ASC 740 Income Tax Provision Engine
Done-for-you, CPA-signed corporate income tax provisions (current & deferred, ETR rate reconciliation, valuation allowance, uncertain tax positions, and the ASU 2023-09 / Pillar Two disclosure footnote) — delivered per entity, per reporting period, audit-ready, before the close deadline

1.Title

ASC 740 Income Tax Provision Engine — an AI-native, CPA-supervised tax-accounting service that delivers completed, signed, audit-ready income tax provisions (not provision software the client operates) for the public companies, private-equity portfolio companies, and pre-IPO businesses that must report income taxes under U.S. GAAP (FASB ASC 740) every quarter and year-end.

Slug: asc740-tax-provision-engine • This is the corporate income-tax accounting / tax provision business (FASB ASC 740, formerly FAS 109 / FIN 48), explicitly distinct from prior runs in the manifest — it is not the ASC 842 lease-accounting engine, not the R&D tax credit study, not cost segregation, not sales-tax nexus, not transfer-pricing documentation, and not SOX controls testing. It produces the tax-expense number and footnote that flow into the financial statements.

2.Final decision: Blueprint

3.Executive summary

Every company that issues GAAP financial statements must report income taxes under ASC 740: the total tax expense (current + deferred), deferred tax assets and liabilities with a valuation-allowance assessment, an effective-tax-rate reconciliation, uncertain tax positions under the FIN 48 recognize-and-measure model, and a detailed income-tax footnote Verified [21]. It is recomputed every interim quarter and at year-end, it feeds directly into reported earnings, and it is consistently one of the hardest, most error-prone areas of the close.

The work is already outsourced at scale — a stable category of "ASC 740 tax provision services" sold by the Big 4 plus RSM, CBIZ, Cherry Bekaert, Weaver, Wipfli, and Baker Tilly, because companies increasingly prefer to hand the obligation to a specialist rather than build and retain a full tax-accounting team Verified [1][2]. The downside of getting it wrong is severe and public: PetMed Express disclosed a FY2025 material weakness in its income-tax-provision review control [3]; ICON plc restated income tax expense by $13.1M (2024) and $6.6M (2023) [4]. Tax-related deficiencies are a persistent material-weakness category even under heightened PCAOB scrutiny Verified [5][7].

4.Thesis

The income tax provision is an excellent AI-native service unit because the atomic object — one legal entity (or consolidated group) for one reporting period — is a discrete, evidence-grounded computation: given the trial balance, the prior-period provision, the book-to-tax differences (Schedule M adjustments), the deferred inventory, and current tax law, compute current tax, roll forward deferred tax assets/liabilities, assess the valuation allowance, schedule uncertain tax positions, build the ETR rate reconciliation, generate the journal entries, and draft the footnote. An estimated ~70–80% of that work is structured extraction, deterministic calculation, and prior-period rollforward — precisely where frontier models plus a tax-calc engine excel Inferred [21]. The residual is genuine judgment — is a valuation allowance needed given the weight of positive/negative evidence, does a tax position meet more-likely-than-not, how does an OBBBA election or a Pillar Two top-up flow through — concentrated at a chokepoint owned and signed by a senior tax-accounting CPA.

Because the work is already outsourced, because GAAP reporters cannot avoid the obligation, and because the customer experiences the company as an expert service (they receive a finished, signed provision package and a footnote they can drop into the 10-Q/10-K), trust burden is low and the budget already exists Verified [1]. Pricing is naturally per entity, per reporting period (interim + annual), with modular add-ons — never hourly. As models improve at long-context document understanding and numerical reasoning, the autonomous share rises and cost-per-provision falls while the human concentrates on a shrinking judgment residual (Sam Altman test: pass). The durable asset is not a model — it is the provision operating system: the trial-balance and M-adjustment extraction templates, the maintained current-law tax-rule library (federal, 50 states, Pillar Two), the per-entity deferred-continuity ledger (each period builds on the last), the audit-tie-out QA that drives error toward zero, and the CPA sign-off that makes each package defensible to the external auditor and the SEC.

