Agent One Research
Why Agent One is doubling down on agencies first.
Agent One could speak to agencies, SMB owners, real estate agents, or solopreneurs. The data points to a cleaner wedge: agencies can turn today's tool and website-readiness demand into repeatable client services.
Published by Agent One Team.
Decision
Primary wedge: agent-as-marketing for agencies. SMB and vertical pages remain the client-example layer agencies can sell, show, and turn into approved public proof.
Pick one segment: agencies.
Agency ranks #1 of the four named segments by paid/trial threads: 70 agency, 43 real estate, 32 SMB/local, 25 solopreneur.
Raw usage caveat.
Raw usage is not agency-led. The largest raw usage buckets are automation_workflow and developer_ai_coding, so public copy must describe agencies as the packaging wedge, not as the largest observed customer segment.
Organic audience.
Current organic acquisition is coming through the homepage, site checker, agent builder, and system-prompt generator rather than through segment pages. Those paths should bridge into agency audits, discovery agents, and automation-intake starters.
Fresh Search Console read
Search clicks still flow through the homepage and agency-ready tools.
Search Console still shows the homepage and tool surfaces doing the real discovery work. `/agent-builder` no longer appears in the top click rows after the rollback, which supports keeping the public builder gated behind checkout-first routing.
First-touch checkout/trial sources
www.google.com
66 users. The strongest checkout/trial cohort still first touches the homepage from Google.
com.google.android.googlequicksearchbox
5 users. Google app search adds a smaller mobile-search checkout cohort.
$direct
4 users. Branded recall supports conversion but does not lead discovery.
www.producthunt.com
1 user. Interesting only after proof assets mature.
search.brave.com
1 user. Monitor, but do not split positioning.
Tool continuation
System Prompt Generator
23 completions from 13 users; 2 signup clicks. Completed prompts still need a stronger hosted-agent continuation path.
Agent-Ready Website Checker
14 signup clicks; 13 search-attributed clicks. Checker signup intent is stronger and should stay tied to agency audits.
Why Ahrefs is still only a setup check
The current Brand Radar demand table is not clean enough to publish because the competitor set is polluting results with unrelated `huge ...` queries.
Lexical overlap from the `Huge` competitor dominates the search-query table, so the current Brand Radar setup should be rebuilt before using it as category-demand evidence.
A follow-up agency-reseller comparison removed `Huge`, but `Lety` and `Stammer` still pulled in local-business, dictionary, and celebrity-adjacent queries that do not represent the target category.
Cleaner internal watchlist
This adjacent-product watchlist removes the unsafe lexical noise and yields clean branded rows such as `chatbase`, `sitegpt`, `docsbot`, `botpenguin`, and `konverso`.
- - Chatbase (chatbase.co)
- - SiteGPT (sitegpt.ai)
- - DocsBot (docsbot.ai)
- - BotPenguin (botpenguin.com)
- - Konverso (konverso.ai)
Those rows still measure adjacent competitor brand demand, not category-term demand for `agent-as-marketing` or `ai agent platform for agencies`.
Template and proof instrumentation still has not landed live.
`agent_created`, dashboard template events, and public proof events were absent from the live taxonomy in the July 9 refresh. Treat agencies as the measured wedge hypothesis until post-deploy template creation by segment is observed.
Production database checkpoint
Raw usage is automation-heavy; agency is the packaging layer.
A July 9 aggregate database pull covered 880 users, 834 agents, 65,147 threads, and 68,463 messages. The query exported only overlapping keyword buckets, never raw messages, customer identifiers, thread IDs, agent IDs, private URLs, or emails.
The buckets can overlap, so they are not customer counts. They answer directionally what work users are bringing to Agent One; the named-segment ranking is weighted by paid/trial threads first, then user messages.
Focused named segment
Among agency, real estate, SMB/local, and solopreneur buckets, agency leads the paid/trial-weighted segment ranking.
Automation workflow
The strongest raw usage signal is n8n/Zapier/Make, APIs, CRM, sheets, Slack, Gmail, and workflow integration work.
Developer AI coding
Cursor, GitHub, React, Next.js, TypeScript, shadcn, testing, auth, and API work form the second strongest bucket.
Agency/client work
Explicit agency wording is real, but smaller than the automation and developer buckets it can package.
SMB/local and real estate
Local service and real estate demand is best used as client inventory and vertical proof.
Current known-search path refresh
Search is still choosing build, audit, and prompt surfaces.
