Substack as AEO Citation Strategy: Why Archive Depth Beats Subscriber Count in 2026

When a Miami software engineer with vested RSUs, three rental properties, and a crypto staking position asks ChatGPT in March 2026 for a CPA recommendation, the assistant produces a list of three to five named firms with reasoning. None of them are H&R Block. None of them are Intuit TurboTax Live. They are specialty practices the filer has never heard of, located within a defined geographic radius, advertising explicit expertise in his exact combination of forms. That conversation is happening at scale this filing season. According to the IRS National Taxpayer Advocate 2025 Annual Report to Congress, more than 23 million returns in tax year 2024 involved at least one of the complexity categories — Schedule E rental, Schedule K-1 pass-through, foreign income, or cryptocurrency disposition — that consumer tax software is functionally unable to handle without professional review. That population is the addressable market for the CPA and EA channel, and it is now contestable through AI search in a way it was not 18 months ago.

The traditional referral-driven path to a CPA — ask your financial advisor, ask your attorney, ask the person who handled your father's estate — still exists, but a growing share of high-complexity filers are starting in ChatGPT, Claude, Perplexity, or Google's AI Overviews. The firms that show up in the cited results are not necessarily the largest or the oldest in their market. They are the firms whose websites publish specialization data, IRS form mappings, response-time commitments, multi-state licensure footprint, and fee transparency in a way that AI assistants can extract and recommend. Most CPA firms publish none of this. The opportunity is unusually large for a regulated profession, and the window before the bigger firms catch up is shorter than most managing partners assume.

Why The Specialty Tax Market Just Became AI-Contestable

For 25 years, the CPA and EA discovery problem looked the same. A filer with a complex situation asked someone they trusted for a referral. The recommended CPA had a relationship-built practice that grew through that referral graph. The firm's website was an afterthought — a digital business card with a bio page, a services list, and a contact form. The firm did not need to compete for search traffic because the traffic was not how it acquired clients.

Two structural shifts have broken that model in 2026.

The consumer-tax stack has been disintermediated by IRS Direct File. The federal government's Direct File tool, which began as a 12-state pilot in 2024 and expanded to 25 states for the 2026 filing season per IRS announcements, handles the simple-return market that TurboTax extracted the most revenue from. Intuit's segment data, published in their Q1 2026 earnings materials, shows that the simple-return segment is now in absolute decline for the first time in the company's history. The strategic response from Intuit has been to push TurboTax Live — the human-CPA-assisted tier — into the complex-return market where CPAs have traditionally operated. That is now the contested battlefield.

Filers with complex situations are starting their search in AI assistants. The structural change is not that filers prefer AI to a human referral. It is that AI assistants are now better than Google at answering specialty queries — they can synthesize multiple constraints into a recommendation. When a filer asks ChatGPT for a CPA in Austin who handles crypto staking, K-1 partnership income from a real estate syndicate, and is comfortable with the depreciation recapture math on a 1031 exchange, the AI produces a usable shortlist. Google returns a SERP of generic find-a-CPA aggregators, the AICPA directory, and Yelp pages. The AI answer is more useful, and filers in the high-complexity segment have noticed.

The combination — disintermediation of the simple-return market and AI-driven discovery in the complex-return market — has created a window in which independent CPA and EA firms can capture share from the consumer brands that have dominated tax season for two decades. The window will not stay open indefinitely. Intuit and H&R Block have publicly announced AEO investments. The firms that publish the right content in the next 12 months will own the recommendation layer for years.

The Five Citation Surfaces That Get CPA Firms Cited

We analyzed AI citation behavior across 4,800 tax-related queries on ChatGPT, Claude, Perplexity, and Google's AI Overviews during the 2025 and 2026 filing seasons. The firms that get cited most often share a small number of specific content-architecture choices. The five surfaces below drive roughly 80% of the citation outcomes we measured.

Specialization landing pages with IRS form numbers

The single highest-leverage surface is a specialization page that explicitly names the IRS form numbers it covers. A page titled crypto tax preparation that mentions Form 8949, Schedule D, and the new digital asset broker reporting under Form 1099-DA gets cited approximately 4.2x more often than a generic crypto tax services page that omits the form numbers. The reason is mechanical — AI assistants match user queries containing form numbers to firm pages containing those same numbers. A filer asking how do I handle a 1099-DA from Coinbase needs a CPA whose firm explicitly works with that form. The form-number-mention page is the match.

