What Coupa’s Data Reveals: Intake Is the Strongest Standalone Lever for AP Efficiency

Freddie White wears many hats at Coupa: Senior Solution Advisor, Intake & Orchestration SME, and Coupa Spend Lab Strategist. Spend Lab is the research and intelligence arm of Coupa, focused on generating strategic insight from aggregate patterns across one of the world’s largest repositories of business spend data.

Andres Calderon, Coupa Spend Lab Economist, also contributed to the insights in this post.

This post is meant for:

• Finance leaders who have invested in AP automation and are still asking why manual review rates aren’t falling as fast as expected.
• Finance leaders who can see the cost of invoice exceptions on their P&L but are not sure where the problem is actually originating.
• Procurement leaders who want to demonstrate strategic impact beyond savings numbers and rely on a data-backed way to connect their upstream decisions to the financial outcomes their CFO measures.
• Decision-makers planning an AI investment in source-to-pay (S2P) and you’re trying to understand what process foundations need to be in place for that investment to pay off.

Why intake, not downstream fixes, drives AP friction

When AP teams are overwhelmed by manual invoice reviews, the default explanation is downstream: invoice format, supplier behaviour, matching tolerances, payment terms. Just under two-thirds of CFOs plan on applying AI and automation to reduce manual processes and operating costs. But too often, those manual review rates don’t fall as fast or as far as expected, and the question of why rarely gets a satisfying answer.

Coupa Spend Lab set out to find one. Our analysis of Coupa’s $9.5 trillion in 2025 (now $10.5 trillion in 2026) in managed spend across 98,330 monthly observations — the first causal investigation of this kind conducted on the Coupa community dataset — produces a finding that changes the conversation entirely:

The dominant cause of AP friction is upstream, not downstream. Structured intake is the single strongest predictor of AP efficiency, outranking supplier count, invoice volume, payment terms, and every other measurable variable — by a factor of approximately three.

This is Coupa Spend Lab’s first published analysis of the Intake Efficiency Dividend. Coupa’s enormous dataset makes causal analysis at this scale uniquely possible, and we’ll be sharing further findings on a regular basis as the research programme matures. These are initial findings, and as we note later in this post, the data may in fact be understating the effect.

The finding survives 15 successive robustness checks, including tests specifically designed to rule out the possibility that it’s a mathematical artifact rather than a genuine structural relationship.

How Coupa measures intake success

The intake success rate’s definition is precise for a reason. Both conditions — PO-backed and linked to an active contract — are required. The distinction matters because PO compliance alone isn’t enough. The contract link is what confirms the spend was strategically governed before it reached AP. Specifically, the link demonstrates:

• A real business need was recognised.
• The request was either routed to a preferred supplier immediately, or autonomously orchestrated through sourcing and contracting if the requirement was genuinely a new one.
• Compliant suppliers were selected and contracted terms were followed before the invoice ever landed.

Without both elements, an invoice may be structured in form but not in substance, and AP absorbs the resulting cost.

This makes the intake success rate a summary metric for the entire S2P process. It captures how completely upstream procurement decisions are governing downstream transactions at the moment the invoice arrives. For a deeper look at how intake and orchestration work as a connected system, the Spend Matters white paper is a useful starting point.

Structured intake cuts manual review by 19.5 percentage points

Our analysis estimates that without any structured intake in place, approximately half of all invoices require manual intervention. That amount is the process debt of unstructured procurement, in other words, the volume of work landing on AP's desk because upstream decisions were incomplete.

As the Intake Efficiency Dividend callout above shows, organisations running full structured intake see an expected manual review rate approximately 19.5 percentage points lower than those without it and changing nothing else within their process.

What does that mean for AP teams? For an organisation processing 50,000 invoices per month, that translates to approximately 5,850 fewer manual AP touches per month. The annual cost of that reduction is calculable from a simple formula: Multiply your monthly manual touch reduction by 12, then apply your own cost differential between manual and touchless invoice processing. At any reasonable cost-per-invoice benchmark, the resulting figure is significant, and it’s directly attributable to a single upstream process decision. (The cost to process an invoice varies widely, with most estimates ranging from $15 to $35. Research from Ardent Partners states that the average business spends $9.90 to process a single invoice.)

