# Message test — https://iprally.com/ After reading your page, 11 of 15 personas could name what kind of product this is, unprompted. - **Page tested:** https://iprally.com/ - **Audience tested against:** Who buys Corporate IP departments at large, R&D-heavy enterprises. These are the ABM targets. The account briefs we built covered Intel, Sandvik, Shell, Sonos and Volvo Cars, and Unilever is a reference customer. They’re typically engineering, manufacturing, energy, consumer tech and automotive companies with big patent portfolios and in-house IP teams. IP law firms and patent attorney practices doing prior-art, invalidity and drafting work. Perl IP is a reference customer here, and the D.C. Bar and AIPLA activity is aimed at US practitioners. IP service providers and search firms. IP Service World is now framed as a new-logo event, and there’s the RPX partnership. Buying committee Users and champions: patent attorneys, IP counsel, patent engineers and searchers. Economic buyers: head of IP or chief IP counsel in corporates, partners in firms. Influencers: R&D leaders. Gatekeepers: security and procurement. That’s why SOC 2 Type II matters. Jobs to be done Faster, more defensible prior-art, invalidity and FTO searches (Invalidity Agent) Drafting support (Protection) AI built into their own workflows (MCP) - **Personas:** 15 simulated - **Report:** https://grader.wynter.com/r/iprally-the-patent-engine-find-understand-and-9A2CWNM > 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? | 11/15 | 82% | 3 without hesitation, 12 with reservations | | 2. Relevance | Can they tell what it solves, and who it's for? | 15/15 | 91% | 9 without hesitation, 6 with reservations | | 3. Value | Do they actually want it? | 15/15 | 78% | all with reservations | | 4. Differentiation | Is there a reason to pick you over the alternatives? | 13/15 | 70% | all with reservations | **Brand alignment** (a side metric, not one of the four layers) — 15/15, 78% 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 4 of the personas who named a category got it wrong: - 3× “AI patent search and drafting platform” - 1× “AI-powered patent search and drafting platform” --- ## 02 · What to change, layer by layer Ordered worst-first. Specific edits, not a restatement of the score. ### Clarity **Replace "The patent engine for humans and their AI agents" with a plain description of the product.** "Patent engine" leaves readers guessing whether this is a search tool, an agent, or a drafting assistant. Say what it is in the H1: patent search and prior-art analysis you can run in the app, in your AI tool, or over the API. *effort medium · impact high · tested against Lead with the use case* **Add a source and date under every percentage or time-saved claim on the page.** The page states savings and statistics without saying where the numbers come from or who measured them. Put the sample, the date and the method in a short line directly beneath each figure. *effort low · impact high · tested against Proof next to the claim* **Define ai_score in a caption beside the first results panel showing it.** The demo output shows "ai_score": 0.82 with no explanation of what the number measures or its range. Add a one-line caption saying what the score ranks and how to read it. *effort low · impact medium · tested against Plain language* ### Differentiation **Replace the generic time-saved testimonial wording with a named firm, role and measured task.** The testimonial and time-saving format matches every other patent vendor's pitch, so it gives no reason to pick this one. Name the customer, the attorney or analyst role, and the specific search that got faster. *effort medium · impact high · tested against Specifics beat superlatives* ### Relevance **Add an industrial or enterprise example alongside the hydration smartwatch demo.** Every worked example is consumer electronics, so corporate IP teams in industrial sectors cannot see their own filings in it. Run one demo on a mechanical or chemical case, such as the pyrolysis claim already on the page. *effort medium · impact medium · tested against Name the audience* --- ## 03 · What is working ### The demo screens do the explanatory work the copy does not Three respondents credited concrete demo artefacts — claim numbers, confidence scores, and the hydration-sensor FTO example — with making semantic search and claim-level mapping credible and understandable. > They showed a concrete example — a wearable hydration-sensing smartwatch FTO screen across US/China/Europe pulling five active families with claim-level mapping — which is the kind of mechanism detail that actually tells me how it works > > — IP Counsel, Energy, 5000+ > semantic search that actually clusters prior art by technical approach (like the hydration-sensing example, splitting optical vs. bioimpedance vs. sweat-based families) and an FTO screen that comes back with jurisdiction-filtered, active-rights-only results in minutes instead of days > > — Patent Searcher, Consumer Technology, 1001-5000 ### The hero line and audience