The business. A done-for-you service that recovers and prevents B2B card interchange leakage for mid-market merchants that accept commercial and purchasing cards. Visa's Commercial Enhanced Data Program (CEDP) retired the legacy Level 2/Level 3 incentive structure; validation began 17 October 2025 and Level 2 fully sunset in April 2026, so any B2B merchant whose billing system is not emitting genuine, validated line-item (Level 3) data is now being automatically downgraded to standard rates — a 0.45–0.60 percentage-point increase on B2B card volume. Verified S3S5S7
What we sell. Not software the merchant operates. We sell a measured outcome: a lower effective processing rate and recovered/avoided interchange, delivered as an expert-run managed service. AI is the internal engine — it parses processor statements, classifies every transaction against the current interchange rulebook, detects downgrades, quantifies leakage, and generates a specific data-remediation plan. A payments expert validates the plan at the judgment chokepoint and oversees implementation. We then monitor CEDP verification status monthly so savings persist.
Why now / why AI-native. The CEDP transition is a live, dated trigger event forcing thousands of merchants off their old rates simultaneously. Processor statements are messy, inconsistent PDFs; interchange has hundreds of categories; CEDP explicitly rejects autofilled/generic data. This is exactly the "high intelligence threshold, low physical labor, regulation-shaped moat" terrain where frontier models plus a human chokepoint decouple revenue from headcount.
Decision: BLUEPRINT — GO.
Mid-market B2B merchants leak 0.45–1.5 percentage points of interchange because their billing/ERP/gateway stack does not emit CEDP-compliant Level 3 data, and the October 2025 → April 2026 CEDP transition just made that leakage larger, automatic, and continuously re-validated by Visa's own machine-learning checks. The leakage is invisible on a P&L (buried inside a blended "merchant fees" line), recurring, and expensive. It is a document-and-data problem with expert judgment at a few chokepoints — the ideal shape for an AI-native managed service that sells the recovered dollars, not a dashboard.
The winning company is processor-agnostic (never resells processing), lands with a free read-only diagnostic, prices on realized savings plus a monitoring retainer, and builds a durable moat in the data-remediation and continuous-verification layer that processors and one-time audit shops leave untouched.
This run inventoried 106 prior blueprint slugs in the manifest (healthcare RCM, tax credits, customs recovery, immigration, IP, construction, ERISA, securities, environmental, government contracting, and numerous compliance/recovery engines). None addressed card-acceptance interchange. Independent web research surfaced a rare, dated regulatory trigger — Visa's CEDP retiring Level 2/3 with validation from Oct 17 2025 and Level 2 sunset in April 2026 S2S3 — creating acute, current, and quantified pain for a large buyer population that already spends money on the problem (payments advisors, statement-audit shops). The category clears the evidence threshold and is novel versus prior output, so it was selected over parcel/freight audit and generic merchant-fee audit (both more commoditized and not tied to a fresh trigger event).
At least five candidates were generated and scored (1–5, higher is better) on demand evidence, whitespace/novelty, margin potential, regulatory moat, MVP clarity, and speed to first revenue.
| Candidate | Demand | Whitespace | Margin | Moat | MVP | Speed | Total |
|---|---|---|---|---|---|---|---|
| B2B Interchange / CEDP Remediation Engine (winner) | 5 | 4 | 5 | 4 | 5 | 5 | 28 |
| Parcel & freight invoice audit / refund recovery | 5 | 2 | 4 | 2 | 4 | 4 | 21 |
| Generic merchant statement fee audit (processor-switch) | 4 | 1 | 3 | 1 | 3 | 4 | 16 |
| Telecom / SaaS expense management audit | 4 | 2 | 4 | 2 | 3 | 3 | 18 |
| No Surprises Act provider-directory verification | 3 | 4 | 4 | 4 | 3 | 3 | 21 |
The winner uniquely pairs a dated trigger event (CEDP) with a clean recovery+prevention outcome, strong existing budget, and a remediation moat that resists commoditization.
