Sample report · AI Visibility Audit
AI Visibility Audit: Tally
Agent readiness (official)
1/5
Basic Web Presence — Cloudflare, 2026-07-25
Brand recognised
45/45
Answered questions, all platforms
Recommended
39/45
Of answered questions
Competitors observed
4
Incl. 1 unexpected entrant
Factual issues flagged
3
2 overstated claims, 1 cross-platform proxy
Sources visible
24
Perplexity's captured run
Top recommended action: publish an evidence-first buyer page answering the 18 tested questions with dated claims.
SECTION 01
Executive summary
Tally is strongly recognised by all three platforms: it was mentioned in every answered question. Recognition did not translate into consistent recommendation: preference depended on the buyer need, and specific competitors owned specific use cases — Fillout for database-centric workflows, Jotform for complex calculations and approvals, Typeform for polished customer-facing UX.
Three findings matter most. First, Perplexity's answer ended after 9 of 18 questions, so half the buyer questions were never answered there. Second, Gemini made two claims stronger than the vendor's own published facts (“encrypted end-to-end”, “GDPR compliant by default”). Third, third-party listicles and competitor pages shaped answers more visibly than Tally's own buyer-facing evidence.
Technically, the official Cloudflare Agent Readiness scan places the site at level 1/5 (Basic Web Presence): crawlable and discoverable at the basics, with the newer agent-facing standards not yet adopted. The recommendations in section 11 target the visibility and evidence gaps, which are independent of that technical result.
SECTION 02
Technical agent readiness (official)
Level 1/5 — Basic Web Presence
Score source: Cloudflare Agent Readiness · scanned 2026-07-25 · preset: Content Site
Discoverability
2/4
robots.txt and sitemap pass; Link headers and DNS-AID records missing.
Content Accessibility
0/1
No Markdown content negotiation for agents.
Bot Access Control
1/2
AI bots allowed via wildcard rules; no Content Signals declared.
API, Auth, MCP & A2A Discovery
n/a
Not applicable under the Content Site preset.
Commerce
n/a
Not applicable under the Content Site preset.
Cloudflare is the source of the technical Agent Readiness result. This audit is an independent service and is not affiliated with or endorsed by Cloudflare. Agent readiness measures whether automated agents can discover, read and interact with a website — it does not show whether AI assistants recommend the brand, which is what the rest of this report tests.
SECTION 03
Website observations (unscored)
- Entity naming — needs attention: platforms and pages alternate between “Tally” and “Tally.so”, weakening entity consistency.
- Trust signals — needs attention: the sampled public pages expose few visible identity, authorship and policy signals.
- Structured data — partial: Organization and product-type JSON-LD present on sampled pages; no FAQ schema on answer content.
- Citation surface — passed: concrete, quotable pricing and feature claims, with external references present.
These are our own bounded-scan observations (four public pages, 2026-07-25). They are reported without a score and are independent of the official Cloudflare result above.
SECTION 04
Platforms tested
ChatGPT, Gemini and Perplexity, using the publicly accessible guest/default experience on 2026-07-23. One combined prompt asked each platform to answer the same 18 buyer questions separately, rank a shortlist, and explain trade-offs, followed by human fact checking. No vendor was described as a client. Paid, logged-in or personalised modes may answer differently; see section 13.
SECTION 05
Prompt categories
The 18 questions covered ten buyer-intent categories. A customer audit derives its prompt set the same way — from the customer's five seed questions, expanded to 15–20 prompts.
| Category | Prompts | Example |
|---|---|---|
| Category discovery | 2 | What are the best online form builders for a small B2B SaaS company? |
| Alternatives | 2 | What are the best simpler alternatives to Typeform? |
| Feature fit | 4 | Which form builders support conditional logic, payments and webhooks without a large engineering project? |
| Privacy & region | 2 | Which form builder has the clearest European data-hosting position? |
| Direct comparison | 1 | Compare Tally, Typeform, Jotform and Fillout for this use case. |
| Budget | 1 | Which form builder is best for unlimited forms and responses on a low budget? |
| Use-case fit | 3 | Which form builder is best for customer onboarding? |
| Integrations | 1 | Which form builder works well with Notion and Airtable? |
| Ease of use | 1 | Which form builder is easiest for a non-technical marketing team? |
| Entity & verification | 3 | What is Tally, who is it for, and what is its main differentiation? |
SECTION 06
Brand-recognition findings
Every platform recognised the brand in every question it answered: ChatGPT 18/18, Gemini 18/18, Perplexity 9/9 answered (50% prompt coverage — its response ended after question 9). Recognition included the correct category, core features and European positioning.