segmentation make the buyer and job-to-be-done immediately… Five respondents said the problem and target buyer are obvious without inference, citing the hero line, job titles, use-case questions and segmentation. > between the use-case questions ("Who's patenting near us?", "Is this claim ready to file?") and the testimonial job titles, I had it pinned down within the first screen or two > > — Patent Searcher, Consumer Technology, 1001-5000 > the hero line "Find prior art, build competitive intelligence, and protect new inventions, wherever you work" tells you the job to be done right away, and the logo line "Trusted by patent offices, Fortune 500 companies, and AM Law 200 firms" plus the "For Corporates" / "For Law Firms" sections spell out exactly who it's for > > — Senior Patent Attorney, Engineering, 1001-5000 > The hero line — "The patent engine for humans and their AI agents" with the subline "Find prior art, build competitive intelligence, and protect new inventions, wherever you work" — told me the problem (inefficient, keyword-limited prior art and FTO searching) and the three jobs it does > > — IP Counsel, Manufacturing, 5000+ > there are literally separate sections "For Corporates" ("Your IP function, always on, and always under your control") and "For Law Firms" ("Draft with the closest prior art on the page... thorough enough to defend") > > — Chief IP Counsel, Energy, 1001-5000 > The "for" is spelled out explicitly further down too — "For Corporates" and "For Law Firms" sections literally segment the two buyer types, with tailored pitches > > — Patent Engineer, Consumer Technology, 5000+ --- ## 04 · What the personas said ### Percentage and statistic claims carry no methodology or source Four respondents flagged that quantitative claims appear without disclosed methodology or sourcing, and two noted the contrast between precise technical demo detail and vague marketing percentages. > The "75% time saved" and "millions of patent answers a month" numbers have zero methodology behind them > > — Patent Searcher, Consumer Technology, 1001-5000 > it undercuts itself with unsourced big numbers like "75%" and "Millions" sitting right next to the specific stuff, like marketing got hold of the page after the product team wrote the demo screens > > — Patent Engineer, Consumer Technology, 5000+ > that "three hours per search compared to a day and a half" line from the Melexis guy is the kind of thing that would actually get attention if I could see the underlying numbers, not just a pull quote. But right now it's one customer's testimonial with no methodology behind it, same with the generic "75%" time-saved stat repeated twice with no source. > > — Senior Patent Attorney, Engineering, 1001-5000 ### The product category wobbles between search engine, agent, and drafting assistant Three respondents found the offering muddled: mixed metaphors across product types, jargon cluttering the proof, and a positioning read as an AI layer on existing workflows rather than a new category. > I'd call it a patent intelligence and prior-art search tool, not a new category — it's an incremental AI layer on search and drafting workflows I already use > > — Patent Attorney, Automotive, 5000+ > the page mixes metaphors — "the patent engine for humans and their AI agents" plus "Launch agent" plus "Ask anything" plus the MCP/API line all describe the same product from three different angles, so I had to mentally merge "search engine," "agent," and "drafting assistant" into one thing myself instead of being told directly. > > — Senior Patent Attorney, Manufacturing, 1001-5000 ### The examples skew to law firms and consumer products, leaving enterprise and industrial… Two respondents wanted proof the tool handles enterprise scale rather than law firm caseloads, and found the smartwatch demo example consumer-friendly but irrelevant to industrial sectors. > I'd need my own company or a close comparator named on the page — not just Unilever and Melexis as logos, but something like "used by in-house IP teams at consumer electronics companies managing 5,000+ patent families" > > — Patent Engineer, Consumer Technology, 5000+ ### Time-savings claims will not be believed until tested on the buyer's own prior art Four respondents said the stated time savings are testable but require validation against their own patents and claims before they would accept them. > The "75% time saved" and "three hours per search compared to a day and a half" quotes are the kind of number I'd want to pressure-test with our own prior art set before I believe it, not take on faith from a case study blurb. > > — Patent Attorney, Automotive, 5000+ > I'd want them to run it live on two or three of our actual energy-sector patents and compare the output to what our last outside counsel search found, not another case study with Unilever or Melexis numbers I can't audit > > — Chief IP Counsel, Energy, 1001-5000 > that "three hours per search compared to a day and a half" line from the Melexis guy is the kind of thing that would actually