Visa's CEDP restructured commercial interchange (validation from Oct 17 2025; Level 2 sunset April 2026), and Visa now validates each transaction with ML and monitors merchants' data-quality consistency. Commercial/virtual-card B2B volume is growing (~$662B US virtual-card volume in 2025, +25% YoY). More card volume + stricter data rules = larger, automatic leakage. Verified S3S4S14
Most mid-market merchants accept commercial cards through billing/ERP/gateway stacks that emit incomplete or autofilled data, which CEDP explicitly disqualifies. The old "your processor handles it" assumption is now false; processors pass through downgrades and a 0.05% participation fee. The remediation work (making genuine line-item data flow) is unowned by processors and skipped by one-time audit shops. Verified S6S17
An established advisory/audit industry already sells against this pain on contingency (Redbridge ~20% recurring savings; P3/Cost Analysts, WeAudit, Optimized Payments). Processors, gateways, and consultancies are publishing CEDP explainers because merchants are actively asking. Existing category budget is being spent today. Verified S11S13S8
US interchange ~$130–145B/yr; interchange is 70–95% of acceptance cost. If B2B/commercial card volume is even a low-hundreds-of-billions and mid-market merchants leak 0.45–1.5pp, the recoverable/avoidable pool per year is plausibly in the billions. A single mid-market merchant with $20M annual B2B card volume leaking 0.6pp loses ~$120K/yr; at a 25–30% success fee that is $30–36K of first-year revenue from one account. Inferred from S5S20
| Gate | Score | Rationale |
|---|---|---|
| 1 — Low trust burden | 4/5 | Fee audit/optimization is already commonly outsourced on contingency; buyer wants the result, not the process. Slight trust drag from the category's spammy reputation — mitigated by processor-agnostic positioning. |
| 2 — Low task-level judgment | 4/5 | Statement parsing, transaction classification, and downgrade detection are decomposable and mostly automatable; true judgment (remediation strategy, processor negotiation) is concentrated in a few steps. |
| 3 — High intelligence threshold | 5/5 | Hundreds of interchange categories, inconsistent statement formats, and CEDP data rules require synthesis across rulebook + client data + gateway capabilities. |
| 4 — Regulation as moat | 4/5 | Network rules (CEDP), PCI scope, and validated-data requirements raise buyer WTP and deter casual entrants; it is network-rule, not statutory, so slightly softer than a licensed field. |
| 5 — No physical labor | 5/5 | Entirely data/document/workflow based; delivered remotely. |
| 6 — Sam Altman test | 4/5 | Better models = cheaper parsing/classification and faster remediation specs. Anti-commoditization rests on the integration/monitoring layer (below). |
Total: 26/30. Clears the bar strongly.
ICP: US mid-market B2B merchants with $5M–$100M annual card volume where a meaningful share is commercial/purchasing/corporate cards — distributors and wholesalers, industrial/MRO suppliers, manufacturers, dental/medical/vet supply, building materials, and B2B SaaS/services billing large tickets. Economic buyer: CFO / VP Finance / Controller / Treasury. Champion: AR / billing manager or payments/ops lead who feels the "merchant fees keep rising" pain. Why reachable: concentrated in trade associations, ERP/AR user communities, and LinkedIn finance titles.
B2B interchange leakage is invisible, recurring, and now automatic. Card fees appear as a single blended line; the merchant cannot see that transactions are downgrading to standard rates because their data is incomplete. CEDP made this worse: from Oct 17 2025 Visa validates the data before granting lower rates, and from April 2026 Level 2 is gone entirely, so merchants who "used to qualify" are now paying 0.45–0.60pp more with no notice. Verified S3S5S7 Autofilled or generic data is explicitly disqualified, so the fix is a genuine data-pipeline remediation, not a checkbox. Verified S17
A lower, defended effective processing rate: (1) a recovered/avoided-interchange number verified on settled statements, (2) a remediated data pipeline that qualifies B2B transactions for the best available (CEDP Level 3) rates, and (3) monthly monitoring that keeps the merchant "verified" and flags re-leakage. The merchant experiences an expert who makes their card costs go down and stay down — not a tool they must run.
