Order to Cash
OTC Agent in Microsoft Teams
Sales and service teams ask about orders, deliveries, credit and invoices in plain English inside Teams. The answers come back from live S/4HANA data.
Enterprise Architect · AI on SAP
I build AI agents that work inside SAP.
I'm a Senior Managing Consultant at IBM with 14+ years of SAP S/4HANA work. These days I spend most of it building AI agents that run against live ERP data. They answer the questions finance, sales and procurement teams ask every day, find revenue leakage, and act on what they find. Nothing gets written back to SAP until a person approves it.
14+years in SAP
4end-to-end S/4HANA implementations
15+AI agents and apps built on SAP
~$40Min 2025 pursuit wins supported
Selected AI work
Agents, copilots and control towers
I built these for IBM's SAP practice and for global clients in consumer goods, med-tech and manufacturing. Each one runs against a real S/4HANA system rather than mocked-up screens. Client names are left out on purpose.
Order to Cash
AI Orders Cockpit
One screen for every order, delivery and invoice. AI sorts out which orders need attention, works out how much value is at risk, and explains why each one is stuck.
Order to Cash
Order-Intake Agent
Reads unstructured customer emails and purchase orders and turns them into SAP sales orders. This takes manual keying out of order entry.
Revenue Assurance
Contract Compliance Agent
Checks sales orders and invoices against the contracted price book and finds revenue leakage. It explains each gap and proposes a correction for a person to approve.
Revenue
Revenue Opportunity Finder
Finds revenue in open orders that could be pulled forward. It comes with a dashboard and a conversational agent that explains the reasoning behind each opportunity.
Pricing
Profitability & Pricing Copilot
Shows margin and pricing insight for each product and customer combination, so commercial teams can price with evidence. A related customer-scoring assistant guides upsell and discount decisions.
Finance
Finance Agent
Answers the finance desk's daily questions: Where is my payment? Why is this invoice blocked? Which customers are overdue? It grounds each answer in company policy and live accounting data.
Record to Report
Next-Gen RTR Agent
A record-to-report proof of concept that speeds up period-end work by having an agent investigate, reconcile and explain entries.
Source to Pay
Invoice Control Tower with Autonomous Agents
An observability tower for supplier invoices. Five agents analyse real SAP documents for blocks, mismatches and risk, and every fix goes back to SAP only after a person approves it.
Source to Pay
Ask S2P Agent
A procurement assistant that answers questions about purchase orders, suppliers and invoices for buyers and AP teams. It is published to Microsoft Teams.
Supply Chain
ATP Explainer
Explains in plain language why an order was confirmed late or partially, by working through advanced available-to-promise (aATP) results.
Quality & Compliance
Serial Number Audit
A read-only audit agent on BTP for regulated (GxP) environments. It needs no custom code in SAP, which keeps the validation scope small.
Delivery
AI-Driven SAP Regression Testing
Turns hundreds of manual order-to-cash test scripts into automated tests. Natural-language test steps are converted into repeatable checks.
Platform
Principal Propagation for Agents
A proof that a person signed in to Teams reaches SAP as their own SAP user. Every agent action then respects that person's own authorizations. No shared technical user is involved.
Platform
SAP AI Accelerator Hub
A secure microsite on SAP BTP that collects every AI accelerator above. Each guest gets a scoped, time-limited view of the demos that are relevant to them.
Pursuits
Deal-Prioritization Agents
My entry for IBM's 2026 watsonx Challenge: four cooperating agents that score a consulting pipeline from CRM and ERP data so teams can focus on the deals most likely to win.
How I build
From the chat window to the SAP core
Most of these agents share one secure architecture. People work where they already are, the agent reasons over live data, and SAP stays clean.
- Teams / Joulewhere people ask
- AgentCopilot Studio, watsonx, LangGraph
- MCP toolson SAP BTP
- Cloud Connectoruser's own identity
- S/4HANAlive data, clean core
Live data, not mockups
Every number on screen comes from a real SAP document, so business users can trust what they see.
A person approves every write-back
Agents investigate and propose. A person approves before anything changes in SAP.
Clean core
Extensions live side-by-side on BTP, which keeps upgrades simple and validation scope small.
Experience
Fourteen years of SAP
I know how order-to-cash actually runs in large companies, and I bring that to every AI build.
- 2025 – present
IBM · SAP OTC Practice, Generative AI Squad
I co-lead the GenAI asset pipeline for the practice: I run ideation workshops, assess feasibility, and architect reusable AI agents that we deploy across clients.
- 2022 – present
IBM · Senior Managing Consultant
I am enterprise architect for Lead-to-Cash on a global S/4HANA template for a Fortune-100 consumer goods company. I advise the design authority on clean-core standards and lead a 20+ person delivery organization.
- 2017 – 2022
Cintas · SAP Architect
I led solution design for a greenfield SAP implementation that replaced a mainframe serving 150,000+ customers a day. I also architected an RFID + IoT garment-tracking system and a B2B/B2C e-commerce platform. SAP recognized my design contributions to S/4HANA Fashion.
- 2011 – 2017
Infosys · Lead Consultant / Project Manager
Greenfield SAP SD and GRC programs across North America and Latin America, covering EDI, e-invoicing and trade promotion.
Let's talk AI on SAP
Open to conversations about agentic AI, S/4HANA transformation and clean-core architecture.