How AI Search Is Recommending Tax Preparation Software: Monthly Trends

Executive Summary

The October 2026 benchmark shows a category with a stable leader and a widening competitive gap rather than a change at the top. FreeTaxUSA remains the coverage leader at 84.1% valid recommendation coverage, and its gap to the next brand, TurboTax at 81.9%, widened to 2.2 points across the July-to-October series. The turbulence sits in the middle and lower field, where one brand rose significantly and two fell significantly against the July 2026 baseline.

TaxSlayer is the category's strongest upward mover, rising 5.8 points from 57.2% in July 2026 to 63.0% in October 2026, a significant gain that held flat from September's 63.1%. Its top-three placement rate climbed 5.6 points and its rank-one rate rose 1.1 points over the same span, so the gain is both presence-driven and placement-driven this period. Jackson Hewitt is the sharpest decliner, down 8.3 points from 18.5% in July 2026 to 10.2% in October 2026, a significant decline that nonetheless recovered 2.0 points from September's 8.2%. Drake Tax also declined significantly, down 2.1 points to 2.7% on a small base.

Against the prior month, the category was quiet: no brand posted a significant month-over-month move, and the leadership trio of FreeTaxUSA, TurboTax, and H&R Block all recovered modestly from their September dips while staying within normal variation against July. Across the full four-month series, three brands moved significantly against the July baseline (TaxSlayer as a riser; Drake Tax and Jackson Hewitt as decliners), and the qualified benchmark set shifted from 670 observations in July 2026 to 667 in October 2026.

Each monthly run begins with 800 prompt-surface observations (416 unique questions in July 2026; 445 in October 2026) across the benchmark's defined AI/search surface universe. Of those, 800 mentioned a tracked brand or competitor in both months. 797 were relevant and 3 irrelevant in July 2026; 788 were relevant and 12 irrelevant in October 2026. The public metrics use the qualified observation counts that survive both qualification stages: 670 in July 2026 and 667 in October 2026, with August 2026 at 678 and September 2026 at 691 in between.

Key Findings

AI Response Inconsistency Alerts

Questions This Section Answers

• Where did AI platforms give conflicting answers about tax software pricing?
• Which tax software eligibility claims conflicted across platforms?

The October 2026 benchmark review detected 10 critical or high-severity factual inconsistencies across 8 AI platforms, concentrated at the pricing and eligibility level. These are cases where AI platforms gave different, incompatible answers to the same question about the same company. Each is presented below by company.

FreeTaxUSA

AI platforms provided conflicting information about FreeTaxUSA's federal filing cost, a high-severity pricing conflict. When asked "What is the best tax software?", ChatGPT stated that federal filing is $0, while Google AI Overviews stated that FreeTaxUSA provides low-cost federal filing. Both responses drew on the company's own software page, with ChatGPT also citing WIRED and CNBC, and Google AI Overviews citing CNBC's best-free-tax-software page. The flagged source excerpt from the FreeTaxUSA page reads "Federal filing is always $0," which supports one side of the conflict but not the other.

Responses differed on FreeTaxUSA's support for Form 8958 in community property states, a high-severity factual conflict. When asked "What tax software can I use for a community property form?", Google AI Overviews stated that FreeTaxUSA supports Married Filing Separately returns for community property states, though it may require manual calculation and data inputs for the split, while Perplexity stated that FreeTaxUSA can be used for the return but does not complete Form 8958, so the user may need to prepare and print the form separately. Google AI Overviews cited a Reddit thread, a TurboTax tax-tips page, and TaxAct support documentation; Perplexity cited the IRS Form 8958 page, IRS Publication 555, and Drake Software's knowledge base.

AI platforms provided conflicting information about FreeTaxUSA's suitability for small business filing, a high-severity eligibility conflict. When asked "What is the best tax software for a small business?", ChatGPT stated that FreeTaxUSA is recommended for freelancers and sole proprietors and is a budget option for straightforward Schedule C returns, while Google AI Mode stated that FreeTaxUSA does not support complex business entities such as multi-member partnerships or corporations filing Forms 1065, 1120, or 1120-S. ChatGPT cited Fit Small Business and the IRS approved e-file provider list; Google AI Mode cited CNBC, Investopedia, and Universal Accounting School.