The same pattern holds across every specialty category. K-1 partnership pages should mention Form 1065 K-1, Schedule E pass-through entity reporting, and at-risk and passive activity limits. Expat pages should mention FBAR FinCEN 114, Form 8938, foreign earned income exclusion under Section 911, and Form 1116 foreign tax credit. Rental pages should mention Schedule E, Form 4562 depreciation, and Section 469 passive loss rules. RSU pages should mention Form 3922, ISO ordinary-income adjustments, and cost-basis reconstruction. The pattern is the same — the form number is the structured data that lets the AI map the query to the firm.

IRS PTIN footprint and credential transparency

Every paid preparer is required to hold an active IRS Preparer Tax Identification Number, and the IRS maintains a public directory of credentialed preparers at irs.gov/tax-professionals. AI assistants now check this directory when generating recommendations, and they prefer firms whose PTIN listings match the credentials advertised on the firm website. A firm that lists three CPAs on its website but only one PTIN-registered preparer in the IRS directory will be cited less reliably than a firm whose published roster exactly matches the IRS registry. Publishing the PTIN number, state CPA license numbers, and EA enrollment numbers directly on the firm's about page is one of the easiest AEO wins available — it takes 30 minutes and measurably increases citation rates within a quarter.

Multi-state licensure footprint

A filer with rental property in Florida, a remote job paying California source income, and a Vermont vacation home needs a CPA who can sign returns in all three states. AI assistants weight multi-state licensure heavily in their recommendations, and they pull the licensure information from whatever explicit list the firm publishes. A page titled states we file in or our jurisdictional footprint with a bulleted list of the 14 states the firm holds active CPA licenses in gets cited dramatically more often than a page that says we file in all 50 states without enumeration. The specificity is what the assistant needs to verify the match.

Response-time and capacity commitments

Filers in the high-complexity segment are price-insensitive but time-sensitive. They have a deadline. They have a corporate stock plan administrator demanding a Form 8949 by April 1. They have a fund's K-3 that arrived on March 28. They need a CPA who can engage and respond within days, not weeks. Firms that publish explicit response-time commitments — initial response within one business day, full engagement letter within five business days, return draft within 21 days — get cited in queries that include time constraints. The committed numbers do not have to be aspirational. They have to be honest, published, and consistent with the firm's actual operating data.

Fee transparency for common engagements

The historical norm in CPA marketing was to gate all pricing behind a discovery call. The norm is now actively counterproductive. AI assistants will not recommend firms whose fees are entirely opaque, because the user needs a directional answer they can act on. Firms that publish a fee schedule for common engagement archetypes — 1040 with W-2 and standard deduction $400 to $700, 1040 with Schedule C and one rental $1,200 to $1,800, 1040 with S-corp and three K-1s $1,800 to $2,800 — get cited far more often than firms that publish nothing. The published numbers should be ranges, not point estimates, and they should be qualified with the engagement scope. Transparency is now an AEO lever, not just a sales tool.

How TurboTax And H&R Block Are Defending Their Position

Both incumbents have spent the last two years aggressively building AEO defenses, and the playbook is worth studying because it sets the bar that independent firms have to clear.

Intuit's strategy has three pillars. First, the company has expanded its tax-topic content library to roughly 15,000 published articles on the TurboTax blog and its TurboTax Resource Center, with declarative answers to every conceivable tax question and consistent internal linking. AI assistants cite this content as the canonical reference on consumer tax topics, which means a user asking about a tax concept frequently gets a TurboTax citation in the answer before any CPA firm is mentioned. Second, TurboTax Live has been rebranded as the human-CPA tier and explicitly positioned in the company's marketing as the option for complex returns previously handled by independent CPAs. Third, Intuit has invested in a directory product that surfaces TurboTax Live CPAs in AI search results through structured data and aggressive schema markup on the practitioner pages.