get attention if I could see the underlying numbers, not just a pull quote. But right now it's one customer's testimonial with no methodology behind it, same with the generic "75%" time-saved stat repeated twice with no source. > > — Senior Patent Attorney, Engineering, 1001-5000 ### Testimonials and time-saved claims read as interchangeable with competitors Two respondents said the testimonial and time-saving format is identical to every other vendor pitch, and one called the unexplained ai_score metric a rule-out risk. > The testimonials and "75% time saved" are identical noise to what every search vendor on my shortlist says, so they don't differentiate anything > > — Patent Attorney, Automotive, 5000+ > the "ai_score": 0.82 line in the API snippet is a rule-out risk for me — if that score isn't explained (what's it weighted on, claim language, classification, citation graph?) I'll assume it's a confidence number dressed up to look rigorous > > — Patent Attorney, Automotive, 5000+ > they sound like they promise an audit trail back to source documents, but the page never explains what that evidence trail actually contains or how it's validated against examiner reasoning > > — Patent Searcher, Consumer Technology, 1001-5000 ### Company facts are thin and in one place contradictory Two respondents flagged an internally inconsistent founding date and the absence of firmographic detail on headcount, funding, and which offices are actually paying customers. > "Since 2018... Over 8 years of experience" doesn't even add up cleanly > > — Patent Searcher, Consumer Technology, 1001-5000 > What would make me more confident on the "who are they" question is firmographic detail they don't give — headcount, funding, how many of those "six major patent offices" are actually paying customers versus pilots > > — Patent Searcher, Engineering, 1001-5000 --- ## 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 own copy is the weakest asset on it — the screenshots are carrying the argument.** *(high)* Three respondents credited demo artefacts (claim numbers, confidence scores, the FTO example) with making the product understandable, while three others found the written positioning muddled across search engine, agent and drafting assistant. The prose is a… - **Every number on the page is unusable as persuasion because none of it is sourced or tested against the buyer's own portfolio.** *(high)* Four respondents flagged percentage claims with no methodology or sourcing, and four more said time savings would not be believed until run on their own patents and claims. That is over half the sample refusing the quantitative case. - **The proof format is so generic it actively invites comparison shopping rather than ending it.** *(high)* Two respondents said the testimonial and time-saved structure is identical to every other vendor pitch, and the unexplained ai_score metric was named a rule-out risk. Undifferentiated proof plus an opaque metric gives buyers no reason to stop looking. - **The page offers no path to the validation it admits buyers require.** *(high)* Four respondents said they must test time savings against their own prior art before accepting them, yet the page leans on unsourced percentages and interchangeable testimonials instead. The stated objection has no on-page answer. - **The page sells to a narrow slice of the market it claims to address and tells everyone else they are not the customer.** *(medium)* Two respondents wanted evidence of enterprise scale rather than law firm caseloads and dismissed the smartwatch demo as irrelevant to industrial sectors. The same demo screens praised for clarity are the ones scoping the audience down. - **Basic credibility is unforced-error territory: a contradictory founding date undercuts the precision the demo works to establish.** *(medium)* Two respondents found an internally inconsistent founding date and no firmographic detail on headcount, funding, or paying offices, while others contrasted precise technical demo detail against vague marketing percentages. Sloppy facts make the vague numbers… --- ## 06 · Who answered | # | Role | Industry | Company size | | --- | --- | --- | --- | | 1 | Patent Searcher | Consumer Technology | 1001-5000 | | 2 | Patent Attorney | Automotive | 5000+ | | 3 | Senior Patent Attorney | Engineering | 1001-5000 | | 4 | IP Counsel | Manufacturing | 5000+ | | 5 | Chief IP Counsel | Energy | 1001-5000 | | 6 | Patent Engineer | Consumer Technology | 5000+ | | 7 | Patent Searcher | Automotive | 1001-5000 | | 8 | Patent Attorney | Engineering | 5000+ | | 9 | Senior Patent Attorney | Manufacturing | 1001-5000 | | 10 | IP Counsel | Energy | 5000+ | | 11 | Chief IP Counsel | Consumer Technology | 1001-5000 | | 12 | Patent Engineer | Automotive | 5000+ | | 13 | Patent Searcher | Engineering | 1001-5000 | | 14 | Patent Attorney | Manufacturing | 5000+ | | 15 | Senior Patent Attorney | Energy | 1001-5000 | --- ## 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 2026-10-26, then deleted along with the personas and their answers.