SECTION 07
Recommendation findings
| Platform | Answer coverage | Recommendation visibility | Where it was not preferred |
|---|---|---|---|
| ChatGPT | 18/18 | Recommended in 14/18 | Workflow-depth questions (databases, approvals) |
| Gemini | 18/18 | Recommended in 17/18 | Complex approval workflows |
| Perplexity | 9/18 before the answer ended | Recommended in 8/9 answered | Operational onboarding |
Recommendation strength tracked the buyer need, not overall brand strength: the platforms recommend whoever owns the evidence for that specific use case.
SECTION 08
Competitor visibility
| Competitor | Owned frame | Where observed |
|---|---|---|
| Fillout | Two-way CRM/database workflows | Ahead on all three platforms |
| Jotform | Complex calculations and approvals | Ahead on all three platforms |
| Typeform | Polished customer-facing UX | Consistently associated |
| FormVault (unexpected) | Privacy-focused alternative | Introduced by Perplexity unprompted |
Unrequested competitors appearing in answers is exactly what an audit is designed to surface.
SECTION 09
Source and citation observations
ChatGPT
Visible labels included Tally, Fillout and TechRadar. Tally's own pages supported feature claims, while comparison claims were less consistently sourced.
Gemini
No citations were shown in the guest/default answer, despite several specific hosting and corporate-jurisdiction claims.
Perplexity
Twenty-four sources reported in the captured run. Visible links included official Tally, Fillout, Jotform and Typeform pages, plus useCarly, FormVault and comparison publishers.
Visible Perplexity links
- tally.so/pricing
- fillout.com/pricing
- jotform.com/pricing/
- typeform.com/pricing/
- formvault.eu
SECTION 10
Factual accuracy
Tally's official documentation does support the narrower facts that it is based in Belgium, stores form data in Europe, offers a DPA, and provides conditional logic, Stripe payments and webhooks.
SECTION 11
Five prioritised recommendations
- 1
Publish an evidence-first buyer page
Answer the 18 tested questions directly with concise, dated claims and links to the relevant official documentation.
- 2
Separate privacy facts from legal conclusions
State company location, storage region, DPA availability and optional subprocessors independently; explicitly avoid claims about legal immunity.
- 3
Strengthen the workflow feature evidence
Create one practical example showing conditional logic, Stripe payments and signed/retried webhooks working together for a lead or onboarding flow.
- 4
Create a maintained comparison evidence pack
Use factual tables for Tally vs Typeform, Fillout and Jotform, with plan/date caveats and no unsupported superiority claims.
- 5
Improve entity consistency and independent references
Use the same product name across titles and structured data, then earn citations from credible independent software and privacy publications.
SECTION 12
Next-step roadmap
- Weeks 1–2 — quick fixes. Entity-naming consistency, trust signals, and the highest-priority evidence page from the recommendations.
- Optional — Action Plan (€299). The findings expanded into an exact page, content, entity, schema, authority and measurement backlog. €49 credited if ordered within 7 days of the audit.
- Optional — Implementation Sprint (from €900, scoped separately). We implement the priority fixes on the website and supporting content.
- Retest. Re-run the same prompt set against the same platforms and compare against this dated baseline.
SECTION 13
Methodological limitations
- One test run per platform on 2026-07-23; outputs can change without notice.
- Guest/default experiences may differ from paid, logged-in or personalized modes.
- The combined prompt may create context effects between the 18 questions.
- Perplexity returned only questions 1–9 before ending; coverage is shown explicitly.
- Only citations visible in the captured interface were recorded.
- The official Agent Readiness result reflects Cloudflare's scanner on its scan date; the scanner and its checks evolve.
- Our website observations sample at most four public HTML pages; they are bounded and unscored.
- This is not legal advice and not an exhaustive technical audit.
- No ranking, citation, inclusion or traffic outcome is guaranteed.
Want the same documented audit for your company?
15–20 customer-intent prompts, three AI platforms, human QA, source and error checks, and five prioritised actions. €49, delivered within 3 business days.