5.Discovery rationale

This run independently scanned the AI-native services terrain — tax/audit, financial-compliance back-office, healthcare administration, and regulated finance ops — screening against the six gates and the evidence threshold, and explicitly checking the 57 prior blueprints in the manifest for overlap. Discovery converged on the ASC 740 income tax provision for three reasons. First, demand is mandatory, recurring, and intensifying: every GAAP reporter must compute the provision every quarter and year-end, and 2025–2026 layered three rule changes on top of it at once — the One Big Beautiful Bill Act (signed July 4, 2025, recognized in Q3 2025 interims), the ASU 2023-09 expanded income-tax disclosures (effective for 2025 PBE annual periods), and the OECD Pillar Two global minimum tax Verified [10][12][14]. Second, the labor that does this work is disappearing: the U.S. CPA population has fallen to ~653k with exam sitters down 30%+ since 2016 and ~73-day fill times for credentialed roles — a structural shortage that pushes the obligation outward to specialists Verified [8][9]. Third — decisively — the incumbents are legacy human-analyst firms billing a scarce senior professional per engagement, while the work is ~70–80% deterministic, and the highest-intent prospects are public on EDGAR (companies disclosing a tax-related material weakness, an auditor change, or a recent IPO). The scoping is deliberate: this is the income tax provision (ASC 740) business, distinct from the prior ASC 842 lease engine and from every prior tax run.

6.Candidate comparison

Six AI-native service candidates were generated and scored (1–5, higher better) across demand evidence, gross-margin potential, MVP narrowness, licensing/regulatory safety, and whitespace vs. existing tools/vendors and the 57 prior blueprints. None of the prior blueprints duplicated the winner.

The winner leads on demand, margin, and MVP clarity. Its only soft spots — licensing nuance (independence/preparer boundary) and whitespace (an established outsourced category) — are precisely the seams this blueprint is designed around: a strict preparer-for-management scope that avoids attest/independence conflicts, and an AI cost structure plus EDGAR-driven distribution that legacy firms cannot match. ASC 718 stock-comp is retained as a natural second module, not a competing run.

7.CODE validation

C — Consumer / buyer trend

Two trends collide. (1) The CPA pipeline is collapsing. Licensed CPAs have fallen to ~653k (Aug 2025), exam sitters are down 30%+ since 2016, BLS projects ~120k accounting/audit openings/year, and 62% of finance leaders report hiring/retention difficulty — tax-accounting talent is the scarcest, most burnout-prone slice Verified [8][9]. (2) The work just got harder. OBBBA (bonus depreciation, §174 R&D, §163(j)) forces deferred-tax remeasurement and valuation-allowance re-scheduling; ASU 2023-09 expands the rate-reconciliation and jurisdictional disclosures; Pillar Two and CAMT add multinational interim volatility Verified [10][11][12][14]. More work, fewer people: the obligation flows outward.

O — Opportunity

The underserved problem: PE-backed portfolio companies, pre-IPO companies, and smaller SEC registrants that are too complex for spreadsheets but too small to staff a dedicated tax-accounting team, and whose audit firm is independence-barred from preparing the provision. Their alternatives are a Big 4 / national-firm provision team (capable but expensive and capacity-constrained) or fragile in-house spreadsheets (a leading material-weakness source). A faster, cheaper, equally-defensible AI-native service wins the long tail.

D — Demand

Buyers are visibly spending and procuring: an established "ASC 740 tax provision services" category sold by every national firm; companies publicly disclosing tax-provision material weaknesses (PetMed Express) and restating tax expense (ICON plc); advisory firms publishing "top 10 year-end provision pain points" content aimed at buyers actively struggling with the close Verified [1][3][4][19].

E — Economic sizing

The adjacent tax-management software market alone is ~$30.8B in 2025 growing ~11.5% to ~$73.5B by 2033 — context for how much is spent on the tax-compliance/reporting stack Verified (analyst est.) [16]. The service opportunity: there are several thousand SEC registrants plus a far larger population of PE-backed and pre-IPO companies needing a provision every period. A conservative serviceable wedge — ~5,000 PE-backed / smaller-registrant entities × ~$25–60k blended annual provision spend (interims + year-end + disclosures) ≈ $125–300M — captures a single niche, against a much larger national provision-services budget Inferred [1][2]. A small share of a market this large supports a meaningful, high-margin business. Figures are illustrative, not a forecast; per-engagement prices vary widely by complexity and entity count.

