# Message test — https://www.prophecy.ai/ After reading your page, only 9 of 15 personas could name what kind of product this is, unprompted. - **Page tested:** https://www.prophecy.ai/ - **Audience tested against:** Mid-Market to Enterprise B2B SaaS companies, vertical agnostic, leaders of data teams, IT leaders, CDAIOs, but also the analyst and business data analyst themselves doing the work. - **Personas:** 15 simulated - **Report:** https://grader.wynter.com/r/prophecy-ai-data-preparation-analysis-tool-ad1ZNys > 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? | 9/15 | 78% | all with reservations | | 2. Relevance | Can they tell what it solves, and who it's for? | 15/15 | 90% | 8 without hesitation, 7 with reservations | | 3. Value | Do they actually want it? | 14/15 | 74% | all with reservations | | 4. Differentiation | Is there a reason to pick you over the alternatives? | 10/15 | 59% | 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 1 of the personas who named a category got it wrong: - 1× “AI data preparation / data pipeline tool” --- ## 02 · What to change, layer by layer Ordered worst-first. Specific edits, not a restatement of the score. ### Clarity **Replace the H1 'Trusted by the Largest Enterprises' with a line naming what Prophecy is.** The page opens with a trust badge and then a dictionary definition, so a reader scrolls several screens before learning Prophecy is an AI data prep tool. Say in the first line that it preps, analyzes and reports on enterprise data inside ChatGPT, Claude or… *effort low · impact high · tested against Show the product early* **Rewrite 'fix it in plain language' under Visually Verify to say who edits and how.** Readers cannot tell whether fixing a step means typing a correction in chat or handing it back to an engineer. State that an analyst describes the change in chat and the workflow and code update, no engineering ticket. *effort low · impact medium · tested against Plain language* **Delete the duplicated 'Accelerate Data Work \\' block repeating the four feature descriptions.** The four-step explanation appears twice, the second time with stray backslashes, which makes the page read as broken and buries the enterprise section. Keep one clean version. *effort low · impact medium · tested against Headings stand alone* ### Differentiation **Move the 'Execution in Your Infrastructure' claim up into the How Prophecy Works section.** The one thing that separates Prophecy from every chat-based data assistant, that compute and access control run on the customer's own platform rather than in the chat app, sits far down the page. Lead with it where the product is first explained. *effort medium · impact high · tested against Give a reason to choose you* **Under 'Moving Off Alteryx?', state what converts automatically and what does not.** 'Bring your existing workflows with you' leaves a buyer guessing whether migration is a button or a rebuild. Name the share of Alteryx tools that convert and the typical time to move a workflow. *effort medium · impact high · tested against Answer the live objection* **Replace 'fast on managed DuckDB' with a named benchmark or row-scale figure.** 'Fast' is the kind of word every competing tool uses, so the speed claim does no work. Give a concrete number, such as the row count or runtime of a typical join on managed DuckDB. *effort medium · impact medium · tested against Specifics beat superlatives* ### Value **Replace 'Trusted by the Largest Enterprises' with named customers or logos.** An unattributed trust claim at the very top invites doubt rather than settling it. Name two or three enterprises, ideally regulated ones, or cut the line. *effort medium · impact high · tested against Proof next to the claim* ### Brand alignment (side metric) **Add a named customer outcome beside the 'One Governed Path' enterprise section.** The enterprise section asks a data platform lead to trust governance, audit and access-control claims with nothing behind them. Put one customer's result next to the claim, such as security reviews avoided or time to first governed workflow. *effort medium · impact medium · tested against Proof next to the claim* --- ## 03 · What is working ### Respondents could repeat back the core description: AI data prep that clears team backlogs Five respondents restated the product as an AI data prep layer automating joins, reconciliation and reporting through a chat interface, framed as a modernized Alteryx. Inspectable, editable visual workflows were named as a trust-building mechanism. > an AI layer that does data prep, cleaning, joining and reporting work that normally sits in a data team's backlog > > — Data Engineering Manager, Technology, 501-1000 > It's an AI layer that sits on top of your data stack (Snowflake, Databricks, BigQuery, or its own DuckDB) and does the data prep, cleaning, joining, and reporting work your analysts currently do by hand in Alteryx or similar tools — but you drive it from ChatGPT, Claude or Copilot > > — Business Data Analyst, Healthcare, 1001-5000 > Basically a governed, AI-assisted Alteryx replacement that plugs into ChatGPT/Claude/Copilot and runs on Snowflake/Databricks/BigQuery or DuckDB. > > — Chief Data and AI Officer, Retail, 201-500 > the idea that every AI-generated step comes back as an inspectable, editable visual workflow instead of just a chat answer—that's the one concrete mechanism that would actually get me comfortable handing this to a data team with governance obligations > > — Senior Chief Data Officer, Financial Services, 501-1000 > I'd call it an "AI-assisted data prep and workflow automation tool"—basically