| ICP | US distributor/wholesaler, $10–50M annual card volume, high commercial-card mix |
|---|---|
| Trigger event | Post-April-2026 CEDP downgrade; "our merchant fees went up and we don't know why" |
| Pain | 0.45–1.5pp interchange leakage, invisible on the P&L, now automatic |
| One-feature MVP | The Interchange Leakage Diagnostic: upload 3 months of processor statements → dollar-quantified leakage report + top-5 remediation fixes |
| Input | 3 recent merchant processing statements (PDF/CSV) + processor/gateway name |
| Output | Effective-rate baseline, category-level downgrade map, annualized leakage $, prioritized remediation plan, CEDP-readiness score |
| Human chokepoint | Payments expert validates classifications & savings range before the report is released |
| Success metric | Diagnostic → paid engagement conversion ≥ 25%; median identified leakage ≥ $40K/yr |
| What they ask for next | "Do the remediation for us" and "monitor it every month" — the recovery + monitoring managed service |
The core claims — interchange scale, the CEDP timeline, the size of the downgrade penalty, the disqualification of autofilled data, ML-based validation, and the existence of a contingency-priced advisory market — are each supported by multiple 2025 sources (vendor, processor, consultancy, and trade press). Vendor-reported "we find savings 90–99% of the time" figures are labeled Inferred (marketing self-report). Precise B2B-only interchange leakage pool is Inferred from unit math, not a single published figure, and is presented as a range.
| Claim | Label | Source | Type | Date | Confidence | Used in |
|---|---|---|---|---|---|---|
| US card processing fees totaled ~$185B in 2025; interchange is 75-85% of that (~$130-145B). | Verified | S6, S20 | Processor/consultancy | 2025 | High | Market & demand |
| Level 2/Level 3 data unlocks ~0.5%-1.5% interchange savings on B2B card transactions. | Verified | S5, S9 | Vendor explainers | 2025 | High | Outcome / pricing |
| Visa retired Level 2/3 and launched the Commercial Enhanced Data Program (CEDP); data validation began Oct 17, 2025. | Verified | S2, S3, S4 | Vendor/consulting | 2025 | High | Trigger event |
| Visa sunsets Level 2 entirely in April 2026; CEDP-compliant Level 3 data becomes the only path to reduced interchange. | Verified | S3, S7 | Vendor/processor | 2025 | High | Trigger event |
| Failure to submit full Level 3/CEDP data triggers automatic downgrade to standard rates — a 0.45-0.60 pp cost increase on B2B volume. | Verified | S5, S7 | Vendor explainers | 2025 | High | Painful problem |
| Autofilled / generic / default data does not qualify for CEDP incentives; genuine line-item data from the billing system is required. | Verified | S3, S17 | Processor docs | 2025 | High | AI-native advantage |
| Visa Commercial Solutions validates each transaction with machine learning and classifies merchants as verified / non-verified with ongoing monitoring. | Verified | S3, S4 | Vendor/consulting | 2025 | High | Ops as product |
| A 0.05% CEDP participation fee applies to qualifying transactions and may be passed to merchants. | Verified | S6, S7 | Processor explainers | 2025 | Medium | Unit economics |
| Independent payments advisors report ~20% average recurring savings on acceptance cost. | Verified | S13 | Consulting advisory | 2025 | Medium | Competitor / budget |
| Interchange audit firms report finding savings on 90-99% of merchant statements reviewed. | Inferred | S11, S12 | Competitor claims | 2025 | Low | Demand (vendor-reported) |
| Performance-based / share-of-savings pricing is an established, legal engagement model in payments consulting. | Verified | S11, S12, S13 | Competitor/consulting | 2025 | High | Pricing legality |
| US virtual-card B2B volume ~$662B in 2025 (up 25% YoY); commercial card acceptance is growing. | Verified | S14, S15 | Trade press/issuer | 2025 | Medium | Market growth |
| Merchant processing statements are inconsistent, processor-specific formats where fees hide as interchange, assessments, or markup. | Verified | S19, S20 | Competitor/consultancy | 2025 | High | AI-native advantage |
| Downgrade review is commonly triggered when non-qualified/standard categories exceed ~3% of volume. | Inferred | S19 | Competitor site | 2025 | Low | Ops as product |
| Interchange represents 70-95% of total card acceptance cost for merchants. | Verified | S20 | Payments consultancy | 2025 | High | Market |
The Claim Table above is the source-claim matrix: every material claim carries a Verified / Inferred / Unverified label, its supporting source IDs (mapped to the Source List), source type, date, confidence, and the section where it is used. No material claim relies on an unlabeled source.
