# Message test — https://userpilot.com/ After reading your page, only 0 of 15 personas could name what kind of product this is, unprompted. - **Page tested:** https://userpilot.com/ - **Audience tested against:** Product and growth leaders at B2B SaaS companies - **Personas:** 15 simulated - **Report:** https://grader.wynter.com/r/product-growth-platform-userpilot-62n90sA > These answers are generated by AI, scored on Wynter's B2B Message > Layers framework using behaviorally-diverse simulated personas. The > methodology is real and the critique is directional. What a simulated > persona cannot have is a live budget, a renewal coming up, or a boss > asking about this quarter. --- ## 01 · The scores Every persona answered all four questions. These are four independent proportions of the same panel, not stages of a funnel. | Layer | Question | Cleared the bar | Strength | Of those who passed | | --- | --- | --- | --- | --- | | 1. Clarity | Do they understand what you do? | 0/15 | 81% | 2 without hesitation, 13 with reservations | | 2. Relevance | Can they tell what it solves, and who it's for? | 15/15 | 85% | 5 without hesitation, 10 with reservations | | 3. Value | Do they actually want it? | 12/15 | 67% | all with reservations | | 4. Differentiation | Is there a reason to pick you over the alternatives? | 7/15 | 50% | all with reservations | **Brand alignment** (a side metric, not one of the four layers) — 11/15, 63% strength (all with reservations). Does the page read like the company you actually are? **Fix first: Clarity.** Earliest failing layer, walking the sequence in order — not simply the lowest score. ### What they thought you sell 10 of the personas who named a category got it wrong: - 5× “Digital adoption / product growth platform” - 4× “Digital adoption / product analytics platform” - 1× “Digital adoption / product onboarding platform” --- ## 02 · What to change, layer by layer Ordered worst-first. Specific edits, not a restatement of the score. ### Clarity **Replace the Lia description with one worked example of a decision it makes.** "Analyzes your product data, generates content, and executes actions" tells a reader nothing about what Lia actually does. Show one case: it spots a drop-off at a step, drafts a tooltip, and you approve it. *effort medium · impact high · tested against Concrete over abstract* **Name the product category in the H1 or the line beneath it.** Nothing on the page says what Userpilot is; readers piece it together from "AI Agent Lia" and scattered feature names. State plainly that it is product analytics plus in-app engagement software. *effort low · impact high · tested against Lead with the use case* **Rewrite "Userpilot AI spots issues, predicts outcomes, and build fixes in-app" with a named outcome.** The claim is unverifiable and also has a grammar error. Say which issue it spots and what the fix was, so the AI section reads like the rest of the page. *effort low · impact medium · tested against Specifics beat superlatives* ### Differentiation **Move the MCP Server block above the AI Intelligence section.** MCP with Claude, ChatGPT and Cursor is the one thing here a competitor cannot copy-paste, and it sits far down the page. Lead the AI story with it instead of with Lia. *effort low · impact high · tested against Give a reason to choose you* **Add company size and industry beside each customer logo in the stories grid.** Names like Relitix, Cuvama and Smoobu mean nothing to a reader who cannot tell whether they are a 20-person startup or a 2,000-person SaaS. Add a one-line descriptor so buyers can see themselves. *effort low · impact medium · tested against Proof next to the claim* **Add one line under "Run the entire product experience from one place" naming the tools replaced.** Every analytics and onboarding vendor claims one place. Say which separate tools a buyer drops when consolidating here. *effort low · impact medium · tested against Give a reason to choose you* ### Value **Add baseline, timeframe and team size to each customer-story stat.** "36% increase in customer lifetime value" has no starting point, period or company context, so it reads as cherry-picked. Write it as from what, to what, over how long. *effort medium · impact high · tested against Proof next to the claim* **Add one short case summary under "Real results from real customers" showing before and after.** The grid of percentages gives no story a buyer can check. One customer, the problem, what they built, the number, and how long it took. *effort medium · impact medium · tested against Conclusion first* ### Brand alignment (side metric) **Cut "AI that analyzes, forecasts, and automates" and replace with what the agent did for a customer.