At our first virtual Product Roundtable, we asked the room a direct question: how much of your indirect spend actually goes through a competitive sourcing event?

The answers were striking. 72% of attendees run less than a quarter of their indirect spend through any formal sourcing process. Nearly four in ten run less than 10%. Only 17% run more than half.

Flip those numbers and the picture is stark: for most procurement teams, more than three quarters of indirect spend goes to a vendor without a single competing bid. No market pressure. No competitive tension. No savings lever pulled.

This matters because competitive sourcing, when run well, routinely delivers better commercial outcomes.

The question this post answers: why is this happening, and what actually changes it?

Two reasons most indirect spend never gets sourced

The economics problem

Most procurement teams already know their longtail exists. The 75% to 85% of indirect spend sitting below their strategic categories is not invisible. It just is not worth the effort to source, given current tools and team sizes.

Running a proper sourcing event on a $100K category takes roughly the same procurement effort as one on a $500K category. The questionnaire still needs writing. Suppliers still need identifying and briefing. Responses still need scoring. The process is the process, regardless of the spend value underneath it.

So rational procurement leaders draw a threshold, and everything below it gets a preferred vendor and a purchase order. No competitive process, no market pressure, no savings. This is worth naming clearly: it is a rational individual decision that produces a bad collective outcome.

The bandwidth problem

Even above the threshold, capacity is the real constraint. We asked attendees about the biggest pain point when they last ran a sourcing event. The responses split almost exactly in two: 44% cited manual process and admin burden, and 44% cited internal alignment and stakeholder buy-in.

These are not separate problems. They compound each other. Admin burden consumes the time that should go toward alignment. Poor alignment creates more admin. Meanwhile, according to the Hackett Group, procurement workloads are growing at 10% a year while budgets grow at 1%. The gap is not closing.

Solve only the economics problem and bandwidth remains the bottleneck. Solve only the bandwidth problem and the math still does not work on longtail spend. Both need to move simultaneously, or the sourcing gap stays where it is.

A customer case study: How they solved it, and what they learned the hard way

Vikram is a customer of Omnea and leads Digital Procurement Operations in a business of 35,000 to 40,000 people operating across 64-plus countries. Before working with Omnea, they had already tried to solve this.

Several years ago, they implemented a traditional sourcing suite from one of the well-known players in the market. It was technically live. Templates were configured. Users were trained. Adoption was near zero, and it stayed that way.

The tool was not broken. The barrier to using it was simply too high, for procurement teams and for suppliers alike. When the pressure was on, people defaulted to email. The tool existed; the behaviour change did not follow.

When they came back to the market, they had one non-negotiable: simplicity. And one structural requirement that shaped everything else. Sourcing had to feel like a natural extension of something teams already used, not a separate system requiring a separate decision to engage.

They had already adopted Omnea for intake. When sourcing was added as a capability within the same platform, the behaviour change required was minimal. Teams knew the tool. The expansion felt like a next step, not a new product.

Three things Vikram would tell any team considering a similar move:

Start narrow and repeatable. Three bids and a buy in one pilot country. A single use case that can be adopted quickly and works for common procurement scenarios. The MVP becomes the standard faster than expected. What looks like a temporary starting point often becomes the model everything else is built on.

Adoption beats features, always. A simpler tool that gets used beats a sophisticated tool that does not. The investment in change management, identifying change champions, running deliberate training, building internal momentum, matters more than configuration. Without adoption, you have nothing.

Train your scorers before you launch. The step most teams skip. Scoring RFP responses is simple once someone has been shown how, but if it is the first time a requester does it mid-event, it creates friction and delays. A short upfront session removes that bottleneck entirely.

The teams that get sourcing adoption right do not introduce a new tool. They extend one people already trust.