For organisations already running on Coupa and looking to expand their intake coverage over time, Coupa Spend Lab has a separate within-customer estimate that is more precisely calibrated for that conversation. Speak with your Coupa team or customer value manager for that analysis.

A note on the finding itself: The intake success rate in this analysis is measured using invoice-level data — specifically, the proportion of invoices that are both PO-backed and linked to an active contract. This means organisations that have achieved strong contract and PO coverage through means other than a formal intake module will appear as high-intake in the data, regardless of how that coverage was achieved. Because of this, the true impact of structured intake on AP efficiency is probably larger than the figures above suggest. Our measurement approach is, if anything, conservative. We’ll continue to refine this measurement as the research program develops.

Intake gains hold at every stage of growth

The most commercially significant finding in our analysis is what happens when intake coverage scales.

Our analysis tested directly whether the efficiency gain from improving intake coverage fades as coverage increases within an organisation. It doesn’t. Moving from 20% to 40% intake coverage delivers the same efficiency gain as moving from 60% to 80%. Each additional percentage point of intake coverage delivers the same efficiency gain as the last — the improvement doesn’t slow down as coverage scales.

For midsize and emerging enterprises in particular, this removes the two most common reasons to delay:

• An intake investment pays off more or less at different scales.
• It makes more sense to address intake once the organisation is larger or further along its digital maturity journey.

Our data doesn’t support those assumptions. The dividend is the same now as it will be later. The only thing that grows with delay is the cost of operating without it.

The same applies to AI investments in the source-to-pay space. Structured intake is the process foundation that determines whether AI tools in procurement and AP deliver genuine value or simply automate a fragmented baseline. A benefit that’s often overlooked here is structured intake’s ability to ensure that key data and metadata seamlessly pass downstream. This enables autonomous agents to reason effectively over newly created objects through governed, autonomous orchestration. If your organisation is planning AI investment in this area, we strongly recommend that you understand the process foundations that need to be built before you commit resources.

Why intake coverage doesn’t always improve in a straight line

When we compare organisations at different intake coverage tiers rather than tracking individual organisations over time, a pattern emerges that initially appears counterintuitive: Organisations in the 20-40% intake tier show slightly higher observed manual review rates than those at 0-20%.

How is this not a reversal? Organisations in the 20-40% range tend to be in active implementation. Specifically, they have enough contract and PO coverage to register in the intake metric, but the downstream AP benefit has not yet fully materialised. When statistical controls for scale and digital maturity are applied, we see a clearer negative correlation between intake and manual review rates. When we track the same organisations over time as their intake coverage increases — rather than comparing different organisations to each other — the relationship is consistently linear throughout. This is the most direct and reliable way to test the no-diminishing-returns finding, and the result is unambiguous. If this sounds like your organisation, the Power of Smart Intake & Orchestration webinar covers what the transition from active implementation to realised benefit typically involves.

3 intake objections the data rules out

“Our supplier base is too complex for intake to work at scale.”

Our analysis finds that supplier count has effectively zero steady-state impact on AP friction when intake quality is controlled for. An organisation managing 10,000 suppliers can be as frictionless as one managing 100, provided those suppliers are governed by structured intake. Complexity isn’t a barrier to intake investment. For most organisations, it’s the strongest argument for prioritising it sooner.

“We’ve already captured the easy wins.”

When we tracked the same organisations over time as their intake coverage increased, we discovered that each percentage point of additional intake coverage is worth the same as the last, across all observed intake levels. There’s no point at which the remaining gains aren’t worth pursuing.

“AP efficiency will improve on its own as we mature.”

Put plainly, it won’t. We ran a separate analysis specifically testing whether AP efficiency improves passively over time — without any change to intake — and found no evidence that it does. Efficiency only moves when upstream intake actively changes. Every gain is earned, and every gain is directly attributable to deliberate process decisions. That’s important for any finance or procurement leader who needs to demonstrate the ROI of structured intake investment: The data makes it attributable.

For more on what it takes to actively improve structured intake — and what the most mature organisations are doing differently — the webinar linked above covers this in detail.