** The AI sections swap the page's concrete voice for slogan language, which reads as hype to a skeptical buyer. Keep the same plain register used in the Session Replay bullets. *effort low · impact medium · tested against Plain language* --- ## 03 · What is working ### The hero problem statement and audience framing land Five respondents said the hero line states the problem clearly and relevantly to their core workflow, and that product, growth, and CS teams are named as the audience. One called the clarity achievable without inference. > "feature discovery into adoption" and "SaaS teams" in the hero/brandkit made it clear. > > — Director of Growth, Technology, 11-50 > the hero line "You ship the feature. Userpilot AI gets it adopted." tells you the problem (features shipped but not adopted) within a few seconds > > — Director of Product, Technology, 501-1000 > the "Teams" section (Product, Design, Customer Success, Marketing cards) plus "1,200 companies... SaaS teams" in the brandkit spells out who it's for. I didn't have to hunt for either — it's explicit, not inferred. > > — VP of Product, Software, 5000+ > I'd need a line naming our actual situation — something like "migrating off spreadsheets/Intercom-bolted-on-tours" or a stat on how long teams like mine take to go from feature ship to measured adoption > > — Director of Product, SaaS, 51-200 ### MCP Server integration with named AI tools is the one differentiator respondents could… Six respondents independently named MCP Server integration — citing Claude, ChatGPT, and Cursor by name — as a concrete, verifiable advantage over Pendo-style competitors. Two said it is the sole differentiator on the page. > The thing that'd actually move the needle for me is the MCP Server — "Bring your Userpilot data into any AI tool you use" with named integrations (Claude, ChatGPT, Cursor, Copilot) is a concrete, checkable differentiator > > — Senior Product Manager, Technology, 11-50 > The MCP Server line — "Bring your Userpilot data into any AI tool you use" — is the one concrete differentiator I'd flag against a shortlist of Pendo/Appcues/WalkMe, because it's a specific, checkable architectural claim > > — Director of Product, SaaS, 51-200 > The MCP Server bit — "Bring your Userpilot data into any AI tool you use," with Claude, ChatGPT, Cursor, Copilot listed — is the one thing that'd actually tip me toward this over a Pendo-style competitor > > — Head of Product, Software, 201-500 > The MCP Server line - "Bring your Userpilot data into any AI tool you use" with named integrations (Claude, ChatGPT, Cursor, Copilot) - is the one concrete differentiator I'd flag against Pendo or WalkMe, because it's a specific, checkable technical claim rather than a vague adjective > > — VP of Product, Technology, 501-1000 > The MCP Server line — "Bring your Userpilot data into any AI tool you use" with logos for Claude, ChatGPT, Cursor, Copilot — is the one concrete differentiator I'd flag against a competitor, because it's a specific, checkable integration claim > > — Senior Growth Manager, Software, 5000+ --- ## 04 · What the personas said ### The product category is never stated; readers assemble it from scattered terms Eight respondents said the page never commits to a category name upfront, forcing them to infer it from use cases and scattered terminology. Several noted the AI-agent framing actively obscures that this is product analytics. > No single phrase names the category outright, so I had to infer it from the mix of terms. > > — Product Manager, Software, 5000+ > Basically Userpilot vs. Pendo/WalkMe territory: no-code in-app guides and usage analytics for SaaS product teams. The customer logos (Doppler, Jiminny, Osano) and their named metrics are the only thing that gives it real shape — the rest of the "AI" framing is marketing gloss over a fairly standard category. > > — Product Manager, SaaS, 51-200 > the page not committing to one word for it — it's called a 'platform,' then broken into five tabs (Product Analytics, User Engagement, Feedback, Session Replay), then an 'AI Agent,' so I had to mentally merge four product categories into one pitch instead of being told up front > > — Senior Product Manager, Software, 201-500 > the "1,200 companies" and all the percentage stats (36%, 75%, 99%) have zero methodology or source attached, so while the problem/audience framing is clear, the proof backing it up is not > > — Head of Product, SaaS, 1001-5000 ### AI claims arrive without definitions, examples, or evidence of what the agent actually… Six