8.Rubric scorecard (six gates)

Total: 29 / 30. Anti-commoditization check (Gate 6 corollary): even if general models make a one-off calculation self-serve, the durable wins are the maintained multi-jurisdiction tax-rule library, the per-entity deferred-continuity ledger (each period depends on the last), the audit-tie-out QA and reviewer accountability, the relationships with audit firms who must rely on the workpapers, and the regulatory liability the buyer is paying to offload — none of which a raw model provides.

9.Target buyer

10.Jobs-to-be-Done

• Functional: "Compute this period's tax provision, give me the signed workpapers, journal entries, and a footnote I can drop into the 10-Q/10-K, before my close deadline."
• Risk/compliance: "Keep me out of a material weakness or restatement — make the provision survive my auditor's review and any SEC comment."
• Financial: "Get my ETR and deferreds right, capture the OBBBA elections and credits, and don't leave money or accuracy on the table."
• Emotional: "Stop the quarterly dread that the provision is the one number on the financials nobody truly stands behind."

11.Painful problem

The income tax provision is a perpetual, deadline-bound obligation that lands on the most scarce people in finance, every quarter and year-end. The mechanics are genuinely hard: bridge book income to taxable income through dozens of permanent and temporary differences, roll forward every deferred tax asset and liability, assess whether a valuation allowance is needed by weighing positive and negative evidence, recognize and measure uncertain tax positions under FIN 48's more-likely-than-not model, build an effective-tax-rate reconciliation that ties, and draft a footnote that satisfies the new ASU 2023-09 disaggregation — all under current law that changed three times in 2025–2026 (OBBBA, Pillar Two, CAMT) Verified [10][12][14][21]. Get it wrong and the consequences are public and expensive: income-tax errors drive roughly 12% of all restatements [6], tax provision review is a recurring material-weakness finding (PetMed Express, FY2025) [3], and restatements of tax expense run into the millions (ICON plc, $13.1M / $6.6M) [4]. Lean tax functions — squeezed by a CPA shortage that has cut the profession to ~653k and stretched fill times to ~73 days — cannot reliably absorb this, so the obligation gets outsourced, or it breaks Verified [8][9].

12.The outcome we sell

We do not sell a provision-calculator tool, a co-pilot, or a dashboard the client operates. The client experiences an expert tax-accounting service; the AI is the internal production engine. The provision number remains management's, adopted by management; we are the independent preparer behind it.

13.First one-feature MVP wedge

14.Evidence summary

• Mandatory, recurring demand — every GAAP reporter computes the provision each interim and year-end Verified [21].
• Already outsourced — a mature national-firm "ASC 740 provision services" category; companies prefer to offload rather than staff it Verified [1][2].
• High and public failure rate — ~12% of restatements are tax-related; live material-weakness and restatement disclosures Verified [3][4][6].
• Acute labor scarcity — CPA population ~653k, exam sitters −30%+, ~73-day fills Verified [8][9].
• Three concurrent 2025–26 rule changes raising work-per-provision — OBBBA, ASU 2023-09, Pillar Two/CAMT Verified [10][12][14].
• Deterministic share ~70–80% — Inferred from provision mechanics [21].
• Unit economics / pricing — per-engagement prices are custom and non-public — Inferred; carried as pilot kill-criteria.

15.Claim table (Verified / Inferred / Unverified)

16.Source-claim matrix

17.Market and demand evidence

Demand is structural and growing. The obligation is universal among GAAP reporters and recomputed every period [21]. The adjacent tax-management software market (~$30.8B in 2025, ~11.5% CAGR) signals the scale of spend on the tax-reporting stack that surrounds the provision [16]. Three concurrent rule changes — OBBBA, ASU 2023-09, and Pillar Two/CAMT — sharply raised work-per-provision in 2025–2026, exactly when the CPA labor pool contracted to ~653k with ~73-day fill times [8][9][10][12][14]. National firms market a dedicated "ASC 740 provision services" line, and advisory firms publish buyer-facing "year-end provision pain points" content — both signals of an active, budgeted buyer [1][2][19].

