Spreadsheets vs AI Coordination: Which Survives a 40-Supplier Project (And Why $1.5M a Year Is Just the Opening Number)
Ivalua put a number on something most ops leads already feel in their gut: businesses are burning more than $1.5 million a year on manual procurement work that software could be doing instead. Not on bad decisions. Not on bad suppliers. On the grind of chasing a confirmation email, re-keying a PO into a second system, or hunting through a WhatsApp thread for the delivery date someone mentioned three weeks ago.
Here's the part the vendor blogs recycling that stat don't say out loud: 86% of companies are already deploying, piloting, or at least talking about AI in procurement. Only 34% have it running in production. The gap between those two numbers isn't a technology problem. It's a data problem, and it's a coordination problem — and on a job with 40 active suppliers, both problems compound fast.
"86% of organizations say they're doing something with AI in procurement. Only 34% have actually shipped it. The other 52% are the ones still re-typing PO numbers into a second system."
Background and Context
The Ivalua report isn't wrong about the $1.5 million figure, but it's a company-level average. It doesn't tell you what happens on a single project with real complexity — a retail fit-out with 18 trade suppliers, an industrial install with 25 equipment vendors and three freight forwarders, an energy project touching 40-plus parties across procurement, logistics, and compliance. That's where the manual-work cost stops being an abstract annual figure and starts being a Tuesday afternoon spent cross-referencing four different inboxes to find out why a delivery is three days late.
Nearly half of procurement teams report their data sits in four or more separate systems. Half say that data isn't reliable enough to actually drive AI insight. That's the real reason the 86%-to-34% gap exists. You can buy the smartest AI assistant on the market, but if your supplier confirmations live in email, your delivery updates live in WhatsApp, and your PO status lives in a spreadsheet nobody's updated since Monday, the AI has nothing coherent to work from. The automation stat everyone's quoting assumes clean inputs. Most teams don't have clean inputs. They have scattered ones.
One Capture Point, Not Four
Requests arrive by email, WhatsApp, a PDF attachment, or a line item dropped into a project system. PashX reads all of it into one record instead of leaving each channel as its own silo someone has to manually reconcile.
Supplier Follow-Up That Doesn't Sleep
On a 40-supplier job, half the "manual work" Ivalua is pricing out is just chasing confirmations and status updates. PashX fires the follow-up, logs the reply, and flags the ones that go quiet instead of waiting for a human to notice.
POs Tracked From Issue to Close
Every PO is tied to its supplier thread, its delivery status, and its invoice — so there's no separate spreadsheet tracking "where did we land on PO #4417" three weeks after it was sent.
Exceptions Surface Before They're a Fire
Late delivery, mismatched invoice, missing documentation — PashX flags the exception and chases the fix. A human still signs off on anything judgement-based; the system just makes sure it gets to that human in time to matter.
One Operational Workspace, Not Four Systems
Half of procurement teams say their AI insight is undermined by fragmented data across systems. PashX fixes that at the root by keeping supplier, PO, delivery and invoice data in one place instead of asking AI to stitch together four disconnected tools.
| Task on a 40-Supplier Project | Spreadsheet + Inbox Approach | PashX Coordination |
|---|---|---|
| Tracking who confirmed a PO | Manual check across email threads and a status column someone forgets to update | Confirmation status attached to the PO record automatically |
| Chasing a quiet supplier | Relies on someone remembering to follow up | Automatic follow-up cadence with escalation flag if no response |
| Reconciling delivery vs. invoice | Done at month-end, after problems have compounded | Flagged as an exception at the point of mismatch |
| Finding the latest WhatsApp update on a shipment | Scroll through a chat thread | Captured into the project record the moment it's sent |
| Reporting project status to leadership | Manual roll-up from four sources | Live view from a single workspace |
A Closer Look: Why the Gap Between 86% and 34% Is the Real Story
Everyone quoting the $1.5 million figure skips past the number that actually explains it. Almost nine in ten procurement teams are "doing something" with AI. Only a third have it in production. That's not a budget problem — most of these companies have already spent the money. It's an infrastructure problem. You can't automate decision-making on top of data that's scattered across four systems and half-trusted by the people who'd have to rely on it.
- Fragmented data: Supplier comms in WhatsApp, PO status in an ERP, delivery notes in email, invoices in accounting software — four sources of truth that disagree with each other by the time anyone checks.
- Pilot purgatory: A lot of AI procurement tools get piloted on a single category or a single supplier relationship and never scale past it, because the integration work to connect every channel is treated as a separate project.
- Trust in the data, not the model: Half of teams don't trust their own data enough to act on AI output. The fix isn't a better model. It's cleaner, centralized input.
- Manual work hiding as "coordination": Chasing confirmations, re-entering PO details, cross-checking invoices against deliveries — none of this shows up on a job description, but it's most of what eats the $1.5 million.
How PashX Outperforms the Competition
- vs Coupa / SAP Ariba: These platforms assume your suppliers will adopt a portal. On a construction or fit-out project, half your suppliers are a foreman replying from a personal phone. PashX meets suppliers where they already communicate — email and WhatsApp — instead of requiring portal logins nobody will maintain.
- vs JAGGAER: JAGGAER's strength is structured sourcing workflows for large enterprise categories. PashX is built for the messier middle: fast-moving, multi-trade, document-heavy projects where requests arrive unstructured and someone has to turn them into a tracked PO within the hour.
- vs generic AI-research tools (procurement AI agents, chatbot layers): A lot of the newer entrants bolt an AI chat interface onto existing procurement data without fixing the fragmentation underneath. PashX centralizes the intake and coordination first — the AI layer sits on top of one coherent operational record, not four disconnected ones.
Key Details
- No single source of truth, no reliable AI: If supplier, PO, delivery and invoice data live in four places, any automation layer on top is guessing. Fix the capture point before adding the model.
- The manual-work cost is mostly follow-up, not decision-making: Most "procurement work" that's eating budget is chasing status, not making judgement calls. That's exactly the part worth automating first.
- Human approval still matters: Automating the chase doesn't mean automating the decision. Exceptions, pricing disputes, and judgement calls still route to a person — the system's job is making sure they reach that person with enough time to act.
- Pilot scope determines pilot success: The 34%-in-production figure suggests most pilots die when they try to scale across every supplier and channel at once. Start with the coordination layer for one active project and expand from there.
Availability and Next Steps
If you're running a project with more than a handful of suppliers and you're still coordinating it across email, WhatsApp, and a spreadsheet someone updates "when they get a chance," you already know where the $1.5 million figure is coming from in your own operation. It's not one big inefficiency. It's forty small ones, repeated across every supplier, every week.
The fix isn't a bigger procurement suite. It's getting every request, every PO, every delivery and every invoice into one place that chases the follow-ups automatically and only asks you to weigh in when a real decision needs making.
About PashX
PashX is a procurement and project coordination autopilot. It captures requests from email, WhatsApp, documents and project systems, then coordinates suppliers, purchase orders, deliveries, invoices and exceptions in one operational workspace. It chases the follow-ups; you approve the judgement calls. Visit pashx.com.
Ready to get started?
See how PashX coordinates your suppliers, POs and deliveries in one workspace.
Open Admin Dashboard →