AP and AR Automation: A Strategic Intelligence Brief

Two Crossover studies, run nine months apart for two different sponsors, converge on the same buyer reality. The CFO buying AP and AR automation in 2026 is not the buyer the AI-native disruptor narrative is targeting. They want stability, integration depth, and demonstrated production performance. They are willing to pay a meaningful premium for AI capability, but only on renewal, never as the basis for selection. They want a human in the loop on the parts that matter, and they want a vendor who has been deploying finance software long enough to deploy theirs without breaking it.

The strategic implication for Esker is that two decades of incumbency, deep ERP integration, and a global production footprint are structural advantages, not legacy baggage. The pitch that wins this market reframes Esker as the answer the CFO has already trusted, with the AI roadmap layered on top. The pitch that loses this market is the one that tries to compete with Rillet, Campfire, and Basis on AI-native ground.

This brief walks through six strategic insights, three white space opportunities, and seven recommended actions. The data is drawn from the AP Automation VoC (Apr 25) for behavioral signal in the AP automation category, and from the Office of the CFO engagement for the broader buying environment that frames every AP and AR conversation right now.

AP Automation Voice of Customer engagement (Apr 25)

A sell-side category VoC. n=21 customers and decision-makers across AP automation platforms, with a Medius-customer overweight to support competitive positioning. Vendors covered include Medius, Bill.com, Coupa, Basware, Oracle AP, AvidXchange, and Yooz. The engagement quantifies vendor selection criteria, mission criticality, willingness to pay, migration difficulty, renewal intent, and competitive differentiation in the AP automation category.

Office of the CFO AI Adoption engagement (Mar 26)

A market intelligence engagement. n=129 CFOs and senior finance leaders across companies with $50M to $5B+ in revenue. The engagement quantifies AI adoption pipeline, vendor selection criteria, build-versus-buy strategy, willingness to pay, board pressure, autonomous-versus-assistive preferences, and the LLM-versus-purpose-built-tool architecture debate.

Together they provide both behavioral signal (Medius customers in the act of buying, retaining, and expanding) and strategic context (CFOs describing how they will allocate budget, evaluate vendors, and respond to board mandates over the next two to three years).

The two studies use different selection-criteria instruments but produce the same finding. AI ranks last or near-last among factors CFOs weigh when picking a vendor. Integration depth, ease of use, and demonstrated proof rank highest.

The CFO engagement calls this finding the Proof Paradox. References and POC results rank 2.21 points higher than accuracy as a standalone metric, the largest spread in the entire evaluation framework. Yet 71% of CFOs cite inaccuracy as their biggest concern. CFOs are not skeptical of accuracy as a value, they are skeptical of accuracy claims that are not backed by demonstrated production performance.

The AP Automation VoC produces the same conclusion in different language. AI/ML capabilities ranked 6.8 out of 10 in the selection decision, second-to-last among nine factors. But on willingness to pay, Medius customers consistently said they would pay more for AI features. The pattern is clear. AI is the renewal and expansion lever. Stability, integration, and proof are the selection drivers.

Three findings from the Office of the CFO engagement, read together, define the buyer's posture toward AI-native vendors and the LLM-only architecture pitch.

The numbers describe a market that wants AI-enhanced finance workflows, not AI-replaced finance workflows. They want a vendor who already understands GAAP, SOX, audit, multi-entity consolidation, tax, and the operational quirks of SAP, Oracle, NetSuite, and Microsoft Dynamics. They want approval matrices, exception handling, and human review on the parts that carry audit risk.

This is a description of the architecture Esker has been deploying for two decades. The AP Automation VoC confirms the same pattern from the customer side. Medius customers reach 89% touchless processing on routine invoices while maintaining configurable approval matrices on the exceptions. That is not autonomous AI. It is purpose-built workflow automation with AI inside it. The exact model 33% of CFOs in the engagement say they want.