18.Active buyer conversations

• SEC disclosures (EDGAR): companies filing 8-K/10-K/20-F item-9A disclosures of income-tax-related material weaknesses and tax-expense restatements are publicly announcing pain — a directly targetable, high-intent list [3][4].
• Advisory pain content: "Top 10 pain points for the 2025 year-end income-tax provision" and quarterly "ASC 740 provision considerations" series are written because buyers are searching these terms during close [19][20].
• Recruiting market: the volume and duration of open senior tax-accounting roles is itself a demand signal — companies that cannot hire the work in must buy it as a service [9].
• Rule-change forums: CFO/controller communities and webinars on OBBBA, Pillar Two, and ASU 2023-09 adoption are full of "how do we actually do this provision now" questions [10][12].

19.Competitive landscape

20.Competitor and budget validation

Existing budget source: companies already pay national firms for outsourced/co-sourced provisions and license six-figure provision software — the line item exists [1][2][16]. Why alternatives are insufficient: national-firm teams price a scarce senior human per engagement and are capacity-constrained at exactly the busy-season peak; software shifts the work back onto the client's own missing experts; in-house spreadsheets are the leading material-weakness source [3][6][9]. Why we win: an AI-native cost base lets us deliver an equally-defensible, CPA-signed provision faster and cheaper, and an independent-preparer posture solves the auditor-independence problem the audit firm cannot. Not a clone: we are neither a software vendor (we sell the outcome, not a tool) nor a generic outsourced-accounting shop (we are a specialist ASC 740 production engine with a maintained rule library and reviewer accountability).

21.Pricing evidence and proposed pricing

Per-engagement provision prices are custom and non-public (firms quote by complexity and entity count) [1]. We price per entity, per reporting period — never hourly:

Pricing is outcome/per-unit, anchored to the deliverable (a signed provision), never to hours. Margin improves as the autonomous share rises and reviewer minutes fall. All figures Inferred and carried as pilot kill-criteria.

22.Regulatory and compliance considerations

The governing framework is FASB ASC 740 (income taxes), interpreted under U.S. GAAP, audited under PCAOB standards, and enforced for registrants by the SEC (item-9A ICFR, restatements). The 2025–2026 overlays — OBBBA, ASU 2023-09, Pillar Two, and CAMT — change the inputs but not the structure of the obligation [10][12][14]. Key compliance guardrails: (1) the provision is management's, prepared by us for management's adoption — we issue no audit/assurance opinion; (2) we are not the client's external auditor, so no auditor-independence conflict arises (this is in fact a tailwind — the audit firm is independence-barred from preparing the provision it audits); (3) all judgments (valuation allowance, UTP) are documented with evidence memos for auditor reliance; (4) data handling follows confidentiality and SOC-2-aligned controls given the sensitivity of pre-release financials (MNPI). No UPL/unauthorized-practice issue arises — tax-provision preparation is accounting/tax work, not legal practice or attest.

23.Licensing boundary

Pricing is per-deliverable, not contingency — no fee-sharing or success-fee structure that could create independence or ethics issues. The CPA sign-off is the trust and liability anchor.

24.AI-native advantage

This is AI-native, not "use ChatGPT." AI changes the economics: the deterministic ~70–80% (trial-balance ingestion, M-adjustment classification, deferred rollforward, multistate apportionment math, ETR rate-rec assembly, footnote drafting, tie-outs) runs at near-zero marginal cost, while the CPA's time concentrates on the judgment residual [21]. AI tasks: document extraction, book-tax difference classification, deferred continuity, rate-rec construction, ASU 2023-09 disaggregation, first-draft VA/UTP memos. Human tasks: VA realizability conclusion, UTP recognition/measurement, law-change elections, final sign-off. Deterministic rule tasks: statutory rates, apportionment formulas, FV/PV and rollforward math, disclosure templates. QC: tie-out checks, prior-period continuity, auditor-style review. Inputs: TB, prior workpapers/returns, fixed-asset detail, M-1/M-3, entity/state map. Outputs: signed provision, JEs, footnote, workpaper file. Never fully automated: the valuation-allowance and UTP conclusions and the final sign-off.