US merchants paid ~$185B in card fees in 2025, with interchange the largest component at 75–85% (~$130–145B) S6S20. Interchange is 70–95% of a merchant's total acceptance cost S20. Commercial/virtual-card B2B usage is growing fast (US virtual-card volume ~$662B in 2025, +25% YoY; 70% of US corporations had adopted virtual cards by 2024) S14S15, expanding the interchange base to which leakage applies. The Parcel/Freight audit-and-payment market (an analog "recover leakage from complex invoices" category) is growing ~8% CAGR, evidencing durable demand for outsourced spend-recovery services generally S1.
Demand is visibly active: dozens of processors, gateways, and consultancies published 2025 CEDP explainers (Finix, Versapay, Redbridge, Stripe, NMI, Priority Commerce, Wind River) precisely because merchants are asking what changed and what to do S2S3S17S18. Optimized Payments published "early observations" on real Product 3 rate changes post-CEDP S8. Merchant-fee audit shops run active contingency outreach S11S12. These are money-and-labor demand signals, not mere trend chatter.
Independent payments advisors (Redbridge, Optimized Payments): sophisticated, enterprise-focused, strategy-led; strong but priced/scoped for large enterprises, thin on hands-on data remediation for mid-market. S8S13 Merchant-fee audit shops (P3/Cost Analysts, WeAudit, Merchant Fee Savers, Apropos): contingency audits, but many monetize by switching the merchant's processor (channel conflict / reputation risk) and treat data remediation as out of scope. S11S12S19 Gateways/processors (Stripe, PayTrace, NMI): offer Level 3 auto-population for merchants on their modern rails, but do not help the long tail on legacy/ERP-originated flows and have no incentive to minimize their own interchange pass-through. S17S18
Budget already exists and is being spent on contingency and retainer models: Redbridge reports ~20% average recurring acceptance-cost savings S13; audit shops operate 100%-contingency engagements S11. The buyer is not being asked to create a new budget line — they are being offered a better, processor-agnostic, remediation-inclusive alternative to tools and consultants they already pay. Existing alternatives are insufficient because enterprise advisors ignore mid-market, and audit shops stop at the report (or steer a processor switch) rather than making genuine Level 3/CEDP data flow and monitoring it.
Evidence: performance-based / share-of-verified-savings pricing is standard and legal in payments consulting S11S13. Proposed model (per-unit / outcome, never hourly):
This is payments/network-rule consulting, not a licensed profession. Key obligations: PCI DSS scope (never store PAN; work from tokenized/aggregated statement data), card-network rules (accurate representation of CEDP; no facilitation of prohibited surcharging/miscoding), and data-protection (DPAs, least-privilege access to statements and gateway configs). Share-of-savings pricing is permissible; contracts must define "realized savings" and the measurement baseline to avoid disputes. No unauthorized practice of law/tax/finance is involved — we do not give tax, legal, or investment advice.
| Layer | Boundary |
|---|---|
| AI system may | Parse statements; classify transactions; detect downgrades; quantify leakage; draft remediation specs and monitoring alerts. |
| Trained operator may | Validate classifications; confirm savings ranges; produce client reports; coordinate remediation with client IT/gateway. |
| Payments expert must | Approve savings estimates before release; own processor/gateway negotiation strategy; sign off remediation plans. |
| Company must NOT | Resell/broker processing for undisclosed residuals; guarantee specific rates; give tax/legal/investment advice; touch or store cardholder PAN; advise miscoding to obtain rates the transaction doesn't qualify for. |
| Required controls | DPAs, PCI-aware handling, audit logs, disclosure of independence/processor-agnostic status, documented savings baseline. |
Merchant statements are heterogeneous, processor-specific PDFs where fees hide across interchange, assessments, and markup S19S20; interchange has hundreds of categories; CEDP disqualifies generic/autofilled data S17. AI changes the economics: it ingests any statement format, normalizes it, classifies every transaction against a current interchange rulebook, flags downgrades, and drafts a remediation spec in minutes rather than an analyst-day. AI tasks: extraction, normalization, classification, downgrade detection, leakage quantification, drafting. Human tasks: validate edge cases, set remediation strategy, negotiate. Deterministic rules: interchange rulebook + CEDP field requirements. Never fully automated: savings sign-off and any client-facing dollar promise.
