Contacts
+1 234 567 890
Follow us:
Get in Touch
Close

Contacts

Ahmedabad, India

+ (123) 1800-234-5678

info@theaidivision.com

Call us: +1 234 567 890
Follow us:
Get in Touch
Get in Touch

SpendOpt: Autonomous Procurement Negotiation Multi-Agent System

Projects

Brief and project description

A multi-agent system that autonomously negotiates discounts on long-tail vendor renewals, driving direct bottom-line savings..

SpendOpt is an autonomous procurement negotiator focused on the "Long-Tail" of vendor spend. While large contracts get human attention, thousands of small SaaS and service renewals often get auto-paid without question. SpendOpt changes that. This agentic swarm reads renewal invoices, initiates email negotiations with vendors, and utilizes proven negotiation tactics to secure discounts—impacting the bottom line without human effort.

Let's Connect

The Challenge

A Logistics Enterprise had an uncontrolled “SaaS Sprawl” problem.

  • The Long-Tail Problem: They had 400+ small vendors (software, subscriptions, maintenance) costing less than $10k each. The procurement team didn’t have the bandwidth to negotiate them, so they just accepted the 5-10% annual price hikes.

  • Auto-Renewal Waste: Many contracts auto-renewed for unwanted services simply because no one caught the deadline in time to cancel or renegotiate.

  • Invisible Inflation: These small increases aggregated to over $2M in unnecessary annual spend inflation, eating into the company’s net margin.

The long-tail of procurement is where margins go to die. SpendOpt shines a light on the invisible inflation of auto-renewals, turning neglected invoices into a newfound source of capital efficiency.

Procurement Lead, The AI Division
TellerSense AI Voice Agent The AI Division Infographic

The Solution

The AI Division deployed an Autonomous Negotiation Multi-Agent System to manage this tail spend proactively.

The Workflow Breakdown

1. In-Flow Analysis:
The agent monitors the Accounts Payable inbox for renewal notices. It extracts the renewal date, price increase, and usage data. It checks against the employee database: “Is anyone actually using this tool?”

2. Strategic Negotiation:
If the tool is needed, the agent drafts and sends a negotiation email to the vendor. It uses dynamic strategies based on the context (e.g., “Budget cuts require us to freeze spend, or we will have to churn,” or “We are considering a competitor”). It understands the vendor’s reply and counters effectively.

3. Approval & Payment:
Once a discount is agreed upon (or if the vendor refuses but the tool is critical), it summarizes the negotiation outcome and presents it to the Finance Manager for a one-click “Approve Payment.”

Technology Stack

  • Agent Framework: LangGraph (Stateful multi-turn negotiation logic)

  • Communication: Gmail API

  • Financial Data: QuickBooks / Netsuite API

  • LLM Logic: Claude 3.5 Sonnet (Selected for its excellent nuance in polite, yet firm persuasive writing)

The Results

  • $450k Annual Savings: In the first year, the agent secured an average discount of 12% across 300+ small contracts.

  • Zero Missed Cancellations: The agent flagged unused software before the renewal deadline, allowing the company to cancel waste proactively.

  • Procurement Bandwidth: The human team could focus entirely on million-dollar strategic contracts, while the AI handled the massive volume of small operational spend.

Annual Savings
0 k
Missed Cancellations
0 %
Subscribe