a modernized, AI-native Alteryx for enterprise data teams, which honestly is exactly how they position it themselves ("Moving Off Alteryx?") > > — Senior Chief Data Officer, Financial Services, 501-1000 ### The problem statement and target audience land immediately Eight respondents said the page names the enterprise data team backlog problem and identifies the data platform lead buyer explicitly in the opening section. This was the most consistently praised element. > the "FOR ENTERPRISE LEADERS" header plus lines like "Start with one pilot team" and "Admins can add it centrally for a pilot team" made the buyer clear — an enterprise data/platform lead rolling this out to teams, which matches me > > — Data Team Lead, Financial Services, 201-500 > the "What is Business Data Work?" section literally names the problem in the first line: prep, reconciliation, analysis and reporting is "still manual, slow to deliver, and stuck in data team backlogs." That's my exact pain. The intended reader is spelled out too: "FOR ENTERPRISE LEADERS" as a section header > > — Business Data Analyst, Healthcare, 1001-5000 > It's obvious fast — the "What is Business Data Work?" section spells it out in the first paragraph: manual prep, reconciliation, reporting stuck in data team backlogs, and the FAQ "What is Prophecy?" confirms it's for "enterprise teams." > > — Chief Data and AI Officer, Retail, 201-500 > "What is Business Data Work?" block names the problem outright: this stuff "is still manual, slow to deliver, and stuck in data team backlogs," which is literally my complaint > > — Senior Chief Data Officer, Financial Services, 501-1000 ### Governance and access control messaging hits a real enterprise pain point Four respondents said the governance integration, audit trail and native platform execution directly address actual pain, specifically reducing data movement and the number of security reviews required. > the real change is governance catching up to what my analysts are already doing in ChatGPT off the books — instead of ad hoc prep happening invisibly, it runs through "your governed infrastructure," with access controls applied and a reusable workflow I can schedule and audit, which is a genuine pain point for us > > — Data Team Lead, Financial Services, 201-500 > our analysts stop burning weeks reconciling the same Excel/CSV/PDF messes and stuck-in-backlog requests get turned into governed, rerunnable dashboards fast, with access controls intact and IT not freaking out about shadow AI use on company data — that's a real problem for us and the "your access controls apply" and "execution in your infrastructure, not the chat" lines speak to the actual governance headache. > > — Senior Business Analyst, Manufacturing, 5000+ ### Execution staying on the customer's own infrastructure is the differentiator respondents… Five respondents pointed to the architectural claim that compute, governance and access control run in the user's platform rather than the chat platform, alongside native app-store integrations and Alteryx workflow portability. > if that's true and low-friction, it's a real switching-cost advantage over a competitor who'd make me rebuild from scratch, and it's concrete rather than a vague promise > > — Data Team Lead, Financial Services, 201-500 > The thing that would actually pull me toward this one over a generic AI-prep tool is the "Execution in Your Infrastructure" line — "Work runs through Prophecy, not in the chat: fast on managed DuckDB, native on your data platform for dev and prod" plus "Your Access Controls Apply." > > — Chief Data and AI Officer, Retail, 201-500 > "Execution in Your Infrastructure: Work runs through Prophecy, not in the chat: fast on managed DuckDB, native on your data platform for dev and prod." That's a concrete architectural claim, not marketing fluff > > — Senior Chief Data Officer, Financial Services, 501-1000 --- ## 04 · What the personas said ### What the product actually is stays ambiguous Two respondents could not tell whether this is a product, a feature, or an API layer, and one questioned what "fix it in plain language" means in practice versus still requiring engineering work. > It wasn't hard to figure out what it does, but what it IS was fuzzy — is it a product, a feature, an API layer? > > — Senior Business Analyst, Manufacturing, 5000+ > I can't tell if "fix in plain language" means retyping a prompt or something closer to actual low-code editing > > — Senior Chief Data Officer, Financial Services, 501-1000 ### Every claim is vendor assertion with no customers, metrics, or case studies behind it Eight respondents flagged missing proof: no named reference customers, no regulated-industry references, no migration numbers, no quantified speed benchmark, and no verification of the metadata-only security boundary. > What's missing is proof it actually beats what I do today: no benchmark on speed-up, no customer number, no "X hours to Y minutes" claim — just "accelerates," which is the word that would make me ask for a demo > > — Data Team Lead, Financial Services, 201-500 > the Alteryx-replacement claim is asserted twice ("Can Prophecy replace Alteryx? Yes.") with zero numbers — no benchmark, no "X% of workflows migrate cleanly," nothing on how a gnarly nested macro survives the import; without a named case study or a migration success rate, that's just marketing copy > > — Business Data Analyst, Healthcare, 1001-5000 > But the page gives me zero numbers — no "X hours saved," no case study, no customer name — just "accelerates delivery" and "governed." Before I take a meeting I'd want one proof point: a