The AP Automation VoC includes the single most useful retention proof point in any AP study Crossover has run. The migration difficulty chart compares how customers of seven AP vendors describe the difficulty of switching to a different platform.

Medius is the only vendor in the entire study where any customer describes switching as difficult. The mechanism is identifiable. 88% of Medius customers integrate AP automation with their procurement and PO system, versus 67% across all vendors. Once that integration is wired in, the cost to unwind it is structural, not contractual.

The CFO engagement confirms the same dynamic from a different angle. 76.3% of CFOs prefer embedded over standalone solutions. Embedded means more integrations, more touch points, and ultimately more switching cost. The Medius retention story is just the embedded preference made measurable.

The CFO engagement quantifies what every finance vendor has been observing: pilots are not converting to production. 36% of CFOs do not pilot AI at all. Among those who do pilot, 25% see zero pilots reach production. Another 24% report under 25% pilot success. Only 4% of all respondents report pilot success rates above 50%.

Why pilots fail

Money is not the constraint. Trust, integration, and capacity are. The top three barriers (ROI proof, integration, skills) are all things the vendor either controls or can supply. Skills gaps in particular are an opportunity. CFOs cannot hire AI talent and they cannot retrain finance teams fast enough. Vendors who arrive with implementation services, embedded enablement, and managed deployment offerings remove the single biggest internal blocker.

This is why Esker's services and customer success organization is a competitive asset. The AP Automation VoC reinforces it. 50% of Medius customers cite implementation methodology as a vendor strength versus only 29% across all vendors. That gap is exactly the pilot-failure problem solved on the other side.

The CFO engagement is unusually direct on pricing power. 95% of CFOs will pay an AI premium. Only 5% refuse one. 28% will accept 15 to 30% above their current pricing. The gating factor is not budget. It is value demonstration.

The driver is named explicitly in the CFO dataset. PE and VC boards are mandating cost reduction through AI-enabled headcount cuts and benchmarking AI adoption rates across portfolio companies. The framing in the open-text responses is strikingly consistent: "don't fall behind competitors" combined with "cut costs through labor replacement." Less innovation, more economic necessity.

The AP Automation VoC captures the willingness-to-pay signal directly. Three Medius customers told the interviewer in their own words what they would pay more for.

This is the most counterintuitive finding in the Office of the CFO engagement, and it should change how Esker (and Bridgepoint) think about platform pitches. The data is unambiguous. CFOs are buying the killer feature first and the platform second. They want to see one capability demonstrated at depth, not a 47-tile product map.

The CFO engagement identifies the killer features that would unlock budget today.

The number one killer feature is sitting directly in Esker's wheelhouse: automated invoice processing with 99% or better accuracy. This is not a category Esker needs to build into. It is the one Esker has owned for two decades. The pitch should be a single number with a credible reference behind it.

The AP Automation VoC reinforces the discipline. Where Medius is praised, the praise is specific and operational ("OCR accuracy and ERP integration are night-and-day better"). Where competitors are critiqued, the critique is generality ("Coupa's solid performance aligns with the functionality offered by other top tools, making it a reliable but not uniquely differentiated solution"). The vendors who win specificity, win the deal.

Three opportunities show up in the Medius data that Esker can either attack as competitive openings or address as roadmap priorities, depending on Esker's current capabilities.

Tax compliance integration

21% of all AP automation customers integrate a separate tax compliance software. 0% of Medius customers do. A Medius retail user explicitly asked for sales tax detection on invoices in the willingness-to-pay section. Medius customers do not have a separate tax tool because they expect Medius to do it, but Medius does not. If Esker has tax automation today, this is a head-to-head differentiator that nobody is currently using. If Esker does not, the willingness-to-pay signal in the Medius data makes the build case easier than usual.

Fraud prevention

0% of Medius customers cite fraud prevention as a vendor strength, versus 5% across all vendors. For an AP and payment platform sitting on the supplier-payment flow, this is an exposed flank. The CFO data shows 34% of CFOs say real-time anomaly and fraud detection would unlock budget today. Two ways to play this. Either Esker has fraud prevention features and is under-marketing them, or Esker should build them. Both are addressable.