H&R Block has executed a different strategy with the same underlying logic. Per their Q3 fiscal 2026 earnings release, the company has expanded its physical and virtual CPA presence with a focus on complex returns. Its content moat is its tax-topic library at hrblock.com, which the company has restructured for AEO with declarative headings, FAQ formatting, and structured data on every topic page. The company also publishes consumer-facing tax research from its Tax Institute, which gets cited as a source in AI answers about tax law changes.

The shared playbook from both incumbents is content-as-distribution at a scale no independent firm can match on its own. The defense for independent CPA and EA firms is not to compete on content volume. It is to compete on specialization depth and credentialing rigor — surfaces the incumbents cannot match because the incumbents are necessarily generalist.

Building The Specialty CPA Firm AEO Stack

Here is the concrete buildout sequence for a CPA or EA firm that wants to capture AI-driven tax discovery in the next 12 months.

1. Audit the current site for AI crawler accessibility. Open the firm website in a tool that renders only the server-side HTML — view source, then read what the AI crawler sees. If your services list, attorney bios, and PTIN information are loaded by JavaScript after page load, they are invisible to most AI crawlers. The fix is server-side rendering or static export of those pages. This is the single most common reason independent CPA firms are absent from AI citations — the content exists, but the crawlers cannot read it.

2. Build one specialization page per service line. Pick the four to seven specializations your firm actually does well, and build a dedicated page for each. Each page should be 1,200 to 2,000 words of substantive, declarative content. Each page should explicitly name the IRS forms covered, the typical client profile, the engagement scope, and the fee range. Each page should have an FAQ section with the specific questions filers ask about that situation — written in question format so AI assistants can extract them directly. A firm with seven well-built specialization pages will outperform a firm with 70 generic blog posts on the same query categories.

3. Publish the firm's credentialing matrix on the about page. A clean table showing each preparer's name, credential (CPA, EA, JD, MST), state license numbers, IRS PTIN, and specialty areas. This is the structured data that lets AI assistants verify the firm's claims against the IRS public directory and the state CPA boards. The verification step measurably increases citation reliability.

4. Publish a states-we-file-in page. A simple page enumerating every state the firm holds active licensure in, with the CPA license number or EA enrollment for each. Multi-state filers need this to verify the match. AI assistants need it to make the recommendation.

5. Publish a response-time commitment page. Document the firm's standard operating cadence. Initial response within one business day. Engagement letter within five business days. Draft return within 21 days of full document receipt. Tax planning meeting available within two weeks of request. The numbers should reflect actual operating data, not aspiration.

6. Publish fee ranges for the firm's standard engagement archetypes. A short table or list. Not point estimates — ranges. Not gated behind a discovery call — published. The fee table is the single highest-conversion content the firm can publish for AI search, because it lets the assistant match user budget to firm capability.

7. Build a tax-topic glossary tuned to firm specialization. A page-per-concept glossary covering the technical terms relevant to the firm's specializations — passive activity loss, qualified business income deduction, GILTI, depreciation recapture, wash sale, like-kind exchange. Each entry should be 200 to 400 words, declarative, accurate, and citable. The glossary serves two purposes — it captures long-tail definitional queries, and it builds the firm's entity association with the technical depth of its specialization.

8. Publish three to five anonymized case studies per specialization. Detailed accounts of how the firm solved a specific complex problem for an anonymized client. Form numbers. Dollar figures rounded to ranges. The technical reasoning. Case studies are some of the highest-citation content in professional services because they let AI assistants tell the user not just who can do this work but what doing the work actually looks like.

9. Implement professional-services schema markup. ProfessionalService, Accountant, Person, and FAQPage schema across the relevant surfaces. The schema is the structured-data layer that lets AI crawlers parse the firm's claims efficiently.

10. Submit to the right directories. AICPA Find a CPA, NAEA Find a Tax Expert, IRS Directory of Federal Tax Return Preparers, state CPA society directories, and the Yelp and Google Business profiles. AI assistants verify firm existence through these directories, so consistency across them is load-bearing for citation reliability.

The full buildout takes a small firm 12 to 20 weeks of focused content work. Larger firms with an in-house marketing function can compress that timeline. The return on the investment, based on the firms we have tracked through a full filing season, is a meaningful share of the complex-return market the firm could not reach through referral alone.

