation

Open the Automation tab. Drag the five agents onto the canvas. Connect: Mail Watcher → Stock Checker → Pricing Engine → Document Builder → Outreach → Follow-up. Set the trigger: new row in enquiry table. Hit Run. The workflow executes end-to-end on your PC. You see each step's output in the log pane. If something looks off, say, a freight slab missing, you edit the Pricing Engine's rule sheet (a simple CSV) and re-run just that node.

Keeping data private

All of this runs inside adris. The Vault module blocks telemetry, tracker, and malware domains at the DNS level, so the machine only talks to the AI endpoints you configured (Groq, local Ollama, etc.). Your enquiry data, pricing logic, and customer list never leave the shop floor.

What the owner actually does

Reviews the daily summary email at 9 AM: "12 quotes sent, 3 pending raw material, 1 capacity clash."
Opens the adris window, clicks the capacity clash, sees the shop floor load chart, decides to outsource or reschedule.
Updates the vendor price CSV when a supplier revises rates.
Adds a new markup rule in the Pricing Engine when GST slabs change.

No coding. No new hires. The agents handle the repetitive reading, matching, calculating, formatting, and chasing. You handle the exceptions.

Start small, expand later

Begin with one product line, say, fasteners. Get the workflow stable for two weeks. Then clone the Automation canvas for fittings, then for fabricated assemblies. Each canvas reuses the same agents; only the rule sheets and templates differ.

When the standard flow isn't enough

If your process needs a custom integration, pulling BOMs from SolidWorks, pushing quotes into Tally, checking vendor portals for live stock, that is a custom build. We scope it in writing, build the agents and workflows inside your adris install, and hand over the source.

Ready to automate your enquiry-to-quotation flow? Request a custom build at https://www.adris.tech/contact?type=custom_build