The client
Toku is a stablecoin payroll platform. Companies use it to pay employees and contractors in USDC and other stablecoins, on top of their existing HR stack: Workday, ADP, Rippling. Built for crypto-native companies, fintech CFOs, and any global team tired of waiting three days and losing 4% on wire fees.
We worked with Toku in 2024 on the redesign: 38 pages in three to four weeks, new positioning, messaging that turned vague "blockchain payroll" into "pay your global team in stablecoins, through your existing HR system." That case study lives on our site. It promised a marketing asset. This is the sequel that proves it.
The problem
Stablecoin payroll as a category is being defined right now. Not on Google's first page. Inside AI answers.
When a CFO at a Web3 company asks ChatGPT "how do I pay my team in USDC without replacing my payroll system," the model picks three to five vendors and renders them as the answer. Those vendors become the shortlist. Everyone else gets one chance to come up later as a "have you considered…" footnote.
Toku had the product, the customers, and the 2024 redesign behind them. What they didn't have in early 2026:
- a clear answer to which AI engines were surfacing them, on which prompts, against which competitors
- a content architecture sized for AI parsing, not just Google ranking
- a way to measure whether AI visibility was actually feeding pipeline
The category's established leaders already owned the generic EOR prompt cluster. Trying to take share from them on "best EOR for startups" or "best EOR for Brazil" was a fight Toku wasn't going to win. The crypto-payroll wedge was wide open: USDC payments, stablecoin payroll, crypto-native EOR, "Workday plus stablecoins."
The strategy
Same dual-track frame we run for every SEO/AEO client, retuned for B2B fintech. Every page has to rank on Google and earn AI citations. Different optimization, same content.
Four moves across the site:
- A long-form resources hub. Crypto payroll guide. Token vesting schedules. Pay-in-crypto tax mechanics. DUNA primer. Written for the actual buyer (finance lead, founder, HR), not for keyword density.
- A structured
/answersdirectory. Short, parseable answers to the questions that come up inside AI prompts: "how to handle probation periods globally," "are stablecoin payroll taxes different from fiat," and so on. Built so an LLM can pull a clean quote without wading through marketing copy. - A programmatic
/rates/{role}-{country}tree. Software-engineer-Portugal. Software-engineer-India. Software-engineer-Nigeria. Data-analyst-Portugal. Pages that target the exact compensation queries every cross-border employer Googles before making an offer. - An integrations directory. One page per HR stack Toku slots into. Rippling, ADP, Workday. The page that turns "can I keep my payroll software" into yes.
None of this is novel SEO. The unlock was sequencing. We built the AI citation surfaces first, because AI engines update their training and live indexes faster than Google, and a single citation in a ChatGPT answer compounds into branded Google search the same week.
The results
AI share of voice: climbing the leaderboard from near zero
30 days. 75 prompts. 12 brands tracked. 6,868 AI responses sampled across Google AI Overviews, ChatGPT, Perplexity, Claude, Gemini, and Grok. Tool: Peec AI. Toku entered the window with effectively zero share of voice. It has been climbing since.
| Brand | Visibility | Share of Voice | Avg Position |
|---|---|---|---|
| Competitor A (category leader) | 64% | 38% | 2.2 |
| Competitor B (category leader) | 62% | 31% | 2.3 |
| Competitor C | 32% | 11% | 3.5 |
| Toku | 18% | 8% | 2.3 |
| Competitor D | 16% | 6% | 2.5 |
| Competitor E | 10% | 3% | 4.7 |
| Competitor F | 5% | 2% | 3.9 |
| Competitor G | 2% | 1% | 4.4 |
Position 4 on a leaderboard Toku was effectively absent from a year ago. Read down the second column. When AI engines decide to mention Toku, they place it at 2.3 on average. Same average position as the two category leaders sitting above it. Those two have been in market for years and spend nine figures a year on awareness. Toku is the youngest name in the top half of the table. The line is still moving up.
The crypto-payroll prompt cluster: Toku owns it
Inside the wedge we targeted, the picture flips. On any prompt that includes "stablecoin," "USDC," "crypto payroll," or "token compensation," Toku is the answer the AI returns.
| Prompt | Toku Visibility | Avg Position |
|---|---|---|
| What are the best stablecoin payroll solutions for crypto and Web3 companies? | 86% | 2.4 |
| What payroll APIs allow integrating stablecoin payments into my product? | 84% | 1.8 |
| What global payroll software integrates with ADP and supports crypto payments? | 82% | 1.8 |
| Which EOR providers support stablecoin or crypto payroll? | 81% | 2.3 |
| Can I use Workday and still pay my team in stablecoins? | 60% | 1.8 |
| Best stablecoin payroll solution for EU companies | 59% | 2.8 |
| How do I pay employees in USDC without replacing my existing payroll system? | 51% | 1.8 |
| How do Web3 startups pay their global teams compliantly? | 47% | 1.6 |
Top-cited prompt in the category: Toku in 86% of AI responses, position 2.4. API prompt for fintech embedders: 84% at 1.8. Integration prompts for buyers with an existing HR stack: low 80s at sub-2.0. These aren't pageviews. These are the moment a buyer asks an AI for a recommendation and gets handed Toku.
The wedge insight: Toku is at 0% on generic EOR, by design
Same Peec AI dataset, different prompts: "best EOR for startups," "best EOR for Brazil," "best EOR for Argentina," "best global EOR providers," "best contractor management for SMBs." Toku visibility on all of them: 0%. The category's established leaders own those prompts.