reference customer our size, in retail, replacing Alteryx or manual CSV work, with an actual time or cost delta. > > — Chief Data and AI Officer, Retail, 201-500 > this claim is still just asserted on the page; I'd need the Trust & Security page or a live demo to show me the metadata-only boundary actually holding up under a real access-control test > > — Senior Chief Data Officer, Financial Services, 501-1000 > the page leads with "Trusted by the Largest Enterprises" but names nobody, which is exactly the gap that would make me hesitate > > — Data Engineering Manager, Healthcare, 5000+ > it's still marketing copy pattern-matching to my concerns rather than proving them > > — Data Engineering Manager, Technology, 501-1000 ### The missing proof actively undercuts the enterprise positioning Three respondents said the absence of named customers and enterprise credibility markers contradicts the enterprise maturity the page claims, leaving the copy pattern-matching buyer concerns without demonstrating capability. > What would make me doubt the size/maturity guess is the total absence of named customers or case studies anywhere on this page — a company that's actually landed enterprise logos usually can't resist naming one > > — Data Team Lead, Technology, 1001-5000 > it still feels like a company earlier in its enterprise credibility journey than it wants to sound - real enterprise vendors selling to healthcare-scale accounts usually have a logo wall or a named case study by now > > — Data Engineering Manager, Healthcare, 5000+ ### Buyers would not act without seeing it run on their own stack Three respondents said value depends on a live demo of a messy multi-source join on their own infrastructure, or a production audit of the governed workflow and open code claims. One noted ERP/CRM references are generic with no named systems. > Seeing it actually run a messy multi-table join or reconciliation on our own Snowflake data live, with the generated workflow correct enough that one of my engineers would sign off on scheduling it in prod without rewriting it first. > > — Data Engineering Manager, Technology, 501-1000 > I'd want to see the "open code" they promise so my team can audit it before I let it anywhere near production. I'm not taking this to a steering committee off the FAQ copy alone > > — Data Team Lead, Technology, 1001-5000 > I'd want a line naming my actual stack friction — something like "works with Salesforce exports and SAP tables" or a named ERP/CRM, not generic "ERP and CRM Reports" in a diagram box > > — Data Team Lead, Technology, 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 cannot survive a second read: comprehension collapses the moment anyone asks how the claims are substantiated.** *(high)* Eight respondents flagged missing proof — no named customers, no migration numbers, no speed benchmark, no verification of the metadata-only security boundary — while the praised elements are all restatements of the problem and the pitch. - **The enterprise positioning is self-defeating: the page raises the exact credibility bar it then fails to clear.** *(high)* Three respondents said absent named customers and regulated-industry references contradict the enterprise maturity claimed, leaving the copy pattern-matching buyer concerns without demonstrating capability. - **The differentiator is unverifiable architecture, so it reads as a spec sheet claim rather than a reason to switch.** *(high)* Five respondents named in-platform execution of compute and governance as the differentiator, but three said they would need a production audit of the governed workflow and open-code claims before acting. - **The page converts nobody on its own — it hands every decision to a demo it does not offer.** *(high)* Three respondents said value depends on seeing a messy multi-source join run on their own infrastructure, and one noted ERP/CRM references name no actual systems. - **Governance, the strongest value claim, is asserted at exactly the level where enterprise buyers demand evidence.** *(medium)* Four respondents credited the audit trail and reduced security reviews, yet the metadata-only security boundary went unverified for eight respondents and three wanted a production audit of the governed workflow. - **Buyers can describe the product without being able to place it, which stalls procurement before pricing.** *(medium)* Five respondents restated it as a modernized Alteryx, but two could not tell whether it is a product, a feature, or an API layer, and one questioned what fixing things in plain language means versus engineering work. --- ## 06 · Who answered | # | Role | Industry | Company size | | --- | --- | --- | --- | | 1 | Data Team Lead | Financial Services | 201-500 | | 2 | Data Engineering Manager | Technology | 501-1000 | | 3 | Business Data Analyst | Healthcare | 1001-5000 | | 4 | Senior Business Analyst | Manufacturing | 5000+ | | 5 | Chief Data and AI Officer | Retail | 201-500 | | 6 | Senior Chief Data Officer | Financial Services | 501-1000 | | 7 | Data Team Lead | Technology | 1001-5000 | | 8 | Data Engineering Manager | Healthcare | 5000+ | | 9 | Business Data Analyst | Manufacturing | 201-500 | | 10 | Senior Business Analyst | Retail | 501-1000 | | 11 | Chief Data and AI Officer | Financial Services | 1001-5000 | | 12 | Senior Chief Data Officer | Technology | 5000+ | | 13 | Data Team Lead | Healthcare | 201-500 | | 14 | Data Engineering Manager | Manufacturing | 501-1000 | | 15 | Business Data Analyst | Retail | 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 2026-11-03, then deleted along with the personas and their answers.