Sourcegent automates the three most complex workflows in enterprise procurement — RFP creation, supplier discovery, and evidence-based award scoring — using autonomous AI agents that operate under full human oversight at every consequential decision.
Built on Google Cloud Vertex AI with Anthropic Claude. Designed from the ground up for enterprises where every procurement decision must be faster, smarter, and explainable.
Built for direct procurement too — not just indirect spend
Most procurement AI starts with tail spend and services. ProgentAI’s architecture is designed for the full procurement universe. Sourcegent operates in two configurations: Indirect Spend for operational and services procurement, and Direct Spend for raw materials and components coupled to production demand signals. And as Supplygent matures, the demand plan itself triggers sourcing events automatically — closing the loop between what the business needs to produce and what procurement delivers.
If your procurement challenge involves production schedules, BOMs, or MRP-driven demand signals, we want to talk.
The Problem
Enterprise procurement carries the highest complexity and compliance burden in operations. And the tools have not kept up
In complex enterprise environments, every sourcing event must produce documented evidence of competitive process, supplier evaluation rationale, diversity outreach, and a complete audit trail that withstands scrutiny — from regulators, auditors, boards, or procurement committees. The consequences of a deficient process are not just internal — they can include disallowed costs, contract disputes, regulatory sanctions, and reputational damage.
Yet the tools available to procurement teams — ERP systems, spreadsheets, and email — were designed for transaction processing, not decision intelligence. The result is sourcing cycles measured in months, supplier pools drawn from familiar vendors rather than the full market, award decisions built on institutional memory rather than structured evidence, and audit packages assembled retrospectively under pressure.
Faster
RFP cycles that take weeks compressed to days with autonomous agent-driven sourcing
Smarter
Supplier discovery that searches the full open market, not just your existing vendor master
Explainable
Every decision documented with structured rationale — ready for audit, board review, or regulatory inquiry
Why It Works
ERP and procurement platform vendors will tell you they have AI too. They do — embedded assistants that respond when asked, recommend when prompted, and alert when rules are breached. That is assisted intelligence. It is not autonomous intelligence. The difference is architectural. ERP AI is stateless — it responds to a screen being opened or a query being run. It does not monitor. It does not initiate. It does not reason across supplier risk, market signals, demand forecasts, compliance obligations, and contract terms simultaneously and decide that a sourcing event needs to happen right now. ProgentAI was built for enterprises where every sourcing decision must be faster than the systems beneath it were designed to deliver, smarter than institutional memory, and explainable to whoever asks — a regulator, an auditor, a board, or a procurement committee. And unlike any ERP module or procurement platform, it gets smarter with every event — accumulating a proprietary intelligence graph that no system of record can replicate.
Faster by Design
RFP cycles that take procurement teams 8 to 12 weeks shrink to days when agents handle the drafting, supplier discovery, and evaluation in parallel. Sourcegent does not speed up the existing process — it replaces the manual steps that create the delay.
Smarter with Every Event
Every sourcing event makes the platform smarter. Award outcomes, supplier performance data, and open market signals accumulate into a proprietary intelligence graph that compounds over time. The more you use Sourcegent, the wider the gap between what it knows and what any directory, marketplace, or general-purpose AI tool can replicate.
Every Decision Explainable
Sourcegent generates structured rationale for every recommendation — why this supplier, why this score, why this award. The audit trail, evidence package, and human approval record are assembled automatically, not retrospectively. Every decision is explainable to whoever asks: a regulator, an auditor, a board, or a procurement committee.
The Hardest Problem, Solved First
ProgentAI chose to start at the most complex, highest-stakes point in the supply chain. The compliance architecture, audit trail design, and human oversight framework built for the hardest environment transfer to every other enterprise context — with configuration, not rebuilding.
Human-in-the-Loop
Autonomous intelligence. Human authority
Every consequential decision in Sourcegent requires explicit approval from a named, role-specific human before it takes effect. No RFP is distributed, no supplier is shortlisted, and no contract is awarded without human sign-off. This is not a safety layer added as an afterthought — it is a hard architectural constraint enforced at the platform level. The system cannot bypass it.
01
Before distribution
A named procurement approver reviews and authorizes every RFP before it reaches a single supplier
02
Before shortlisting
A named category owner approves the supplier shortlist before evaluation proceeds
03
Before award
A senior procurement leader personally signs off before any contract is awarded
The audit trail for every human approval — who approved, what they approved, when, and what evidence they reviewed — is preserved permanently with tamper-evident verification.
The Platform – Sourcegent
Sourcegent. Three agents.
One complete sourcing event
Sourcegent is an autonomous procurement intelligence layer that sits above your ERP and handles the three most burdensome workflows in enterprise sourcing — from trigger signal to signed contract — with a named human in control at every consequential step.
RFPgent — Autonomous RFP Engine
Generates complete, procurement-ready RFPs automatically — drawing on historical procurement data, market intelligence, should-cost models, and your organization’s evaluation framework — and stages them for human review before any supplier sees them.
- Full RFP package: specifications, evaluation criteria, SLA framework, supplier briefing
- Automated supplier outreach to both established and net-new market entrants
- Human approval required before any RFP reaches a supplier
Suppliergent — Supplier Discovery & Intelligence
Discovers, qualifies, enriches, and shortlists suppliers from both your existing vendor master and the full open market — including categories of suppliers your team may never have sourced before — with risk, financial health, and compliance screening built in.
- Queries structured registries, enrichment providers, and open web simultaneously
- Identifies net-new suppliers your procurement team has not previously engaged
- Human approval required before the shortlist is finalized
EBSgent — Evidence-Based Scoring & SLA Monitoring
Scores every bid across quantitative and qualitative dimensions with documented rationale, assembles the complete award evidence package, and monitors post-award supplier performance continuously against contractual commitments — automatically flagging corrective action when thresholds are breached.