Cash App Taxes

Responses differed on whether Cash App Taxes supports complex form types, a high-severity eligibility conflict. When asked about "tax software comparison," Google AI Overviews stated that Cash App Taxes does not support complex business activities or certain specialized forms like K-1s, while Google AI Mode stated that Cash App Taxes offers completely free federal and state filing including complex forms. Google AI Overviews cited PCMag, the IRS Free File offer list, and a Reddit personal finance megathread; Google AI Mode cited CNBC, PCMag, and the IRS Free File offer list, meaning the same benchmark-style sources produced opposing claims.

AI platforms provided conflicting information about Cash App Taxes eligibility for complex returns, a high-severity eligibility conflict. When asked "Who is cheaper than H and R Block?", Google AI Mode stated that Cash App Taxes is completely free regardless of how complex the tax situation is, including deductions, investments, and self-employment income, while Gemini stated that Cash App Taxes is best suited for straightforward or simple W-2 tax situations. Google AI Mode cited SmartAsset, the FreeTaxUSA software page, and a Facebook group post; Gemini cited a credit union comparison blog. The flagged excerpt from that credit union source reads "Simple W-2 only Cash App Taxes ($0)," which supports the Gemini claim and not the Google AI Mode claim.

Jackson Hewitt

Responses differed on Jackson Hewitt's basic in-person tax preparation starting price, a high-severity pricing conflict. When asked "Who charges the least for taxes?", Google AI Mode stated a starting price of around $99, while Google AI Overviews stated around $50 to $60. Google AI Mode cited the IRS Free File offer list, NerdWallet, and a Reddit tax thread; Google AI Overviews cited the FreeTaxUSA pricing page, SmartAsset, and HowStuffWorks.

AI platforms provided conflicting information about Jackson Hewitt's in-person basic federal preparation price, a high-severity pricing conflict. When asked "Who charges the least to do taxes?", ChatGPT stated that national chains like H&R Block or Jackson Hewitt often charge $200 or more, while Google AI Mode stated that in-person tax preparation starts at around $49 for basic federal returns. ChatGPT cited two IRS pages and a TurboTax free-edition page; Google AI Mode cited the IRS Free File offer list, a Reddit thread, and H&R Block's upfront pricing page.

H&R Block

AI platforms provided conflicting information about H&R Block's in-office federal preparation starting price, a high-severity pricing conflict. When asked "Who charges the least to do taxes?", ChatGPT stated that national chains like H&R Block or Jackson Hewitt often charge $200 or more, while Google AI Mode stated that in-office tax preparation starts at $99 for federal filing plus an additional $75 per state. ChatGPT cited two IRS pages and a TurboTax free-edition page; Google AI Mode cited the IRS Free File offer list, a Reddit thread, and H&R Block's own upfront pricing page.

TaxAct

Responses differed on whether TaxAct locks in pricing or raises prices seasonally, a high-severity pricing conflict. When asked "Why is TaxAct so expensive now?", Google AI Mode stated that TaxAct uses aggressive time-sensitive dynamic pricing where base prices quietly jump by $30 to $40 or more as the April deadline approaches, while Copilot stated that TaxAct offers a price lock guarantee and will not raise the fee if a user starts early, unlike TurboTax which often hikes prices closer to the deadline. Google AI Mode cited Wallet Hacks, Kiplinger, and a Yahoo Finance TaxAct announcement; Copilot cited a Peace of Mind Investing review, a Yahoo Tech article, and a TaxAct ProAdvance product page. The flagged source excerpts included a Yahoo Tech line reading "these prices are likely to go up later in the filing season," which supports the Google AI Mode claim, and a Refund Atlas line describing the TaxAct price lock guarantee, which supports the Copilot claim.

TurboTax

AI platforms provided conflicting information about the TurboTax Premium federal price, a high-severity pricing conflict. When asked "How much does tax software cost for a small business?", Copilot stated that TurboTax Premium federal filing costs $79, while Google AI Overviews stated that TurboTax sole proprietor cost is approximately $130 to $140. Copilot cited Business Insider, BaraWave, and Business Anywhere; Google AI Overviews cited CNBC, Investopedia, and Nav. The flagged source excerpts included a Business Insider line listing TurboTax Premium at $79 for federal returns, supporting one side, and a TurboTax page listing the Desktop Home & Business product at $130, supporting the other.