respondents said the AI mechanism is asserted but never shown — no worked examples, no concrete decisions it makes, no proof of autonomous execution. Several said the AI layer's contribution to results is undocumented. > "analyzes your data, generates content, executes actions" doesn't tell me what it actually does differently from a rules engine, so I'd want a concrete example of Lia making a decision a human analyst wouldn't have caught. > > — Senior Product Manager, Technology, 11-50 > the AI piece ("Lia", "Agent Analytics") is still vague marketing until I see it actually execute a change without a human writing the logic > > — Director of Product, SaaS, 51-200 > Mainly the AI labels without definitions - "AI Data Analytics," "AI Agent Analytics," "Predictive Analytics" are all named but never explained, so I can't tell if Lia is doing genuine pattern detection or just surfacing dashboards with a chat interface bolted on. > > — VP of Product, Technology, 501-1000 > but the page only asserts that, it doesn't show it. I'd want to see the mechanism (what signals does Lia actually use to decide what nudge to build, and how does it know a flow worked) before I'd trust the AI layer > > — Senior Growth Manager, Software, 5000+ > nothing tying those results to the AI agent specifically versus the plain workflow builder they probably had before Lia existed > > — Growth Manager, SaaS, 1001-5000 ### Who the page is for has to be inferred or scrolled for Four respondents said the target audience is implied through context clues and use cases rather than stated, with personas only identifiable after scrolling past the hero. One noted missing migration-specific context. > I had to infer from context clues like "without a single dev ticket" and the use-case list (onboarding, adoption, churn, support) that this is aimed at product/growth/customer success people, not engineers or marketers. > > — Product Manager, Software, 5000+ > The audience isn't spelled out in that headline, but the "Teams" section a few scrolls down does the job for me — Product, Design, Customer Success, Marketing/Growth each get their own blurb, so I didn't have to guess hard, just scroll. > > — Product Manager, SaaS, 51-200 > I'd need a line naming our actual situation — something like "migrating off spreadsheets/Intercom-bolted-on-tours" or a stat on how long teams like mine take to go from feature ship to measured adoption > > — Director of Product, SaaS, 51-200 ### The statistics are unusable without baseline, timeframe, methodology, or company size Five respondents dismissed the percentage stats as lacking baseline, timeframe, company-size context, and methodology; one called them potentially cherry-picked. Several wanted a before-after timeline or case study isolating the AI layer's impact. > "75% increase in feature usage," "99% reduction in training hours" - with no context on company size or how long that took, so I can't tell if it's comparable to us. > > — Product Manager, Software, 5000+ > those numbers have no baseline or context — increase from what, measured how, over what period — so right now it's marketing copy, not proof. > > — Senior Product Manager, Technology, 11-50 > the page gives me logos and percentage stats (75% increase in feature usage, 5-10x faster adoption) with zero methodology, so I can't tell if that's cherry-picked from one customer or typical. > > — Head of Product, Software, 201-500 > A documented case where a specific shipped feature went from near-zero usage to real adoption within a few weeks, with the before/after numbers tied directly to Lia's nudge, not just a general 'usage went up' stat — that's the difference between a tool I trust and another dashboard I ignore. > > — Head of Product, Software, 201-500 > A documented case where a feature launch went from build to measurable adoption in days instead of weeks, with a before/after number that's attributed specifically to Lia > > — Growth Manager, SaaS, 1001-5000 ### Customer logos do no work because they are unrecognizable and unsized One respondent could not recognize the customer logos or size those companies against their own organization, leaving the social proof inert. > the customer logos (Doppler, Whale, Jiminny etc.) aren't companies I recognise or can size up against us, so I can't tell if they're 50-person startups or actual enterprise peers. > > — Product Manager, Software, 5000+ ### The tone breaks into generic hype whenever AI comes up Four respondents said the AI sections slip into superlatives and marketing language without explaining mechanism, clashing with the page's otherwise precise voice