- Structured evaluation scorecard with explainable, auditable rationale for every score
- Immutable audit log with tamper-evident verification — every decision preserved
- Human approval required before any contract is awarded
Platform Roadmap
From procurement intelligence to the autonomous supply chain operating system
Three platforms. Three years. One architecture that scales from the first sourcing event to full demand-and-supply planning intelligence — covering both direct and indirect procurement across the complete supply chain continuum.
Phase 01
Sourcegent
Autonomous sourcing and procurement intelligence. From RFP creation to contract award to ongoing supplier performance monitoring.
RFPgent · Suppliergent · EBSgent
In pilot · Raising seed
Phase 02
Purchasegent
Procure-to-pay and operations intelligence. Autonomous decision support for purchase orders, invoice anomaly detection, inbound delivery performance, and carrier management.
POgent · Invoicegent · Receivegent · Carriergent
Roadmap
Phase 03
Supplygent
Demand and supply planning intelligence. Translates demand forecasts into sourcing actions, optimizes inventory dynamically, surfaces planning gaps before they become crises, and monitors supply risk continuously. When Supplygent detects a material shortage signal, it automatically triggers a Sourcegent direct procurement event.
Demandgent · SOPgent · Inventgent · Riskgent
Vision
Technology
Enterprise-grade infrastructure.
Purpose-built for complex environments
Sourcegent is built on Google Cloud Platform with Anthropic Claude as the reasoning engine. All data processing, agent reasoning, and storage occurs within the GCP environment — satisfying the data residency, security, and audit requirements that enterprise clients demand. Nothing leaves the boundary.
Google Cloud Vertex AI
All agent reasoning runs within GCP. Data never leaves the enterprise cloud boundary.
Anthropic Claude Opus 4 and Haiku
Opus 4 for high-stakes reasoning — RFP drafting, evidence scoring, award recommendations. Haiku for high-volume classification at scale.
LangGraph — Multi-Agent Orchestration
Coordinates multiple specialized agents across a multi-week sourcing event lifecycle with full state management, failure recovery, and human oversight checkpoint enforcement.
Immutable Audit Infrastructure
Every agent action, every human approval, every document version preserved in a tamper-evident audit log with cryptographic hash verification. Designed for scrutiny from day one.
ERP Agnostic
SAP, Oracle, Workday, and Coupa connector support. The intelligence layer sits above your system of record — adding a new ERP connector does not change the agent logic.
The Vision
Enterprise supply chains run on decisions. ProgentAI is building the autonomous intelligence layer that makes every supply chain decision faster, smarter, and explainable
Every supply chain decision — what to source, from whom, at what price, when to reorder, how much to hold, when a supplier is at risk of failing — is made today by people working from incomplete information under time pressure. ERP systems record what happened. Procurement platforms surface what is overdue. Neither reasons about what should happen next. That is a fundamental architectural difference — not a feature gap. General-purpose AI assistants embedded in ERP and procurement systems respond when asked. They draft faster, summarize better, and alert when thresholds are breached. But they wait to be invoked. ProgentAI’s agentic architecture works differently. Agents monitor the full procurement environment continuously — supplier risk signals, market price movements, demand forecast changes, contract expiry timelines, compliance obligations, and inventory thresholds — and reason across all of them simultaneously. When conditions warrant action, agents initiate the sourcing event, execute the workflow, and stage the decision for human approval. No one has to ask. No screen has to be opened. The intelligence is always running. That gap — between a system that waits and a system that acts — is where procurement value is lost every day.
ProgentAI is building the autonomous intelligence layer that closes that gap across the entire supply chain continuum — from demand sensing and inventory planning through direct and indirect procurement through procure-to-pay and supplier performance management. We started at the highest stakes point: procurement. Every commitment to buy flows downstream into production schedules, cash flow, regulatory audits, and supplier relationships. Win the intelligence layer here and the rest follows naturally. This is what Sourcegent does.
Team
Led by someone who spent decades moving through every layer of the supply chain — from the factory floor through global ERP implementations, IT leadership, and management consulting, to managing strategic customer relationships at the world’s largest enterprise software company.
We are building the team
ProgentAI is actively recruiting Engineering Lead, AI/ML Engineer, Integration Engineer, and additional roles for the pilot delivery team.
Rama Radhakrishnan
Founder & CEO, ProgentAI, Inc.
Rama Radhakrishnan’s career spans every layer of the supply chain and technology enablement — from production planning and warehouse operations at Max Factor and logistics management at Redken, through global ERP implementations at Teradata, to CIO at Tekelec. That foundation took him into enterprise consulting at Coopers & Lybrand and CSC Consulting, where he led Oracle Applications practices and delivered multi-million-dollar systems integration engagements across complex global enterprises.
He then spent over a decade at SAP Ariba and SAP SuccessFactors as Regional Vice President, Customer Management — directly responsible for the success of SaaS Sourcing, Procurement, and HCM applications across 45 strategic enterprise clients transacting over $35 billion in business commerce. A decade of entrepreneurship building and running businesses followed, honing his P&L instincts and startup discipline. That complete journey — supply chain practitioner, technology leader, management consultant, enterprise software executive, entrepreneur — has now brought him to the frontier of autonomous AI and to founding ProgentAI.
Contact
Let’s Have a Conversation
We are in active conversations with seed investors, enterprise procurement leaders, and Google Cloud and Anthropic teams. If you are any of these, reach out directly.
Direct Contact