Cash flow forecasting and predictive analytics on AP and AR data

43% of CFOs in the engagement say cash flow forecasting with variance explanations would unlock budget. Multiple Medius customers explicitly asked for predictive analytics on AP data. Esker's combined AP and AR footprint produces the data for both directions of the working capital cycle, which is exactly the dataset CFOs want to forecast against. If Esker is not currently leading with AI-powered working capital intelligence, this is the highest-margin product expansion opportunity in either dataset.

The seven moves that follow from the data, ordered by speed-to-impact.

1. Restructure the pitch deck around proof and references first. Lead with customer count, invoice volume processed, production accuracy, and named references in the buyer's industry. The CFO engagement's Proof Paradox finding is unambiguous: references and POC results outweigh accuracy claims by the largest spread in the entire evaluation framework. Esker's installed base is the strongest asset Esker has, and most pitch decks under-deploy it.
2. Publish the production accuracy number across the installed base. The killer feature CFOs are buying is "automated invoice processing with 99% or better accuracy." Esker has the right to claim this number with a credibility no AI-native vendor can match. If the number exists internally but is not externalized, externalize it. If it has not been measured precisely, measure it.
3. Quantify migration difficulty and switching cost. The AP Automation VoC made retention defensible to the deal team and to potential buyers because it quantified migration as "very or extremely difficult" for 38% of customers. Esker should commission its own equivalent measurement. The result becomes the centerpiece of LP communications and the strongest defensive asset against any AI-native pitch attempting to displace.
4. Lead with the killer feature wedge, not the platform. CFOs prefer point solutions that excel at one thing. The platform pitch comes after the wedge has been won, not before. The wedge for Esker is invoice processing accuracy, with AR, procurement, tax, and analytics presented as expansion paths, not selection criteria.
5. Price to displaced labor cost, not feature parity. 95% of CFOs will pay an AI premium and 28% will pay 15 to 30% more. The math the buyer is doing internally is "how much FTE cost am I displacing, and what fraction transfers to the software line." Anchoring price to displaced labor produces a multiple of feature-parity pricing. The board has already given the CFO permission.
6. Reposition AI as the renewal and expansion lever. AI ranks 6.8 in selection in the AP automation engagement and 6.67 as standalone accuracy in the CFO engagement. It ranks first in willingness-to-pay verbatims. The implication is that AI capability should sit in the second half of the sales conversation as the value-driver of next year's contract, not the entry pitch. This sequencing also defends against the AI-native vendors who are trying to make the conversation about AI-native versus traditional.
7. Build the Bridgepoint portfolio motion. 57% of CFOs face moderate-to-strong board pressure on AI adoption, and PE boards explicitly benchmark AI adoption rates across portfolio companies. Bridgepoint sits on Esker's board and on the boards of other portfolio CFOs who are getting the same questions. Esker can be positioned as the AI answer Bridgepoint already owns, replicable across the rest of the portfolio. That is a co-marketing motion, a referral motion, and a thesis motion all at once.

Esker's strategic workflow runs across six stages, from Market Intel through Strategic Planning, with Transaction Prep as the gating stage Bridgepoint cares about most as the exit horizon approaches. Crossover Intelligence plugs into each stage through one of two surfaces.

Five offerings, scoped to Esker's planning calendar and Bridgepoint's exit horizon.

Any of the five can stand alone. The Customer VoC plus Competitive Vendor Account Lists pairing is the most common starting point for an operator in Esker's position because it delivers the full picture (own customers + competitor customers) within 6 weeks. The Exit Readiness Snapshot is sequenced separately on Bridgepoint's planning calendar. The Sector Sponsorship can start as a one-off vertical study and convert to a recurring quarterly cadence if it earns its place in Esker's marketing motion. It is the longest-horizon move and the one that compounds the most. Happy to walk through scope, timeline, and pricing on any combination.