25.Internal AI engine architecture

26.AI-vs-human operations pipeline

27.Dynasty translation layer

• Buyer translation: CFO/VP Tax/Controller who must report income taxes and dreads a material weakness — wants a signed, audit-ready provision delivered on time.
• Service translation: done-for-you provision-as-a-service; AI does extraction/calc/rollforward/drafting; humans do VA/UTP/elections/sign-off; client receives the finished package.
• Workflow translation: intake → normalize → compute → review → QA → deliver → auditor support → return-to-provision true-up → renew next period.
• Tooling translation: secure portal/SFTP, a tax-calc engine + LLM extraction, a maintained rule library, a workpaper/document generator, a CRM; favor available tools over custom software at launch.
• Sales translation: "We deliver your ASC 740 provision — signed, audit-ready, before your close deadline — so your team stops fighting the spreadsheet and your auditor stops finding issues."
• Delivery translation: launch by delivering the first 3 provisions semi-manually (CPA + AI assist); automate extraction and rollforward first; harden as volume grows.
• Expansion translation: interims → consolidated/multi-entity → foreign/Pillar Two → ASC 718 stock-comp → return-to-provision automation → a maintained UTP/VA playbook product.

28.Anti-duplication analysis

What exists: national-firm provision teams, provision software (ONESOURCE, Bloomberg, CCH Tagetik, Corptax, Longview), and AI-native tax startups embedded in incumbents (Neo.Tax) [1][17][18]. Why this is not a copy: software sells a tool the client's scarce experts must operate; national firms sell expensive senior hours; we sell a finished, CPA-signed provision-as-a-service to the long tail those two under-serve. Narrow wedge: single-entity U.S. provisions for PE-backed/pre-IPO companies. Under-served segment: companies too complex for spreadsheets, too small for a dedicated tax-accounting team, and blocked from using their auditor. Unsolved pain: the judgment-plus-deadline burden that software doesn't remove. Differentiation: a maintained multi-jurisdiction rule library, a per-entity deferred-continuity ledger, an auditor-reliance-grade QA, and reviewer accountability. Vs. prior blueprints: distinct from the ASC 842 lease engine (leases ≠ income tax), R&D credit, cost segregation, sales-tax nexus, transfer pricing, and SOX testing — this produces the tax-expense number and footnote in the financial statements.

29.Anti-commoditization analysis

Sam Altman test (pass): better models raise the autonomous share and lower cost-per-provision, making the service stronger, not obsolete. If general models make a one-off calculation self-serve, the durable wins remain: (1) the maintained current-law rule library across federal, 50 states, and Pillar Two, updated as law changes (OBBBA-style shocks recur); (2) the per-entity deferred-continuity ledger — each period depends on the last, so the relationship compounds; (3) auditor-reliance-grade QA and reviewer accountability a raw model cannot provide; (4) the independent-preparer posture the audit firm cannot occupy; and (5) the regulatory liability the buyer is paying to offload. A CFO will not bet the 10-K tax line on an unsigned model output.

30.Service delivery workflow

1. Engagement letter (preparer-for-management scope) + secure portal provisioning.
2. Intake checklist: TB, prior provision/return, FA detail, M-1/M-3, debt/comp, entity/state map.
3. AI extraction + normalization; completeness check flags gaps to the client.
4. Compute current + deferred; roll forward; build rate rec; draft VA/UTP memos and footnote.
5. CPA review at the chokepoints (VA, UTP, elections); sign.
6. QA anti-error gate: continuity, tie-out, auditor-style red-team.
7. Deliver signed package + JEs + footnote + workpaper file before deadline.
8. Auditor support: respond to review questions; capture comments into the learning loop.
9. Return-to-provision true-up after the return files; renew next period.

31.Operations as product

The operation is the product: variance is engineered out via standardized intake checklists, required-evidence lists, automated completeness checks, exception queues, reviewer-assignment logic, confidence scoring on extracted data, audit trails, version control of the rule library, gold-standard provision exemplars, red-team checks against auditor findings, customer-ready output templates, root-cause analysis on any auditor adjustment, and a postmortem loop that turns each correction into an SOP, rule, prompt, retrieval source, or QA check. The goal: every provision cheaper, faster, and more consistent than the last.