What autonomous sourcing actually does to the coverage gap

When AI handles the mechanics of a sourcing event, two things happen. The time cost per event falls. And the categories that previously sat below the sourcing threshold, where the effort never justified the return, now have a viable ROI. The $500K category with $50K in savings potential is worth running a process on. So is the one below it.

Triggering an event. Autonomous sourcing removes the decision entirely: the workflow triggers automatically when spend requests cross a threshold. No bandwidth required, no event missed.

Building the questionnaire. AI drafts the questionnaire in minutes from the intake brief. A generalist buyer reviews and edits. What used to take a specialist hours becomes a five-minute task.

Supplier identification. AI surfaces incumbents cross-referenced against existing agreements, plus new options, automatically. No lists assembled from memory. No default incumbent by preference.

Managing responses. A supplier portal with magic links replaces email threads and chasing. No account creation required. One supplier told us they'd blocked four hours for an RFP and finished in thirty minutes.

Scoring and award. AI does first-pass scoring with rationale attached. The human reviews, overrides if needed, and makes the final call. Alignment delays gone; audit trail built in.

As one procurement director put it: "We're not sourcing it at the moment. If we can do that, it doesn't matter how much it generates. It's just free money."

The threshold for what is worth sourcing moves. Spend that was always going to the incumbent now goes through a competitive process, with the same headcount.

Book a demo

If your team is in the 72% running less than a quarter of indirect spend through a competitive process, value is sitting in your longtail unrecovered. Book a demo and we will show you what autonomous sourcing looks like on your spend profile.

Questions from the session, answered

These questions came in live from procurement leaders during the roundtable.

Yes. One of the most effective ways to cut sourcing time is to identify that a need is already covered without running an event at all. Throughout the intake process, if a tool or service matching the request is already in use, the platform surfaces that to the requester. Where an existing agreement is in place, the incumbent supplier is included in any suggested RFP and weighted accordingly. The platform draws from both existing supplier relationships and external market data, so it is not limited to what is already contracted, but it always factors in what is.
This is an area of active development. The distinction worth drawing is between reactive sourcing, responding to requests as they come in, and proactive category strategy, where procurement is identifying opportunities ahead of demand. Omnea is being used for the latter in a number of deployments, including with a large US grocery retailer. The approach uses AI document review to scan contracts and spend data, identify savings opportunities, and raise category strategy requests in the platform. Sub-requests can be grouped under a single strategy, with reporting on how each performs over time. The underlying requirement is good data: supplier lists, spend categorisation, and contract repository access. Where those exist, the platform can start doing the pattern recognition.
Both. The platform draws from incumbent suppliers with existing agreements and actively uses that context to avoid unnecessary new vendor relationships where one already fits. But it also surfaces new suppliers based on the category and requirements, cross-referencing factors like ratings and category fit. The goal is to make sure the event starts with a genuinely competitive field, not just the vendors procurement already knows.
The platform connects to your ERP across entities and subsidiaries, which means policies, delegate of authority structures, and approval thresholds can be configured per buying entity or region. Adecco operates across 64-plus countries with multiple ERPs, and regional policy variation is built into how their deployment runs. On language, the platform currently supports Spanish, French, German, and Japanese for the procurement-facing interface, with supplier-facing translation in active development.
Yes. The platform handles pricing inputs ranging from a single flat fee through to structured annual licence models with multiple line items, and up to full Excel-based pricing templates where suppliers download, complete, and re-upload. In each case, the AI ingests the data and enables comparison across suppliers, including scenario analysis and total cost of ownership calculations where the underlying data supports it.
Three things came up consistently across the session. First, simplicity. Vikram at Adecco had previously implemented one of the market's well-known sourcing suites and seen near-zero adoption. The tool that gets used beats the tool with more features. Second, adoption focus. The implementation approach is deliberately change-management-led, not just configuration-led. Third, the underlying product vision: the goal is not to make existing sourcing events slightly more efficient. It is to make sourcing viable on spend that was never worth the effort before. More sourcing, more savings, same team.