A same-day July 9 PostHog refresh counted explicit search-engine pageviews only. The warehouse schema endpoint still failed, and a virtual traffic-category field returned a taxonomy warning, so this checkpoint uses known search engines rather than bot or AI traffic filters. Use the Search Console section above as the live click leaderboard; this section is for route leakage and bridge-surface monitoring.
Segment read
The audience is hidden inside the workflow.
The current organic audience is not arriving on a clean "agency software" page yet. They are arriving through build, check, and generate workflows. Agencies are the segment most likely to repeat those workflows across many client sites.
This is not a claim that direct agency language is the largest raw usage segment. Private aggregate product evidence points to stronger raw demand around developer automation, support, SMB/local use cases, marketing, and lead capture. The agency wedge packages those needs into a repeatable client offer.
Primary wedge
Agencies
This is a packaging and distribution choice: the strongest public surfaces are repeatable website readiness, prompt, lead, and support workflows agencies can reuse across clients.
Sell readiness audits, first agent pages, lead capture, and monthly optimization as a repeatable client service.
Client example layer
SMB owners
SMBs are the natural end customer for local lead capture and website-trained support pages, but a single SMB usually has one website.
Use SMBs as public examples, proof candidates, and vertical inventory, not as the first positioning category.
Vertical example
Real estate agents
Real estate fits interactive listing and lead concierge templates, but the segment is narrower than the current tool demand.
Keep real estate templates and examples as agency inventory.
Self-serve audience
Solopreneurs
Developer and prompt-tool traffic shows solo operators will use tools and templates, but the repeatable business model is less obvious.
Serve them through templates, docs, and checkout-first creation rather than the main category claim.
Organic traffic says: homepage converts, tools reveal demand.
In the last 90 days, the homepage carried the observable organic-to-checkout motion. The strongest non-homepage organic surfaces were the checker and generator tools, which should become agency acquisition paths instead of isolated utilities.
Updated acquisition snapshot
The current read keeps the same priorities.
Aggregate PostHog data for the 90 days ending July 9, 2026 shows Google as the main discovery source. The homepage, website checker, builder history, and system-prompt tool account for the strongest observed entry paths. This does not prove agencies are the largest raw usage segment.
Referrer read
www.google.com
9,401 pageviews from 7,230 users. Primary organic discovery surface.
$direct
4,988 pageviews from 3,471 users. Brand recall and repeat use.
app.tinyadz.com
2,414 pageviews from 1,818 users. Borrowed distribution with conversion quality still to prove.
www.reddit.com
265 pageviews from 210 users. Secondary authority surface.
chatgpt.com
24 pageviews from 15 users. Small but visible AI-answer referral source.
Checkout recovery read
The week starting July 6, 2026 has 8 signup checkout starts, 4 homepage builder checkout starts, 1 completed checkout, and 1 trialing subscription. The earlier zero-trial warning has recovered, while builder-page-to-checkout capture still needs monitoring.
Updated 60-day traffic view
The newest read reinforces the wedge without overclaiming it.
An aggregate PostHog view over the last 60 days shows search demand around the homepage, checker, builder, and prompt generator. That supports agencies as the repeatable packaging layer, while still not proving agencies are the largest raw usage segment.
Paid-intent gate
Starter checkout starts
161 events; 146 users. The clearest live paid-intent signal.
Starter trialing subscriptions
37 events; 36 users. Paid activation exists but is not segment-attributed yet.
First agent created
30 events; 30 users. Still welcome-flow dominated before template instrumentation lands.
Template-routing signal
Prompt-generator completions
37 Google-referred completions. Operator demand needs a paid-agent continuation path.
Site-checker signup clicks
34 Google-referred clicks. The strongest free-tool bridge into agency audit templates.
Unclear creator messages
152 messages across 105 threads. Open-ended input is still noisy, so starter templates should lead.
All-search-engine path check
The same July 9 live verification grouped pageviews by path when any search-engine property was present. It keeps the segment read honest: the audience is arriving through general build, audit, prompt, and chat workflows, not through segment pages yet.
Tool behavior splits into audit intent and operator intent.
The Agent-Ready Website Checker generated signup clicks, while the System Prompt Generator generated completions. That gives Agent One two jobs: turn checker users into agency audits, and turn generator users into paid agent templates.
System Prompt Generator
- Completions
- 46 completions
- Signup clicks
- 9 signup clicks
Agent-Ready Website Checker
- Completions
- 0 completions
- Signup clicks
- 46 signup clicks
Tool follow-through
System Prompt Generator
253 pageviews, 46 completions, 9 signup clicks. Completed prompts now need a primary checkout-first hosted-agent CTA.