Human minutes are spent only where judgment or client-facing dollar claims live; everything upstream is AI/automation.
1. Buyer translation: CFO/controller of a mid-market B2B merchant; urgent problem = silently rising card fees post-CEDP; wanted outcome = lower, defended effective rate.
2. Service translation: done-for-you recovery + remediation + monitoring. Customer receives dollars saved and a maintained low rate; automation does parsing/classification; humans validate and negotiate.
3. Workflow translation: intake statements → AI analysis → expert validation → diagnostic report → remediation with client IT/gateway → verify realized savings → monthly monitoring → renewal.
4. Tooling translation: statement-ingestion + OCR, canonical data store, interchange rulebook, LLM workbench, report generator, monitoring dashboard, CRM, e-sign, DPA templates. Favor off-the-shelf before custom.
5. Sales translation: "Your merchant fees quietly jumped when Visa changed the B2B rules in 2025–26. We'll show you exactly how much you're leaking — free — and then fix it and keep it fixed. You pay a share of what we actually save."
6. Delivery translation: first 3 clients delivered semi-manually (analyst + expert) with the AI workbench; monitoring is a monthly re-run. Automate ingestion and monitoring first.
7. Expansion translation: productize per-processor remediation playbooks and per-vertical packages; add Mastercard/Amex OptBlue programs; graduate monitoring into a light merchant-facing portal (support layer only, not the product).
Existing: enterprise payments advisors, contingency fee-audit shops, and gateway Level-3 auto-population. Why not a copy: we are processor-agnostic (no processor-switch monetization), mid-market focused, and we do the data remediation and monthly CEDP-verification monitoring that advisors scope out and audit shops skip. Under-served segment: mid-market merchants on legacy/ERP-originated card flows too small for Redbridge and poorly served by switch-driven shops. Unsolved manual pain: making genuine Level 3 data actually flow and stay compliant month over month. Differentiation: the remediation + continuous-verification operating system, not a one-time PDF audit.
Sam Altman test: better models make statement parsing and classification cheaper and faster — good for our COGS. If general models let merchants self-serve the audit: the durable value is not the audit; it is (1) implementing the remediation across a merchant's specific ERP/gateway stack, (2) negotiating with processors, and (3) continuously keeping the merchant "verified" under Visa's ongoing ML monitoring as rate schedules and data rules change. Those are operational, relationship, and freshness moats a generic chatbot cannot hold. We compound a proprietary library of processor statement templates, realized-savings priors, and per-vertical remediation playbooks.
Variance is the enemy. We standardize: SOP per processor statement format; required-evidence checklist at intake; automated completeness checks; exception queue for low-confidence classifications; reviewer-assignment logic by account size; confidence scoring on every transaction; audit trails and versioned rulebook; gold-standard example library; red-team check on any savings estimate above a threshold; customer-ready report templates; and root-cause + postmortem on any account where realized savings miss modeled savings by >20%.
Every unit passes: (1) effective-rate reconciliation (sum of parsed fees must tie to statement totals), (2) CEDP field-requirement check, (3) confidence-threshold gate routing edge cases to human review, (4) conservative savings-range banding, (5) second-reviewer sign-off for accounts >$50K modeled savings, and (6) realized-vs-modeled tracking that feeds the learning loop. No dollar figure reaches a client without expert approval.
| Task | License | Min/unit @ launch | Min/unit @ day 90 | Automation path | Quality risk | Cannot automate | Audit trail |
|---|---|---|---|---|---|---|---|
| Validate classifications / edge cases | None | 45 | 15 | Higher-confidence models + rulebook tuning | High | Novel/ambiguous categories | Reviewer log + confidence scores |
| Approve savings range | None | 20 | 10 | Calibrated priors from realized data | High | Final dollar promise | Signed estimate record |
| Remediation strategy per stack | None | 60 | 30 | Per-processor playbooks | Medium | Non-standard ERP setups | Remediation runbook version |
| Processor/gateway negotiation | None | 40 | 30 | Templates + scripts | Medium | Relationship judgment | Comms log |
| Monthly monitoring review | None | 20 | 8 | Auto-alerts; review exceptions only | Low | Interpreting new schedules | Monitoring alert log |
"The Interchange Leakage Diagnostic + Fix." Free read-only diagnostic quantifies leakage; paid engagement recovers/remediates and monitors. Delivered by one analyst + one payments expert on the AI workbench — no custom platform required to start.