32.No-holes quality engine

• Continuity check: current-period opening balances must tie to prior-period closing (deferreds, NOLs, UTP).
• Tie-out: provision-to-financials and provision-to-return reconciliations must foot and cross-foot.
• Rate-rec gate: the ETR reconciliation must reconcile to the statutory rate within tolerance.
• Evidence completeness: VA and UTP conclusions require a documented evidence memo before sign-off.
• Red-team: an independent reviewer (or a second model pass) attacks the provision as an auditor would.
• Confidence scoring: low-confidence extractions route to human review before they enter the calc.
• Postmortem: any auditor adjustment triggers root-cause → new rule/QA check.

33.What the human expert actually does

Minutes are Inferred launch targets, carried as pilot kill-criteria; the thesis is cohort-over-cohort compression as the rule library and QA mature.

34.Minimum viable offer

"Your year-end ASC 740 provision, signed and audit-ready, before your close date — for one entity, fixed fee." Includes current/deferred, deferred rollforward, ETR rate rec, valuation-allowance memo, journal entries, and an ASU 2023-09-compliant footnote, with CPA sign-off and auditor support through review. One feature, one ICP (PE-backed single-entity), one outcome (a clean provision), priced per period.

35.Fulfillment process

First 3 customers fulfilled semi-manually: a senior CPA drives the engagement while the AI assist handles extraction, rollforward, and first-draft memos/footnote. Day-one tools: secure portal/SFTP, an LLM extraction pipeline, a spreadsheet/calc engine, a document generator, and a maintained rule library. Automate extraction and deferred rollforward first (highest repeatable volume); keep VA/UTP human at launch. Evolve into templates, SOPs, and a software-assisted operation as volume and pattern coverage grow.

36.Tools and systems

• Intake/security: SOC-2-aligned portal/SFTP; MNPI controls; access logging.
• Engine: LLM extraction + a deterministic tax-calc layer (spreadsheet engine or provision calc); model-agnostic prompts.
• Knowledge: versioned federal/state/foreign rate & rule library; OBBBA / Pillar Two / ASU 2023-09 templates.
• Workpapers: document generator for memos, footnote, JEs, tie-out files; per-entity continuity ledger.
• Ops: CRM, exception queue, reviewer-assignment, QA dashboard, audit-trail store.

37.Human-in-the-loop quality control

Two human gates: (1) data-gap resolution before computation, and (2) CPA judgment-and-sign-off at VA/UTP/elections before delivery. Between them sit deterministic QA gates (continuity, tie-out, rate-rec, red-team). Low-confidence extractions and any auditor-adjustment patterns route back through review. The CPA sign-off is the irreducible trust and liability anchor — never automated away.

38.Nonlinear scaling and unit economics

Revenue scales faster than headcount because the deterministic majority of each provision is automated and a CPA touches only the residual. COGS per provision: model inference + document processing (low and falling), data storage/hosting, human review minutes (the dominant cost — compressing cohort-over-cohort), CPA sign-off minutes, QA, support, and rework. Track every line from day one; do not bury labor in "operations." Targets: throughput rising per CPA as the rule library matures; cycle time compressing; rework, quality-failure, and escalation rates falling; margin expanding with volume as fixed rule-library cost amortizes. Funnel assumptions (Inferred, pilot-tested): diagnostic→pilot and pilot→paid conversions and high period-over-period retention (the provision is recurring and switching mid-relationship is costly). Note rising tax salaries (+3.7%) make the in-house alternative more expensive, widening our value gap [22].

39.Distribution proof table

40.Sales and outreach plan

Lead with a diagnosis, not a demo. For EDGAR-triggered prospects, send a short, specific memo referencing their disclosure and the remediation path. For PE firms, pitch portfolio standardization. For auditors, pitch independence-clean reliance. Convert via a free provision diagnostic → scoped fixed-fee pilot (one entity, one period) → full-cycle subscription. Three layers below.