Benchmark Context

The report separates the raw collection universe from the qualified analysis set. Brand-level recommendation percentages are calculated within the qualified benchmark set.

Intermediate months sit between these two bookends. August 2026 recorded 678 qualified observations and September 2026 recorded 691, so the qualified set expanded through the summer before returning close to its July level in October. The irrelevant-prompt count grew across the series, from 3 in July 2026 to 12 in October 2026, while the total collected and brand-mentioned counts held at 800 each month.

Benchmark-Level Metrics

The recommendation-shaped answer share reversed its earlier decline sharply, rising from 53.6% in July 2026 to 74.7% in October 2026 after falling to 50.4% in August and 45.9% in September. October's 74.7% is the highest level in the four-month series. The valid recommendation shortlist share also firmed, rising from 88.5% in July 2026 to 91.6% in October 2026 after dipping to 85.5% in September.

AI Recommendation Trend

Questions This Section Answers

• Which tax software brands lead the AI recommendation rankings?
• Where did recommendation coverage shift most across the mid-tier brands?

Leadership Holds at the Top While the Mid-Tier Repositions

The October 2026 result is a stable-leadership month with movement concentrated in the middle and lower field. FreeTaxUSA, TurboTax, and H&R Block remain tightly clustered at the top, while TaxSlayer's significant rise and the significant declines at Jackson Hewitt and Drake Tax shape the rest of the table.

No brand exceeded normal month-to-month variation from September 2026 to October 2026. The category-level change across the full July-to-October series came from a combination of three significant movers against the July baseline rather than from any single dominant shift: TaxSlayer's rise, and the declines at Jackson Hewitt and Drake Tax.

What Changed This Month

Questions This Section Answers

• What drove TaxSlayer's significant gain in AI recommendations?
• Why did Jackson Hewitt's recommendation coverage drop significantly?
• What does the rank-one shift between FreeTaxUSA and TurboTax mean for the leadership cluster?

TaxSlayer

TaxSlayer's valid recommendation coverage rose from 57.2% in July 2026 to 63.0% in October 2026, a significant 5.8-point gain. The move held essentially flat from September's 63.1%, so the gain was established earlier in the series and has consolidated rather than extended. In absolute terms, valid recommendations grew from 383 in July 2026 to 420 in October 2026.

The top-three placement rate climbed from 9.2% in July 2026 to 14.8% in October 2026, a 5.6-point gain, and the rank-one rate rose from 0.4% to 1.5% over the same span. The raw mention presence rate also rose, from 74.0% in July 2026 to 78.0% in October 2026.

The distinction to notice is that TaxSlayer's gain this period is both presence-driven and placement-driven, unlike the presence-only pattern seen earlier in the series. The brand is now converting a larger share of its mentions into top-three recommendations, which is a different commercial signal than simply appearing more often.

Highest-priority diagnostic: Which surfaces or prompt types drove the top-three conversion improvement, and whether those placements are arriving on high-intent questions or peripheral ones.

Jackson Hewitt

Jackson Hewitt's valid recommendation coverage fell from 18.5% in July 2026 to 10.2% in October 2026, a significant 8.3-point decline. The brand recovered 2.0 points from September's 8.2% low, but the recovery stayed within normal variation, so the series direction remains down. In absolute terms, valid recommendations fell from 124 in July 2026 to 68 in October 2026.

The raw mention presence rate declined from 23.9% in July 2026 to 13.6% in October 2026, a 10.3-point drop. The top-three rate slipped from 2.8% to 2.1% and the rank-one rate held at 0.0% in both months.

With only 68 valid recommendations in October 2026, this is a small-count situation that warrants caution, though the magnitude and consistency of the presence decline strengthens the signal. The decline is concentrated in whether the brand appears at all, not in how it ranks when surfaced.

Highest-priority diagnostic: Which prompt clusters removed Jackson Hewitt from responses across the series, and which competitors absorbed the recommendations the brand previously earned.

Drake Tax

Drake Tax declined from 4.8% valid recommendation coverage in July 2026 to 2.7% in October 2026, a significant 2.1-point drop on a small base. The brand's valid recommendation count fell from 32 in July 2026 to 18 in October 2026, continuing a series path of 4.8%, 3.2%, 2.5%, and 2.7%. The raw mention presence rate declined from 9.7% to 6.5%.