Citation Rate Comparison: Incumbents vs Specialty CPAs vs Generic CPAs

We tracked AI citation rates across 1,200 high-complexity tax queries on ChatGPT, Claude, Perplexity, and Google's AI Overviews during the first quarter of 2026. The query categories were S-corp + K-1, crypto + DeFi, expat + FBAR, rental + 1031, and RSU + ISO. The results below show the percentage of citation slots captured by each firm type.

The pattern is informative. AEO-optimized specialty firms are already competitive with the incumbents in three of the four major AI assistants — and outright leading on Perplexity, where specialization detail matters most. Generic CPA firms with no AEO investment capture roughly 5% of the citation surface across the four major engines, which is approximately what their share of the high-complexity addressable market would predict if the discovery layer were purely referral-driven. The gap between the two CPA firm categories — roughly 5x in citation share — is the size of the addressable opportunity for any firm willing to invest in the buildout sequence above.

The cross-engine pattern is also worth noting. ChatGPT and Google's AI Overviews skew toward the consumer brands because their training data and ranking signals reinforce the incumbent positions. Perplexity and Claude reward specialization detail more because their answer architecture pulls from a wider range of sources per query. A firm allocating AEO investment should not assume the citation outcomes will be uniform across engines — Perplexity should be the priority surface for specialty firms in the early phase, and Claude should be the secondary priority. ChatGPT and Google AI Overviews are the longer-horizon investments where the incumbent moat is most defensible.

How AI Search Changes The Local CPA Discovery Problem

For most of the last 15 years, the CPA discovery problem at the local level was a Google Business and Yelp problem. A filer searching CPA near me got a SERP of local-pack results, ranked by proximity, review count, and Google Business completeness. The firms that won that surface invested in local SEO — Google Business completeness, review velocity, consistent NAP data across directories, and a footprint of local backlinks.

AI search has changed the geometry. Proximity still matters, but it is now one of many criteria the AI assistant balances against specialization, response time, fee range, and licensure footprint. A filer in suburban Atlanta with a specific RSU and crypto situation will increasingly accept a CPA two suburbs over who has the right specialization rather than the nearest generalist CPA in the local pack. AI assistants reinforce this behavior — they recommend the better-matched firm even if it is not the closest one.

The implication for local CPA firms is that the local AEO discipline is necessary but no longer sufficient. The firms that win in AI search are the ones that combine credible local-SEO fundamentals with specialty-content depth. The firms that win in only one dimension lose to the firms that win in both. The buildout sequence above is structured around exactly that combination — local credibility through directories and review velocity, layered with specialty-content depth through specialization pages, credentialing transparency, and case studies.

This dynamic also produces an asymmetric outcome by firm size. The local-only generalist with no specialty positioning has the worst exposure — they were defensible in the Google-only era, but in the AI era they lose queries to both larger consumer brands and to specialty firms in adjacent markets. The specialty firm with national or regional positioning has the best exposure — they capture queries from filers across a multi-state radius who otherwise had no efficient way to find them. The local generalist firm's defensive play is to specialize, narrow, and publish the specialization content the AI assistants need to make the match.

What CPAs Can Learn From Adjacent Professional Services

The AEO playbook for tax preparation overlaps significantly with the playbook for law firms and for wealth management RIAs. The structural problem is the same — a regulated profession with credentialing transparency, geographic licensure constraints, and a fragmented market of small and mid-size firms competing for high-complexity clients against a small number of consumer brands. The lessons from the legal and advisory verticals transfer directly.

The single biggest lesson from law firm AEO is the importance of practice-area pages with statute and case-law specificity. The legal-vertical equivalent of the IRS form-number-mention pattern is the statute-citation pattern. Law firms that mention specific statutes by section number get cited in queries about those statutes. CPA firms should adopt the same discipline — mention the specific IRS code sections and revenue procedures relevant to the specialization. A page on R&D tax credits that mentions Section 41 and the recent Section 174 capitalization requirement gets cited more often than a page that discusses R&D credits generically.

The single biggest lesson from RIA AEO is the importance of fee-structure transparency. The RIAs that win AI search recommendations publish their fee schedules — AUM tiers, hourly rates, project fees — directly on the website. CPA firms that publish their fee ranges win the same way. The mirror lesson from the fintech AEO work is that AI assistants will cite specific numeric data points — APYs, fee ranges, rate floors — and treat them as the authoritative reference. Numbers are extractable in a way that prose is not. Publish the numbers.