41.Founder-led content plan

Publish expert teaching content that builds trust with controllers and VP Tax: how to do the OBBBA-affected provision, what ASU 2023-09 actually requires, the valuation-allowance evidence framework, the top causes of tax-provision material weaknesses, Pillar Two for interim provisions, and worked rate-reconciliation teardowns. High-performing organic posts become paid-ad creative later.

42.First 30 days of content

10 educational posts: (1) OBBBA's deferred-tax remeasurement, step by step; (2) ASU 2023-09 rate-rec disaggregation explained; (3) when you actually need a valuation allowance; (4) FIN 48 UTP in plain English; (5) Pillar Two for interim provisions; (6) the 5 most common provision material weaknesses; (7) interim ETR method pitfalls; (8) return-to-provision true-ups; (9) why your auditor can't prepare your provision; (10) first-time IPO provisions. 3 diagnostic teardowns: a redacted rate-rec that doesn't tie; a deferred rollforward with a continuity break; a footnote missing 2023-09 disaggregation. 2 lead-magnet angles: "Provision Material-Weakness Self-Scan" checklist; "OBBBA + Pillar Two Provision Readiness" assessment. 1 webinar: live "Year-end ASC 740 under OBBBA, Pillar Two, and ASU 2023-09." 1 outbound diagnosis template: EDGAR-triggered remediation memo.

43.Lead magnet and waitlist plan

Lead magnet: a free Provision Readiness / Material-Weakness Self-Scan — the prospect answers a short intake and uploads a redacted prior provision; we return a diagnostic flagging continuity breaks, untied rate recs, missing 2023-09 disclosures, and VA/UTP documentation gaps. It captures the exact pain signal, demonstrates competence, and routes straight to a scoped pilot. Waitlist: "Reserve a year-end provision slot" (capacity is genuinely finite at close) creates urgency and a qualified pipeline.

44.Warm GTM plan

Work diagnostic users, waitlist members, webinar attendees, and existing CFO/controller relationships first. Offer a free provision review and a scoped one-entity pilot. Prioritize anyone who has disclosed a material weakness or is mid-OBBBA/Pillar Two adoption — they have a deadline and a budget.

45.Targeted outbound plan

Build a perfect-fit list from EDGAR (item-9A tax material weaknesses, tax-expense restatements, auditor changes, recent S-1s) and PE portfolio rosters. Personalize each touch around the specific disclosure or trigger and lead with an opportunity/diagnosis memo, never a generic demo ask.

46.Answer-engine / search visibility plan

Publish authoritative, structured answers to the questions buyers ask AI assistants and search engines during close — "how does OBBBA affect my tax provision," "what does ASU 2023-09 require," "do I need a valuation allowance," "can my auditor prepare my provision." Optimize for ChatGPT, Perplexity, Gemini, and Google so the company surfaces when a controller researches the problem at deadline.

47.Pilot design and early-demand trap mitigation

First cohort capped at 5 paying pilots, all matching the narrow ICP (PE-backed, single-entity, U.S. federal + multistate). Fixed-fee, one period. The trap to avoid: taking unlimited custom/complex provisions that don't generalize. Mitigation: a strict eligibility filter, a standard intake, and a rule that any out-of-scope complexity is logged as a future module — not absorbed as bespoke labor.

48.Early-access feedback flywheel

Daily standup on pilot exceptions; weekly review of every auditor comment. Product feedback = recurring extraction misses, rule gaps, QA holes, or template defects → become SOPs, rules, prompts, retrieval sources, or QA checks. Custom work = a one-off client quirk → logged, not systematized. Anything that caused an auditor adjustment must be fixed before the next cohort.

49.Build-before-scale checkpoints

• After 5 pilots: harden intake, required-evidence lists, and QA tie-out checks.
• After 10: harden SOPs, exception queues, reviewer checklists, and delivery templates.
• After 20: pause new pilots until COGS, rework, escalation, and cycle time are measured. Acceptable temporary workaround: manual rollforward for unusual deferreds. Unacceptable (model-not-scalable signal): every provision needs bespoke CPA rebuild with no reusable pattern.