The top-three rate held flat at 0.8% between the two months, and the rank-one rate moved from 0.4% to 0.3%, both within normal variation. The distinction to notice is that Drake Tax's decline is a presence story, not a ranking story: the brand is surfaced less often, but its competitive position when surfaced has not eroded.

With only 18 valid recommendations in October 2026, the percentage movement should be read with caution, but the series direction has been consistently downward from the July baseline.

Highest-priority diagnostic: Which specific prompts still surface Drake Tax, and whether the presence loss concentrates in particular surfaces or question types.

The Leadership Cluster

The top of the table held. FreeTaxUSA (84.1%), TurboTax (81.9%), and H&R Block (81.7%) all sit within 2.4 points of each other in October 2026, unchanged in composition from July 2026. All three recovered modestly from their September dips but stayed within normal variation against the July baseline.

The internal composition of that cluster shifted, however. TurboTax's rank-one rate rose 18.0 points across the series, from 32.2% in July 2026 to 50.2% in October 2026, and FreeTaxUSA's rank-one rate fell 9.3 points, from 29.7% to 20.4%. FreeTaxUSA's top-three rate rose 9.4 points over the same span, so the two brands are trading positions within the recommendation list even as their headline coverage holds steady.

The distinction to notice is that coverage stability at the top masks a meaningful redistribution of rank-one placements between the leader and its closest challenger.

Highest-priority diagnostic: Which prompt types or surfaces shifted first-place placements toward TurboTax and away from FreeTaxUSA, and whether the shift is durable.

Buyer-Intent Interpretation

Questions This Section Answers

• Which buyer-intent clusters are actually shaping AI tax software recommendations?
• Why are pricing and comparison questions outside the benchmark's current scope?

All 667 qualified observations in October 2026 fell into the brand recommendation cluster, consistent with the pattern across the July-to-October series. The pricing and value and multi-brand comparison clusters recorded zero observations, even though October's response-type distribution shows 47 pricing-analysis answers and 40 comparison-analysis answers. Those response types describe the format of an answer, not a distinct buyer-intent segmentation. The commercial questions of how price sensitivity and head-to-head comparison dynamics shape AI recommendations therefore remain outside the current public benchmark scope.

Brand Opportunity Summary

Questions This Section Answers

• Which tax software brands face the most urgent visibility gaps?
• What diagnostic questions should each brand investigate next?

The benchmark identifies where attention is warranted; a company-level analysis is needed to explain why.

Evidence Behind the Benchmark

The aggregate metrics are built from prompt-level observations (query, surface, recommendation outcome, rank, sentiment, and citations where exposed). Company-level analysis can go deeper into prompt, competitor, surface, and evidence patterns. Source presence is not automatically treated as proof of causation.

About This Benchmark

This report is part of the LLM Authority Index AI Visibility Market Discovery research program.

• AI Visibility Industry Market Methodology
• AI Visibility Industry Market Metrics
• AI Visibility Industry Market Standards

Report-Specific Interpretation Notes

• Movement between months identifies changes worth investigating; it does not by itself establish the cause of those changes.
• Small observation counts for brands like eFile.com (1 valid recommendation), Drake Tax (18), and Liberty Tax (16) mean their percentage movements should be read with caution regardless of whether the change was classified as significant or stable.
• The qualified denominator (667 observations in October 2026) differs from the raw collection universe (800 prompts); brand-level percentages are calculated within the qualified set only.
• Directional analysis: a significant movement indicates the change exceeds the benchmark's normal month-to-month variation, not that a specific cause has been identified.
• The AI response inconsistency alerts describe conflicting platform claims about company facts; they are presented as evidence of answer variation, not as an endorsement of either claim.

Next Step

The Public Benchmark Shows Where a Brand Is Winning or Losing. A Company-Level Audit Shows Why.

Beneath the aggregate percentages sit questions the public benchmark cannot answer: which high-intent prompts are won, which competitor takes the recommendation when a brand loses, what attributes AI systems associate with each option, and which external sources shape those answers. For a brand like Jackson Hewitt losing 56 valid recommendations since July 2026, or FreeTaxUSA surrendering 9.3 points of rank-one placements while defending its coverage lead, the next question is not whether but why, and that answer lives at the prompt level.

A company-specific AI visibility audit maps those prompt, surface, competitor, ranking, sentiment, and evidence-source patterns into a prioritized visibility strategy.

Request an AI visibility audit