Rodrigo Coronado, Partner and Head of Innovation at Jarvis Consulting Group, works with finance and supply chain leaders on how to put their data to use. We asked him what JrvsOS is, why it starts with invoices, and what an organization can expect when implementing it. 

What is JrvsOS? 

JrvsOS is the system Jarvis uses to put AI to work on an organization's own data. It connects existing systems of record, such as the ERP, field systems and document libraries, to one governed data foundation without replacing them. It then adds business fluency, meaning the ontology, rules and knowledge of the client's industry and company, along with the skills, agents and tools that each do a specific job. Applications sit on top and solve defined business problems. The first is Invoice Reconciliation, which checks every invoice against the contract behind it before the invoice is paid and places the evidence beside every discrepancy so the accounts payable team can act right away. 

Find the needle (2 min.). Sample data.

What problem does Invoice Reconciliation solve? 

It solves invoice-to-contract leakage. A rate is billed above the contract price, a volume discount is never applied or a charge appears that the agreement does not cover. Each case is small on its own, which is why the total goes unnoticed. World Commerce & Contracting puts the average at 8.6% of contract value. For an organization with $100M in contracted supplier spend each year, that is a number worth knowing. 

If the data already exists, why isn't this being caught today? 

The answer is scattered across systems. The invoice sits in the ERP, the purchase order in procurement, the contract and its amendments in a document library and the vendor ratings in a supplier portal. Checking one invoice properly means opening all of them. Most teams sample a few invoices or rely on three-way matching, which confirms the purchase order and the receipt but rarely goes back to the terms of the contract. 

How does Invoice Reconciliation work? 

It starts by building a knowledge graph around each invoice. The invoice is linked to its contract, master agreement, purchase order, goods receipt and vendor, including that vendor's ratings, rebate tiers and history. With those links in place, Invoice Reconciliation runs its checks the moment an invoice arrives: price against the contract rate, quantity against the purchase order, quantity against goods received, duplicate invoices and arithmetic. Clean invoices move on to payment. Exceptions go to the right person with the evidence attached, and the invoice is filed under its contract once it is resolved. 

Where does AI come in, and how does finance know it can trust the results? 

AI handles the work that used to need a person, such as reading contracts, interpreting terms and connecting records that were never linked. The dollar figures come from fixed rules applied to those records, so every amount Invoice Reconciliation reports can be traced to a clause, a purchase order line or a receipt. That matters because a controller has to be able to explain a recovered amount to a supplier and to an auditor. 

What does it find in practice? 

In the sample data we use for demonstrations, a single invoice from a packaging supplier shows four issues worth $4,128.32, including a price above the contract rate, a freight surcharge the contract does not allow and a volume rebate that was earned but never credited. Invoice Reconciliation stops $2,568 of that from being paid. Run across a full set of invoices, the same checks build a view of leakage by supplier and by type of issue, whether there is one invoice or ten thousand. 

Who should be paying attention? 

CFOs, Controllers and VPs of Finance who are being asked for an AI plan, along with supply chain and procurement leaders responsible for supplier spend. It suits organizations with high invoice volumes and contracts that carry tiered pricing, rebates or schedules of allowed charges, which describes much of the energy and industrial sector. 

What comes after invoices? 

Once contracts, purchase orders, receipts and vendors are connected, the same foundation supports other work, such as preparation for contract renewals and spend analysis. Invoice leakage is the right place to start because the return can be measured in dollars, and that gives finance a clear case for the next step. This follows the path we use with every client: start with one system, prove that the results are traced and tested, expand to more systems and then scale to new applications on the same foundation.