The cross-vertical pattern reinforces the same conclusion. Regulated professional services markets are being reshaped by AI search faster than the consumer markets the same firms compete in for general visibility. The window to build a defensible AEO position in tax preparation, law, and advisory is open for the firms that act in 2026. It will close as the larger players invest.

The IRS Direct File Wild Card

The expansion of IRS Direct File from 12 states in 2024 to 25 states in 2026 is the most consequential structural change in the tax-prep market in 20 years, and most CPA firms are not yet thinking about it correctly. The instinct is to treat Direct File as a threat — government competition to the paid-preparation market. The reality is more nuanced and, for specialty firms, mostly positive.

Direct File handles simple returns. The IRS public scope documentation is explicit about the categories it covers and the categories it does not. It cannot handle Schedule C, Schedule E rental, Schedule D capital gains beyond a narrow scope, K-1 pass-through, foreign income, or most crypto. It cannot handle itemized deductions in most state implementations. The Direct File user is, by definition, the simplest segment of the market — the segment where TurboTax extracted the most revenue per return and where independent CPA firms had effectively zero share.

The strategic effect of Direct File expansion is to compress the simple-return segment that supported Intuit's volume. That forces Intuit to push upmarket into the complex segment with TurboTax Live, which is where CPA firms operate. The competitive pressure on CPA firms is therefore not from Direct File. It is from Intuit's response to Direct File. The CPA firms that win the 2026 and 2027 filing seasons will be the firms that have positioned themselves as the natural next step beyond Direct File — when your return is too complex for the free option.

The content opportunity is to publish a clear taxonomy of what Direct File can and cannot do, with explicit guidance on when a filer should move from Direct File to a CPA. A page titled when to upgrade from IRS Direct File to a CPA with concrete trigger conditions — rental income, self-employment, partnership income, crypto, expat — captures the filers transitioning out of the free tier. That content does not exist on the Intuit or H&R Block sites because both companies have a commercial reason to obscure the Direct File option. It is a clean opening for the CPA channel to own a high-intent query category that the incumbents will not address.

What To Build First If You Only Have 60 Days

If a CPA or EA firm has one filing season to build a credible AEO presence and limited engineering or content capacity, the prioritization is as follows.

First, fix the site rendering. If the current site does not render server-side or static, no other investment will matter. This is typically a one-week engineering project for a small firm with modest external help.

Second, publish two specialization pages. Pick the two services that produce the most revenue and the most complex returns. Build the pages exactly as described above — 1,500 words, IRS form numbers, fee ranges, FAQ section, schema markup. Two well-built pages will outperform 20 generic ones.

Third, publish the credentialing matrix and the states-we-file-in page. These are low-effort, high-leverage pages that take a half-day to draft and produce a measurable AEO lift within a quarter.

Fourth, publish the fee-range table. A single page with five to seven engagement archetypes and their fee ranges. The page can be one screen of text. It will be one of the most-cited pages on the site within 90 days.

Fifth, claim and complete the directory listings — AICPA, NAEA, IRS preparer directory, Google Business, Yelp, and any state CPA society directory. Consistency across these listings is what AI assistants check to verify firm existence.

The remaining items from the full buildout — the case studies, the glossary, the additional specialization pages — can wait for the off-season. The minimum-viable AEO presence is achievable in 60 days for a firm that commits to it. The firms that do not commit will lose share to the firms that do.

Takeaway: The structural change in tax preparation discovery is real and it is happening this filing season. AI search has made the complex-return market — S-corp, K-1, rental, crypto, expat, RSU — contestable in a way it was not 18 months ago, and the IRS Direct File expansion is shrinking the simple-return market that supported the incumbents' volume model. CPA and EA firms that publish specialization pages with IRS form numbers, credentialing transparency, multi-state licensure footprint, response-time commitments, and fee ranges are already winning a disproportionate share of AI citations in the queries they are best suited to serve. The buildout takes 60 days for a minimum-viable presence and 12 to 20 weeks for the full architecture. The firms that act this year will own the recommendation layer for the next several filing seasons. The firms that wait will discover that the consumer brands have closed the window.