First 3 clients: intake statements via secure link → run AI workbench → expert validates → deliver diagnostic → close engagement → remediate with client IT → verify realized savings → begin monthly monitoring. Day-one tools only; automate ingestion + monitoring after the first cohort.
Secure upload/portal; OCR + LLM workbench; canonical data store (Postgres/Sheets initially); interchange rulebook (versioned docs/rules); report generator (HTML→PDF); monitoring scheduler; CRM (HubSpot); e-sign; DPA templates; analytics for realized-savings tracking. Off-the-shelf first; custom only where volume justifies it.
Confidence scoring routes only uncertain items to humans; mandatory expert sign-off on savings; second reviewer for large accounts; realized-vs-modeled feedback recalibrates priors; postmortems convert misses into rulebook and SOP updates.
COGS per diagnostic (launch): model inference + OCR ~$3–10; ~85 human minutes (validate+approve) at loaded cost. Automation: ~55% at launch → ~75% at 90 days → ~85% at 1 year as parsing/classification harden. Throughput: 2–3 diagnostics/expert/day at launch → 6–8 at day 90. Cycle time: 3–5 business days launch → <48h at maturity. Rework target: <8%. Quality-failure target: <3% (savings miss >20%). Escalation: <15% of transactions to human. Margin expansion: monitoring retainer is near-pure-margin recurring revenue that grows the blended margin as the book compounds. Funnel assumptions: lead-magnet (diagnostic) → paid ~25%; pilot → paid ~40%; annual monitoring retention ~85%. Success fee on realized savings decouples revenue from headcount because a $120K-leakage account yields ~$30K with the same ~85 minutes of expert time as a $20K-leakage account.
| Channel | Why ICP reachable | First angle | Conversion assumption | Proof source | Measurement | Follow-up |
|---|---|---|---|---|---|---|
| LinkedIn (finance titles) | CFO/controller titles filterable by industry+size | "Visa's CEDP quietly raised your B2B card costs — here's how much" | 1–2% connect→diagnostic | Active CEDP content demand S2 | Diagnostic bookings | ROI memo + call |
| Trade associations (distribution/wholesale) | Members share fee-pain; newsletters/webinars | Webinar: "The 2026 B2B card-fee reset" | 5–10% attendee→diagnostic | Category budget exists S13 | Webinar→diagnostic rate | Recorded diagnostic offer |
| Answer-engine/SEO | Merchants search "why did my merchant fees go up 2026 / CEDP" | Definitive CEDP remediation guides | Organic → diagnostic | Explainer proliferation S3S18 | Organic diagnostic signups | Nurture sequence |
| ERP/AR communities (NetSuite, SAP B1, Acumatica) | Level 3 data originates in ERP; users feel the gap | "Make your ERP emit CEDP Level 3 data" | Niche but high-intent | Data disqualification rule S17 | Community→diagnostic | Integration checklist |
| Referral partners (fractional CFOs, ERP VARs) | They see merchant fees across many clients | Rev-share on referred diagnostics | High-trust | Advisory market S8 | Partner-sourced deals | Co-branded reports |
| Targeted outbound email | Firmographic lists of B2B merchants | Personalized leakage estimate | Standard cold-email math | Contingency market S11 | Reply→diagnostic | Diagnosis-first memo |
Diagnosis-first, never demo-first. Lead with a specific, dollarized leakage estimate; convert to the free read-only diagnostic; deliver an expert-validated report; propose recovery+remediation+monitoring on realized-savings pricing. Land with recovery, expand into the monitoring retainer.