Intake plus e-invoicing: A near-zero-touch combination

Our analysis also finds that structured intake and e-invoicing operate through independent mechanisms — parallel levers whose benefits compound rather than compete.

What we found: Organisations combining a 68% intake success rate with a 62% e-invoice processing rate achieve a median manual review rate of just 12.5%. To put those numbers into perspective: While structured intake does the heavy lifting by reducing your manual review rate by 19.5%, layering e-invoicing on top drops it by an additional 7%. Neither threshold alone produces that result. Together, they drive AP friction close to zero. This approach effectively elevates your organisation to best-in-class benchmarks.

Your intake roadmap: 4 steps, in order

As Coupa looks at the roadmap for maximising intake success rates, solving the employee adoption puzzle remains a top priority. The addition of Tonkean’s capabilities to the Coupa platform, including natural language intake, no-code workflow building, and over 250 native connectors, directly addresses this barrier. By eliminating the manual handoffs between systems, Tonkean bridges the gap between disparate teams and tools. When the structured channel is fast, frictionless, and available in the apps employees already use to work, intake compliance naturally follows. See the unified Coupa + Tonkean experience in action.

To get there, our data supports a clear order of operations:

1. Choose the right configuration first

Moving to full intake-to-pay from AP-only or PO-only is the highest-impact single decision you can make. Our comparison between organisations running full structured intake and those without it quantifies the expected difference. Everything else builds on this foundation.

2. Expand coverage within your existing setup

Once the right configuration is in place, expand how much of your spend flows through structured channels. The data shows each additional percentage point of coverage delivers a consistent, predictable reduction in manual AP work. To effectively prioritise this expansion, customers can leverage Coupa Spend Pulse and AI Insights to pinpoint categories and commodities that need attention. By watching real-time market signals and correlating external data directly against your enterprise spend, these AI-powered capabilities surface clear, objective alerts when it is time to take action. Equipped with these verifiable insights, teams can focus precisely on where to improve their intake experiences and subsequent orchestration.

3. Close the contract gap, not just the PO gap

Organisations focused only on PO compliance are capturing part of the available efficiency gain. Both elements — PO-backed and contract-linked — are required for an invoice to count as fully structured and for the full dividend to be realised.

4. Shorten your approval chains

Our analysis identified approval chain length as the only variable with a positive friction-increasing coefficient among the top predictors in the 800+ variable dataset. Once you establish the lever of structured intake, focus on optimising the next lever: how you simplify approval chains.

For a deeper look at building a strong P2P foundation that makes this sequencing practical, Coupa’s Beyond Intake playbook covers the full picture.

What it costs your AP team to wait on intake

Our analysis clearly shows that AP efficiency doesn’t drift toward improvement on its own. There’s no passive learning curve or platform-wide trend that will fix your manual review rates while you wait.

If your organisation chooses to maintain the status quo (relying on downstream fixes while leaving upstream intake unstructured), you’re effectively choosing to pay a process tax indefinitely.

Here’s what inaction typically produces:

• Perpetual manual debt: Without structured intake, the volume of manual work landing on your AP team remains static. You will continue to burn resources on exceptions that are, by definition, preventable.
• Capped returns on tech investments: Investing in AI tools without first establishing structured intake is like fitting a high-performance engine to a vehicle with no steering. The automation is real. But it's accelerating a fragmented baseline rather than a structured one, which limits the return on every dollar invested in your technology stack.
• Delayed strategic impact: Procurement leaders who wait for a more convenient time to address intake are simply pushing the bottleneck further into the future. The efficiency dividend we observed is constant; the only variable that changes with delay is the time, money, and focus lost during the wait.

The question isn’t whether your organisation can handle the complexity of structured intake — our data shows that complexity isn’t a barrier. The question is whether you can afford to continue operating with the friction your AP team is absorbing today.

The path to efficiency is active, intentional, and clearly defined by the data. The timing, however, remains yours to decide.

* The intake success rate emerged from a causal analysis of Coupa’s community dataset. It emerged as the dominant lever across all five model specifications, surviving 14 successive robustness checks including penalised Lasso variable selection, fixed effects identification, and distributional specification tests.