Most agencies would treat that as a gap to fill. They'd write generic EOR comparison pages, try to outrank the leaders, and lose. We treated it as the line that defines the play. Toku competes on "global payroll with stablecoins," and nothing else. Every prompt that includes USDC, stablecoin, crypto, Web3, or token compensation: Toku shows up. Every prompt that doesn't: Toku is absent. The positioning is sharp. The AI engines mirror it back.
By AI engine: Google AI Overviews is the dominant surface
Most agencies optimize for ChatGPT because it's the surface they personally use. The actual distribution looks different.
| AI Engine | Toku Visibility | Toku Share of Voice | Avg Position |
|---|---|---|---|
| Google AI Overviews | 35% | 57% | 2.3 |
| ChatGPT | 11% | 23% | 2.6 |
| Perplexity | 9% | 20% | 1.9 |
Google AI Overviews carries 57% of Toku's total AI mentions and cites Toku in roughly one of every three responses on tracked prompts. This is SEO and AEO collapsing into one surface. The same Google query now returns an AI summary above the blue links, and whoever shows up in that summary inherits both the AI visibility and the click.
Page-level Google growth: the programmatic SEO is compounding
First month of GSC tracking (Feb 11 → Mar 12, 2026) versus the most recent month (April 10 → May 9, 2026). Every page in the new content architecture is up double or triple digits.
| Page | Feb–Mar | Apr–May | Change |
|---|---|---|---|
| /eor | 2 | 16 | +700% |
| /token-compensation-primer | 1 | 9 | +800% (position 14.2 → 6.9) |
| /rates/software-engineer-salary-hiring-rates-nigeria | 5 | 29 | +480% |
| /integrations/rippling | 5 | 15 | +200% (position 17.2 → 7.8) |
| /resources/crypto-payroll-guide | 4 | 11 | +175% |
| /resources/the-employers-guide-to-token-compensation | 15 | 29 | +93% |
| /rates/software-engineer-salary-hiring-rates-india | 0 | 17 | NEW |
| /rates/software-engineer-portugal | 0 | 10 | NEW |
The crypto payroll guide (the page ChatGPT cites for the broad "how does crypto payroll work" prompt) is up 175% on its Google clicks over the same window. That's the dual track working. The AI-citation surface pulls one stream of buyers. The Google ranking pulls a second. Same page, two flows of attention, compounding.
Branded search spillover: the AI flywheel feeds back into Google
When an AI engine names a brand, the buyer opens Google and types that brand name. Strongest indirect signal the AI work is landing. Five branded-variant queries grew sharply between February and April:
| Branded Query | Feb | April | Change |
|---|---|---|---|
| toku login | 1 | 8 | +700% |
| toku eor | 8 | 17 | +112% |
| toku web3 | 0 | 11 | NEW |
| toku token | 0 | 6 | NEW |
| toku app | 0 | 5 | NEW |
Nobody types "toku web3" into Google because they saw a billboard. They type it because they read a Reddit thread, watched a YouTube explainer, or, increasingly, asked an AI "what's the best crypto payroll vendor" and were handed Toku. The branded-search lift is the receipt the AEO work leaves behind.
The latent impression bank: AI Overviews is reading Toku's content even when it doesn't click
Four Toku resource pages are sitting on enormous GSC impression counts at very low CTRs:
| Page | Impressions (90d) | CTR |
|---|---|---|
| /resources/how-do-token-vesting-schedules-work | 86,648 | 0.03% |
| /resources/is-it-legal-to-pay-people-with-cryptocurrencies-stablecoins… | 81,435 | 0.17% |
| /resources/crypto-payroll-guide | 36,987 | 0.08% |
| /resources/duna-101-a-founders-guide-to-wyomings-dao-legal-framework | 33,422 | 0.20% |
That's the fingerprint of AI Overviews surfacing a page inside an answer panel without sending the click. Google's AI is reading Toku's content and quoting it inside the SERP. The buyer gets the answer without leaving Google. Bad for direct clicks. Excellent for brand authority. These pages are how Toku ends up in the AI's training and live-index pool.
The pipeline composition: majority-organic, named-domain B2B
On the conversion side we're keeping absolute numbers private. Those are Toku's competitive intel. What we will share is the shape. The majority of tracked B2B meetings booked over the last several weeks were first-touched by organic search or an AI surface. Not paid. Not outbound. Not partnerships.
The breakdown:
- Roughly 56% of first-touch attribution came from Google organic search
- Roughly 33% came in via direct or branded navigation (the spillover signal)
- The remaining ~11% split across Twitter/Grok-era social referrals, Brave Search, and crypto-ecosystem partner referrals
The booked meetings come from named-domain B2B buyers in the right segments: crypto-native infrastructure, EOR consultancies, traditional fintech CFOs. The exact ICP Toku has on file, surfaced through content that the AI engines now reliably cite for the questions those buyers ask.
What made this different
Most B2B SEO case studies stop at traffic. We pushed measurement further down the funnel: is the brand the answer an AI returns when a buyer asks the question that matters?
Three things made it work. The wedge: refusing to optimize for generic EOR prompts the category's leaders already owned, and going deep on the crypto-payroll cluster instead. The sequencing: building AI citation surfaces first (long-form resources, structured answers, programmatic compensation pages), not bolting AEO onto finished SEO content after the fact. The measurement discipline: Peec AI for AI visibility, GSC for Google ranking and impression growth, PostHog for first-touch lead attribution. If those datasets don't move together, the work isn't real.
The 2024 redesign laid the foundation. The 2026 results are what happens when that foundation gets eighteen months to compound, with an AEO-aware content program running on top of it.
Engagement: ongoing since 2024 · Dual-track SEO/AEO growth partnership · toku.com