Expert-led education on the exact pain: what CEDP changed, how downgrades happen, why "your processor handles it" is now false, the cost of doing nothing, and how to read a merchant statement. Repurpose top organic posts as paid-ad creative later.
10 educational posts: (1) What Visa CEDP changed for B2B; (2) The hidden 0.45–0.60pp downgrade tax; (3) How to read your merchant statement; (4) Level 3 fields your ERP probably isn't sending; (5) Why "autofilled" data fails CEDP; (6) The 0.05% participation fee explained; (7) Verified vs non-verified merchant status; (8) 5 signs you're leaking interchange; (9) Processor-agnostic vs switch-driven audits; (10) What monthly monitoring actually catches. 3 diagnostic teardown formats: anonymized statement teardown video; before/after effective-rate table; category-level downgrade heatmap. 2 lead-magnet angles: "Interchange Leakage Diagnostic"; "CEDP-Readiness Scorecard." 1 webinar: "The 2026 B2B Card-Fee Reset: recover what you're leaking." 1 outbound diagnosis template: personalized "we estimate $X/yr leakage on your volume — here's the free diagnostic."
Lead magnet: free Interchange Leakage Diagnostic (upload 3 statements → dollarized report). Waitlist: "CEDP-Readiness Scorecard" for merchants not ready to share statements — captures stack details and volume, scores readiness, and routes to the diagnostic. Both capture the pain signal (leakage size) and qualify sales-readiness (volume + commercial-card mix + downgrade evidence).
Work diagnostic and scorecard users first: any merchant whose diagnostic shows ≥$40K leakage gets a same-week expert call with a realized-savings proposal. Re-engage waitlist scorecard leads with their readiness gaps.
Build firmographic lists (B2B distributors/wholesalers/manufacturers, $5–100M revenue). Personalize each touch with an estimated leakage range from public signals (industry commercial-card mix). Lead with a diagnosis memo, not a demo ask.
Publish the definitive, citable CEDP remediation library so ChatGPT/Perplexity/Google surface the company when merchants research "why did my merchant fees increase 2026", "Visa CEDP Level 3", or "B2B interchange optimization." Structure content as clear Q&A with schema markup; earn citations from processor/ERP communities.
First cohort cap: 8 merchants. Learning objectives: parser coverage across top processors, realized-vs-modeled savings accuracy, remediation cycle time, and which manual steps recur. Early-demand trap: resist bespoke consulting; every fix must become a reusable SOP/playbook. Decline accounts requiring one-off custom work outside the wedge.
Weekly review of realized-vs-modeled savings and every human-corrected classification. Product feedback = recurring classification errors, missing processor templates, remediation steps that repeat. Custom work = one-off asks unique to a single client. Corrections become rulebook entries, new statement templates, remediation playbooks, QA checks, and calibrated savings priors before pilots expand.
After 5 clients: harden intake, required-evidence checklist, and QA reconciliation. After 10: harden per-processor SOPs, exception queues, reviewer checklists, and report templates. After 20: pause new pilots until COGS, rework rate, escalation rate, and realized-vs-modeled accuracy are measured and within target. Acceptable temporary manual steps: new-processor template building. Unacceptable: recurring per-client custom remediation that never becomes a playbook.
Stand up secure upload + DPA; assemble interchange rulebook v1 and top-10 processor statement templates; build the diagnostic report template; wire the AI workbench; publish 3 cornerstone CEDP posts + diagnostic landing page; line up 5 warm merchant contacts for free diagnostics.
Deliver 5–8 free diagnostics; convert 2–3 to paid recovery+remediation; run 1 webinar; publish the 30-day content set; sign 2 referral partners; codify SOPs from the first diagnostics; measure diagnostic→paid conversion.
10–15 paid engagements; first monitoring retainers live; parser coverage across top-15 processors; realized-savings verified on early accounts; automation at ~75%; publish 3 anonymized case teardowns; begin outbound at scale; hit build-before-scale checkpoints before expanding cohort.
Diagnostic→paid conversion (≥25%); median identified leakage (≥$40K); realized-vs-modeled savings accuracy (≥80%); cycle time; automation %; rework rate (<8%); escalation rate (<15%); monitoring retention (≥85%); revenue per FTE (≥$300K); gross margin (55–70%); CAC payback (<90 days).