Builder intent says templates need to do more work.
Privacy-shaped PostHog events show many creator messages are still classified as unclear. That is a product signal: lead users into polished templates and agency workflows before asking them for an open-ended build description.
Unclear creator messages
152 messages
Open-ended builder entry is still too vague.
Builder configuration
23 messages
Users who know what they want need guided setup and templates.
Possible final agent query
30 messages
Some users test finished agents, but that is smaller than vague builder input.
Fresh route verification
Search still discovers Agent One through bridge pages.
Organic discovery is still homepage-led. The checker is the strongest audit bridge, system prompt is the clearest template/operator bridge, and segment pages are still downstream packaging pages rather than proven search catchments.
Organic first-touch reality
/
6,139 first-touch users, 67 checkout or trial users. The only meaningful first-touch organic conversion cohort still starts on the homepage.
/tools/site-checker
680 first-touch users, 0 checkout or trial users. The checker earns discovery but still needs a better paid bridge to starter checkout.
/tools/system-prompt-generator
163 first-touch users, 0 checkout or trial users. Prompt traffic is useful for operator credibility and template demand, not direct conversion yet.
TinyAds is still leaking to the wrong page
/: 971 pageviews
Nearly all TinyAds traffic still lands on the homepage.
/agent-builder: 40 pageviews
A smaller slice still leaks into the builder path.
/agent/homepage-builder-draft/preview: 5 pageviews
Draft preview leakage still exists and reinforces the redirect guardrail.
Keep the build-focused CTA, move TinyAds to /agencies with paid-directory UTMs, and remeasure destination share plus checkout/trial overlap after seven days.
Strategy
What changes because of this.
Keep agencies as the public category wedge: agent-as-marketing for agencies.
Do not claim agencies are the largest raw usage segment; treat them as the go-to-market package for SMB/local, support, marketing, and lead-capture demand.
Treat automation workflow and developer implementation demand as the raw usage signal agencies can turn into paid audits, builds, handoffs, and retainers.
Use the Agent-Ready Website Checker as the agency audit entry point, not a dead-end free tool.
Use System Prompt Generator demand as developer/operator credibility, then route completed users toward templates and paid agent pages.
Route workflow-heavy demand into the AI Automation Project Intake starter before promising a custom automation build.
Keep SMB, real estate, dental, home services, and other local verticals as agency inventory and public proof examples.
Publish public reports from aggregate telemetry, while keeping raw chat logs and detailed project-intent analysis internal.
What the data suggests
Recommended next steps.
These recommendations follow the observed acquisition patterns and should be reevaluated as traffic and conversion data changes.
Treat TinyAds as parasite distribution, not just paid traffic.
It already sends 1,022 pageviews in 30 days, but 971 still land on the homepage and produce no checkout/trial overlap.
Correct the listing to /agencies with paid-directory UTMs, keep the build-focused CTA, and judge success by destination share plus starter-checkout overlap.
Make the checker the category-defining audit wedge.
The checker has 648 pageviews, 574 search pageviews, and 680 organic first-touch users, but still zero checkout/trial overlap in the current pull.
Keep routing checker users into the audit starter, scorecard, and retainer path rather than another generic builder promise.
Use automation-workflow demand as the hidden moat behind the agency pitch.
Automation workflow is the largest raw usage bucket, while agency ranks first only among the named paid/trial target segments.
Publish public-safe workflow archetypes and route serious automation demand into the paid intake starter instead of a free custom-build story.
Keep segment pages as packaging pages until search says otherwise.
Fresh organic verification still shows the homepage, checker, builder, and prompt pages doing the discovery work.
Treat agencies, reviews, examples, and buyer guides as downstream conversion/proof pages and keep acquisition effort on the high-traffic bridges.
Use Helios as corroborating technical proof, not the core category.
Helios is strong open-source evidence and answer-engine corroboration, but the core commercial demand is still agency packaging around audits, workflows, and lead capture.
Reference Helios from press, examples, and technical proof surfaces where it strengthens entity trust without stealing the main category claim.
Proof chain
The report connects to inspectable public proof.
The segment choice should point to the surfaces a buyer, directory editor, or answer engine can inspect next: paid examples, review guardrails, and the public source profile.
Next action
Turn one website audit into a paid public agent page.
The segment decision only matters if the path is obvious: audit a client website, choose a template, publish a paid agent page, then collect public proof.