Industry: Information Technology & Services
Headquarters: Rajkot, India
Challenge: Manual, chat-based procurement leading to missed deadlines, expired quotes, and payment delays.
Solution: An Agentic AI Purchase Manager integrated into Telegram to automate the RFQ-to-Payment lifecycle.
The Challenge
A Procurement Process Stuck in Chat
For a rapidly growing mid-sized enterprise, the purchasing division was entirely dependent on manual communication. The procurement team managed the entire lifecycle—from Request for Quotation (RFQ) to vendor payment—via Telegram chats and email threads.
While chat-based communication was fast, it lacked structure. The organization faced critical bottlenecks:
- Inefficient Sourcing: Buyers manually messaged vendors, often including out-of-scope suppliers in RFQs, leading to irrelevant quotes.
- Approval Bottlenecks: Directors were constantly interrupted for quote approvals, often lacking context on budget impact or quote validity periods.
- Missed Deadlines: Quotes expired due to slow internal approvals, forcing the company to renegotiate at higher prices.
- Payment Friction: Proof of delivery was frequently lost in chat history, delaying vendor payments and damaging supplier relationships.
The company needed a solution that could operate within their existing communication channels without forcing vendors or employees to learn a complex new ERP system.
The Solution: An Agentic AI “Purchase Manager”
The enterprise deployed a multi-agent AI system designed to function as an autonomous procurement manager. Operating entirely within the company’s Telegram interface, the AI utilizes specialized sub-agents (Sourcing, Negotiation, Finance, and Compliance) to execute the entire procurement lifecycle.
The AI operates on a “perception-action” loop, making decisions, communicating with vendors, and enforcing company policy without human intervention.
The 6-Step Autonomous Workflow
1. Context-Aware RFQ Generation
Unlike a mass broadcast, the AI maintains a categorized vendor database. When a requirement is raised (e.g., “Need a load balancer/router for the Mumbai office”), the AI identifies the correct category and sends RFQs only to relevant vendors.
2. Intelligent Quote Comparison & Deadline Tracking
The AI acts as a procurement analyst, extracting data from vendor PDFs. It calculates total costs including taxes, flags scope differences (Supply vs. Install), and—crucially—tracks validity periods. It presents the Director with a concise comparison table, explicitly warning if a quote has a short validity (e.g., “2-day validity” vs. “15-day validity”).
3. Human-in-the-Loop Approval
The AI does not make final financial decisions. It presents the summarized options to the Director via chat. In the example, the Director simply replied, “Budget option is fine.” The AI immediately understood this to mean the Naitri Infotech quote (₹6,195 all-in), demonstrating advanced natural language understanding.
4. Autonomous Negotiation & Terms Enforcement
The AI initiates the next phase automatically. It informed the vendor to proceed but immediately enforced the Director’s new directive: “Payment after completion.” It negotiated the timeline and coordinated the site visit without human intervention.
5. Delivery Verification & Proof of Work
The AI tracks the expected delivery date. Once the vendor confirms completion, the AI requests manual verification of the “Proof of Delivery” (a necessary human check for physical goods).
6. Automated Accounts Payable & Vendor Communication
Once the delivery is manually verified, the AI triggers the final workflow. It passes the invoice to the Accounts department and informs the vendor that payment is being processed. It also logs the response time for the daily check-in, ensuring no vendor is left waiting.
The Results
70%
reduction in procurement cycle time
90%
reduction in Director approval time
100%
compliance with vendor payment terms
Within the first six months of deployment, the Agentic AI transformed the purchasing division’s KPIs. The company eliminated manual follow-ups, ensured 100% compliance with payment terms, and drastically reduced the time spent by leadership on procurement approvals.
Key Wins for the Business
- The “2-Day Validity” Save: In the screenshot example, the AI highlighted that the budget quote was only valid for 2 days. A human manager might have missed this, leading to a re-quote and price increase. The AI moved immediately, locking in the price.
- Policy Enforcement: The AI strictly adhered to the Director’s instruction to change payment terms to “Payment after completion,” overriding the vendor’s initial request for a 50% advance.
- Audit Trail: Every interaction—from RFQ to final payment—is logged in the chat, creating a perfect audit trail for finance and compliance.
Performance Metrics
| Metric | Before AI | After AI | Improvement |
|---|---|---|---|
| Procurement Cycle Time | 5–7 Days | 1–2 Days | 70% Faster |
| Quote Validity Compliance | 65% (Quotes often expired) | 100% | Zero Expired Quotes |
| Director Time Spent | 10+ Hours/Week | < 1 Hour/Week | 90% Reduction |
| Payment Processing | Manual reminders, high error rate | Fully automated post-delivery | Zero Late Payments |
| Vendor Negotiation | Ad-hoc | Structured & Policy-Driven | Higher Margins |
Beyond Automation: A Strategic Shift
By deploying an Agentic AI Purchase Manager, the enterprise did not just digitize a form; they delegated an entire job function. The AI handles the repetitive, high-stakes coordination of sourcing, negotiation, and follow-up, allowing human directors to focus solely on strategic decisions and final approvals. This “Chat-First” automation approach required zero training for vendors and minimal change management for internal staff, proving that the future of procurement is not a complex ERP—it is an intelligent agent in your pocket.