The dominant risks are commoditization (gateways auto-populating Level 3), non-realization of modeled savings, category reputation, and network-rule changes. Each is mitigated by owning the remediation+monitoring layer, billing on realized savings, processor-agnostic positioning, and a monitoring model that adapts to schedule changes. See the full register below.
Description. Visa/Mastercard further restructure commercial interchange, compressing the savings pool.
Mitigation. Diversify across Visa + Mastercard + Amex OptBlue programs; sell remediation & monitoring (durable) not just one-time recovery; reprice quickly.
Description. Gateways (Stripe, PayTrace, NMI) auto-populate Level 3 fields, shrinking the gap.
Mitigation. Target the long tail of merchants on legacy stacks and ERP-originated card payments where auto-Level-3 fails; own the ERP/AR data-remediation layer gateways won't touch.
Description. Merchant-services 'audit' shops are associated with churn-and-burn processor switching and spammy outreach.
Mitigation. Never resell processing; processor-agnostic; publish methodology; contingency only on verified, client-confirmed savings; no bait-and-switch.
Description. Recommended remediation is not implemented by client IT, so modeled savings never materialize.
Mitigation. Do-it-for-them integration support; bill on realized (settled-statement-verified) savings, not modeled; 60-90 day realization gate before invoicing.
Description. Handling merchant statements, processing data, and gateway credentials creates breach and PCI-scope risk.
Mitigation. SOC 2 path; least-privilege read-only access; never store PAN; tokenized/aggregated data only; signed DPAs.
Description. Client claims savings were coincidental (rate cards changed anyway), disputing the fee.
Mitigation. Baseline effective-rate lock at intake; documented before/after category-level attribution; third-party statement as source of truth.
Description. AI mis-classifies an interchange category, over-promising savings.
Mitigation. Deterministic interchange rulebook + human expert sign-off at chokepoint; confidence scoring; conservative savings ranges.
Description. Treasury/finance slow to approve access and remediation.
Mitigation. Free diagnostic (read-only statement scan) as low-friction entry; ROI memo aimed at CFO; land-with-recovery, expand-to-monitoring.
Description. Parsing coverage limited to a handful of processor statement formats.
Mitigation. Build a parser library covering top 15 processors first (covers majority of mid-market); template-driven onboarding of new formats.
Description. Interchange expertise concentrated in one or two reviewers.
Mitigation. Codify rulebook + SOPs + gold-standard cases; cross-train; escalation queue; reviewer redundancy.
Description. Litigation/regulation (e.g., swipe-fee settlements) alters interchange schedules mid-engagement.
Mitigation. Monitoring model adapts to new schedules; contracts reference 'then-current' rates; reprice on schedule changes.
Description. Merchant takes the recovery and cancels before monitoring revenue accrues.
Mitigation. Bundle recovery + 12-month monitoring in one agreement; demonstrate ongoing downgrade leakage to justify retainer.
(1) Networks/processors make CEDP Level 3 qualification automatic and universal, collapsing the leakage pool. (2) A trusted incumbent bundles free remediation into every gateway. (3) The category's spam reputation poisons top-of-funnel trust faster than content can rebuild it. Each is survivable if the company anchors on the durable monitoring relationship and mid-market/legacy-stack long tail rather than one-time recovery.
Clears the evidence threshold: identified buyer (mid-market B2B merchant CFO/controller), specific painful problem (automatic post-CEDP interchange downgrade), evidence the problem exists and is spent-against (contingency advisory market), active demand (CEDP explainer proliferation), competitor/budget validation, a credible reason to win (remediation + monitoring moat), a narrow one-feature MVP, a practical path to first sale (free diagnostic), service-first delivery without a large platform, no fatal regulatory blocker (network-rule consulting, share-of-savings legal), and a credible path to 50%+ gross margin via a recurring monitoring book.
GO. Build the B2B Interchange Optimization & CEDP Remediation Engine as a diagnostic-led, processor-agnostic managed service. Land with an expert-validated leakage diagnostic, convert to realized-savings recovery + data remediation, and compound margin through a monthly CEDP-verification monitoring retainer. Prioritize the remediation and continuous-verification layer as the defensible core.