50.7-day launch plan

• Days 1–2: define the single ICP and one-entity offer; stand up portal + engagement letter (preparer-for-management scope).
• Days 3–4: build the intake checklist, extraction pipeline, and a gold-standard provision exemplar.
• Day 5: publish the lead magnet (self-scan) and 2 teaching posts.
• Days 6–7: pull an EDGAR trigger list; send 25 personalized diagnosis memos; book diagnostics.

51.30-day launch plan

• Run 10–20 free diagnostics; convert 3–5 into fixed-fee one-entity pilots.
• Deliver first provisions semi-manually with CPA + AI assist; capture every auditor comment.
• Publish the full first-30-days content set; run one webinar.
• Codify v1 SOPs, rule library, and QA gates from pilot learnings.

52.90-day launch plan

• Reach 5–10 paying provisions; measure reviewer minutes, rework, cycle time, gross margin.
• Harden intake/SOPs/QA per build-before-scale checkpoints; launch interim + ASU 2023-09 modules.
• Sign 1–2 PE-portfolio or fractional-CFO channel partnerships.
• Decide go/expand vs. fix based on margin and audit-pass evidence.

53.Metrics and KPIs

54.Risks and mitigations

The provision is high-stakes and judgment-laden; the central risks are accuracy/liability, incumbent and software commoditization, the difficulty of compressing reviewer minutes, and data sensitivity (MNPI). Each is addressed in the risk register below with a concrete mitigation.

55.Exhaustive risk register

56.What could kill this

Three things. (1) Reviewer minutes never compress — if every provision needs near-full bespoke CPA effort, the margin thesis fails (measured by 20 pilots). (2) Incumbents/software fully bundle AI provision-as-a-service before we accrue the continuity-ledger and auditor-relationship moat. (3) A high-profile accuracy failure destroys trust in a business whose product is defensibility. Each has a designed mitigation, but these are the real failure modes.

57.Go/no-go reasoning

GO. The candidate clears every evidence-threshold criterion: identified buyer, painful recurring problem, proven existing spend, active demand, competitor/budget validation, a credible win rationale, a narrow MVP wedge, a service-first delivery path, no fatal regulatory blocker, a credible 50%+ gross-margin path, and a believable, data-advantaged distribution path. The rubric scores 29/30 and the Sam Altman test passes.

58.Final recommendation

59.Source list

1. Cherry Bekaert — Tax Provision Services for ASC 740 Compliance
2. Weaver — ASC 740 Tax Provision
3. SEC EDGAR — PetMed Express Form 8-K (FY2025) — income-tax-provision material weakness
4. SEC EDGAR — ICON plc Form 20-F (FY2025) — income-tax-expense restatement
5. Moss Adams — Trends in Public Company Material Weaknesses
6. AuditUpdate — Material Weaknesses Are Increasing and the Accountant Shortage May Be to Blame
7. BDO — Four Reasons to Strengthen Income Tax Accounting Internal Controls
8. Ramp (citing NASBA/AICPA) — The Accountant Shortage: Causes, Impacts & Solutions
9. Talentfoot (Robert Half data) — How the CPA Shortage Is Extending Time-to-Fill
10. Baker Tilly — Impacts of the OBBBA under ASC 740
11. RSM US — Accounting for the Income Tax Impacts of the OBBBA
12. RSM US — ASU 2023-09: Expanded Income Tax Disclosure Requirements
13. PwC — FASB Issues Guidance on Income Tax Disclosures (In-Depth 2023-12)
14. EY — Accounting Considerations for the Global Minimum Tax (Pillar Two)
15. RSM US — OECD's 'Side-by-Side' Package (Jan 2026)
16. Straits Research — Tax Management Software Market Size & Growth
17. Bloomberg Tax — AI for Corporate Tax: Verified Intelligence
18. Thomson Reuters — AI in Corporate Tax: Key Trends, Use Cases, and What's Next
19. Baker Tilly — Top 10 Pain Points for the 2025 Year-End Income Tax Provision
20. RSM US — ASC 740: Q3 2025 Provision Considerations
21. Bloomberg Tax — How to Calculate the ASC 740 Tax Provision
22. Robert Half — The Accountant Shortage Will Persist in 2026