AI agents on top of your LIMS that answer customer enquiries, prepare accurate quotations and draft follow-ups. Your LIMS stays the source of truth. Your staff approve what matters.
Web-chat pilot first · WhatsApp next · deeper LIMS automation after
CustomerNeed pesticide residue testing for 2 samples of basmati rice. How much?
✓ LIMS: service available · method on file · TAT 5 working days
CosmIQ agentSure. Is this for export or domestic compliance? And roughly how much sample can you send?
CustomerExport. 500 g each.
01The problem
Front-desk teams spend hours retyping enquiries, looking up price lists and chasing missing sample details, and past customers rarely hear from you again.
Requests arrive at all hours across personal and shared WhatsApp numbers, with no record of what was asked or answered.
Staff check services, methods, TAT and prices by hand. Every quote is slow, and mistakes cost money.
Order history sits in the LIMS, but nobody has time to suggest the next relevant test.
02Two agents
Both agents read from your LIMS and work through WhatsApp. Your laboratory staff stay in control.
03Enquiry → quotation
When a customer writes in, the agent follows a controlled sequence. Service availability, requirements and prices are always fetched from your LIMS.
A customer asks for a test in their own words, any time of day.
Name, phone number and GST number are checked against customer records.
Availability comes from the service catalogue, never from AI guesswork.
The customer confirms the selected services, sample type and sample count.
The agent asks where the approved quotation should be sent.
A fixed-template quotation uses calculated prices and tax. Staff review it before emailing.
04Marketing
Runs on a schedule your administrator configures. Nothing reaches a customer without approval.
Customer and previous-order history from LIMS, on a daily or weekly schedule.
Matches past orders to relevant tests your lab actually offers.
Personalised WhatsApp promotions, written for each customer.
Nothing is sent until an administrator reviews and approves each draft.
Approved messages go out; the agent learns from the edits your admin makes.
05Guardrails
The AI never invents services, prices, test methods, sample requirements or turnaround times.
The language model handles language. Everything that has to be correct is handled by your data and deterministic code.
| Role | Responsibility |
|---|---|
| AI understands | Interprets the customer's message, extracts information, writes natural follow-up questions and drafts. |
| Workflow controls | A controlled state machine decides each step. The AI does not run the process on its own. |
| LIMS verifies | Services, parameters, sample requirements, turnaround times and prices are always read from your LIMS. |
| Backend calculates | Prices, tax and totals are computed by business logic, never by the AI. |
| Quotation engine generates | Quotations are produced from a fixed template and stored for tracking. |
06How it connects
Nothing to replace. CosmIQ talks to customers on WhatsApp and to your LIMS through its API. Your team approves what goes out.
Under the hood: Next.js · FastAPI · PostgreSQL · OpenAI / Gemini · Vercel · Railway
07Pilot
We prove the full enquiry → information collection → quotation loop before touching live customer channels.
Run the complete enquiry-to-quotation workflow in a web chat, against your real service data.
Move the proven workflow onto your WhatsApp Business number.
Order creation, marketing schedules and more workflows, once you're confident.
08FAQ
No. Prices, tax and totals are calculated by backend business logic from your LIMS or service database. The AI only handles the conversation.
The agent only offers services it finds in your LIMS, so it won't invent one. The enquiry is still saved, so your team can follow up.
Marketing messages never go out until an administrator approves each draft. Quotations are built from your fixed template using verified LIMS data.
No. Staff keep working in your LIMS. Enquiries are saved and orders are created there where needed.
We connect through your LIMS API and agree exactly which data (services, prices, customers, orders) the agents read and write. No API yet? Tell us; the pilot can start from your service and price database.
Access is limited to what the pilot needs and follows your lab's privacy policy. Agent actions are logged for review.
09Request a pilot
Tell us about your lab. We'll reply with a pilot plan scoped to your LIMS and services.