and failing to address buyer skepticism. > it's also overly breezy with the AI-agent framing ("Lia analyzes your product data, generates content, and executes actions") which reads like it's chasing the current AI hype cycle rather than talking to a skeptical buyer who's been burned before > > — Growth Manager, Technology, 11-50 > Tone's generic SaaS marketing-speak, not written for a skeptical technical evaluator like me — too much superlative, not enough mechanism. > > — Director of Growth, Technology, 11-50 > it slips into generic PLG-vendor voice whenever it talks about the AI ("analyzes all your data points, tells you what's going on") and that's where it stops feeling like it was written by someone who's actually used Lia versus someone in marketing extrapolating from a product spec > > — Senior Growth Manager, SaaS, 51-200 > The one place it slips into generic AI-hype register is the Lia/Agent Analytics copy, which reads like every other 2024 AI bolt-on rather than this company's usual precise tone > > — Director of Product, SaaS, 51-200 --- ## 05 · The hardest read An adversarial pass over the findings. Every claim below was checked against the panel's own answers; unsupported ones were dropped. - **The page's central AI positioning is the weakest part of it, and it drags the credible parts down with it.** *(high)* Five respondents found the AI mechanism asserted but never shown, four said the AI sections slip into superlatives clashing with the page's otherwise precise voice, and six said the AI-agent framing obscures what the product actually is. - **Readers have to do the categorization and qualification work the page refuses to do.** *(high)* Six respondents said no category name is ever committed to, and four said the audience must be inferred from use cases or found only after scrolling past the hero. Every reader assembles a different product in their head. - **Every proof element on the page is inert.** *(high)* Four respondents dismissed the percentage stats for lacking baseline, timeframe, company size, and methodology, with one calling them cherry-picked; another could not recognize or size the customer logos. Nothing on the page survives scrutiny. - **The single differentiator respondents found is an integration list, not a product advantage.** *(high)* Six respondents named MCP Server with Claude, ChatGPT, and Cursor as the concrete advantage over Pendo-style competitors, and two called it the only one. A page whose sole differentiator is third-party tool names has no defensible position of its own. - **Buyer skepticism about AI is left entirely unanswered, which converts the page's biggest claim into its biggest liability.** *(high)* Four respondents said the AI sections use marketing language without mechanism and fail to address skepticism, while five wanted worked examples, concrete decisions, or a case study isolating the AI layer's contribution. - **The hero is the only section carrying its weight, and the page squanders the attention it earns.** *(medium)* Five respondents said the hero states the problem clearly and names product, growth, and CS teams. Everything after it — category, AI mechanism, statistics, logos — was flagged as unsupported or inferred. --- ## 06 · Who answered | # | Role | Industry | Company size | | --- | --- | --- | --- | | 1 | Product Manager | Software | 5000+ | | 2 | Senior Product Manager | Technology | 11-50 | | 3 | Director of Product | SaaS | 51-200 | | 4 | Head of Product | Software | 201-500 | | 5 | VP of Product | Technology | 501-1000 | | 6 | Growth Manager | SaaS | 1001-5000 | | 7 | Senior Growth Manager | Software | 5000+ | | 8 | Director of Growth | Technology | 11-50 | | 9 | Product Manager | SaaS | 51-200 | | 10 | Senior Product Manager | Software | 201-500 | | 11 | Director of Product | Technology | 501-1000 | | 12 | Head of Product | SaaS | 1001-5000 | | 13 | VP of Product | Software | 5000+ | | 14 | Growth Manager | Technology | 11-50 | | 15 | Senior Growth Manager | SaaS | 51-200 | --- ## 07 · Before you act on this The methodology is real, and the critique is directional. What a simulated persona cannot have is a live budget, a renewal coming up, or a boss asking about this quarter. **Validate anything you're betting on with real ICPs who are actually in-market.** Being wrong is more expensive than you think. Finding out is cheaper than you'd guess. Wynter runs message testing with verified B2B professionals — trusted by HubSpot, RingCentral, Shopify, Cognism, Paddle, Veeam, Rippling and Miro. This report is kept for 60 days from 2026-10-05, then deleted along with the personas and their answers.