# Message test — https://tana.inc/ After reading your page, only 8 of 15 personas could name a reason to pick you over a similar option. - **Page tested:** https://tana.inc/ - **Audience tested against:** Operations and team leads - **Personas:** 15 simulated - **Report:** https://grader.wynter.com/r/tana-do-work-in-the-meeting-gg0hSbY > 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? | 15/15 | 85% | 5 without hesitation, 10 with reservations | | 2. Relevance | Can they tell what it solves, and who it's for? | 15/15 | 94% | 11 without hesitation, 4 with reservations | | 3. Value | Do they actually want it? | 13/15 | 70% | all with reservations | | 4. Differentiation | Is there a reason to pick you over the alternatives? | 8/15 | 52% | all with reservations | **Brand alignment** (a side metric, not one of the four layers) — 8/15, 52% strength (all with reservations). Does the page read like the company you actually are? **Fix first: Differentiation.** Earliest failing layer, walking the sequence in order — not simply the lowest score. --- ## 02 · What to change, layer by layer Ordered worst-first. Specific edits, not a restatement of the score. ### Differentiation **Add a named customer with a before/after number beneath the "Done before you hang up" section.** Nothing on the page says who already runs Tana on real calls or what changed for them. Name one company, its team size, and a concrete result like hours saved or issues filed per week. *effort medium · impact high · tested against Proof next to the claim* **Replace "Tana doesn't take notes" with a line naming what it writes into Linear, Jira and HubSpot.** A buyer comparing Tana to Granola or Fireflies cannot tell from the hero what is different beyond a negation. Say the agent files the issue in Linear and updates the HubSpot deal during the call, not after. *effort low · impact high · tested against Give a reason to choose you* **Add a short proof line under the integrations logo row showing one filed-issue example from a real customer.** The logo wall shows what Tana connects to, not that anyone relies on it. Attach one named team's live workflow to the row so the integrations read as proof rather than ambition. *effort medium · impact medium · tested against Proof next to the claim* ### Value **Add a pilot offer line next to the primary download button.** The timeline comparison is illustrative, and buyers want to see Tana run on their own messy calls before believing it. Offer a trial on live calls with a stated setup time. *effort low · impact high · tested against Answer the live objection* **Source the 30,000 hours stat inline, or cut it.** The number appears with no basis, so readers discount it along with the surrounding claims. Attribute it to customer call data over a stated period, or drop it. *effort low · impact medium · tested against Proof next to the claim* ### Brand alignment (side metric) **Add a line under "Aligned without another meeting" on where call data is stored and who can see it.** An agent with write access to Linear and HubSpot raises an immediate security question for anyone in a regulated or larger org. State that capture runs locally, name the retention and access controls, and say an approval step precedes every write. *effort medium · impact high · tested against Answer the live objection* **Rewrite "For AI-native teams that build products, not PowerPoints" to name ops and cross-functional leads.** That line reads as a founder-to-founder joke and shuts out ops managers, program leads and compliance-minded buyers who do the follow-up work. Name the roles who own decisions and owners after a meeting, not just builders. *effort low · impact high · tested against Name the audience* **Replace the generic role pitch blocks with one concrete meeting outcome per role.** The role sections read as filler next to the detailed Linear bug example. Give each role a specific artifact Tana produces, such as the CRM field an account manager stops updating by hand. *effort medium · impact medium · tested against Concrete over abstract* --- ## 03 · What is working ### The mechanism is understood and repeatable — an agent that acts, not a notetaker Seven respondents restated the product correctly: an agentic assistant that captures calls locally and files to Linear, Jira and HubSpot in real time. Naming the integrations and the act-vs-transcribe distinction did the work. > files the Linear bug, drafts the follow-up email, updates HubSpot, logs decisions with owners, all before the call ends > > — Team Lead, Software Development, 51-200 > every call feeds one shared "knowledge graph," so you can later ask "what did we decide about SSO" and get an answer with provenance > > — Senior Operations Manager, Engineering, 51-200 > It's an AI meeting assistant that sits in your video calls, listens, and then actually executes tasks off the back of them — files bugs into Linear, drafts follow-up emails, updates the CRM, logs decisions with owners — instead of just spitting out a summary afterward like a notetaker would. > > — Head of Operations, Software Development, 201-500 > It's an AI meeting assistant that sits in on your Zoom/Teams/Meet calls, captures what's said, and actually does the follow-up work in real time — filing a Linear ticket, drafting the recap email, updating HubSpot, logging who owns what — instead of just handing you a transcript afterward. > > — Director of Operations, Technology Services, 501-1000 > It's an AI meeting assistant that sits in on Zoom/Teams/Meet calls, captures what's said, and then auto-files the resulting work into tools we already use — opens a Linear/Jira ticket, updates HubSpot, drafts the follow-up email, logs decisions with owners. > > — Operations Lead, Product Management, 1001-5000 > the page was unusually concrete about mechanism (Linear, Gmail, HubSpot named directly) > > — Operations Manager, Engineering, 11-50 ### The hero line and role-segmented blocks land in the first scroll Five respondents said the problem statement and explicit audience segments were legible immediately and reduced inference burden about who the product is for. > For product/sales teams; role list spelled it out. > > — Team Lead, Product Management, 11-50 > the hero line "Tana doesn't take notes. It captures your calls, files the issue, drafts the email and updates the CRM before you hang up, and remembers what everyone promised" tells you the problem (meetings produce work and context that gets lost/retyped) in the first few seconds. > > — Senior Operations Manager, Engineering, 51-200 > there's a literal row of segments (Product and tech, Consultants, VCs, Founders, Customer success, Sales) each with its own pitch, so I didn't have to hunt > > — Head of Operations, Software Development, 201-500 ### Propose-then-accept defuses the unsupervised-agent objection One respondent singled out the approval step as directly answering the fear of an agent with write access to live systems. --- ## 04 · What the personas said ### "Knowledge graph" is used without definition and creates friction Respondents flagged the term as undefined and structurally unexplained, leaving them to interpret what it does. > 'knowledge graph' itself is the term that makes me pause, since it's used as if it's self-evidently a thing (nodes, entities, provenance?) without ever being defined structurally. > > — Senior Operations Manager, Engineering, 51-200 ### The claims are only believable after a demo on messy, unscripted calls Five respondents withheld belief pending a live pilot on their own real calls, saying scripted demos and the mocked timeline prove nothing. The 30,000 hours stat was called out as unsourced. > if it can only do clean retrieval on curated examples, it's not worth more than a trial, not a full rollout decision. > > — Senior Operations Manager, Engineering, 51-200 > Watch it sit through one of our actual messy, interrupt-heavy engineering calls and file a Linear issue I'd hand to a dev with zero edits — if it does that without me cleaning it up, I'm sold > > — Head of Operations, Software Development, 201-500 > "done before you hang up" needs a live demo proof, not copy. > > — Team Lead, Product Management, 11-50 > "30,000 hours in meetings" is a throwaway stat with no source, and "done before you hang up" is a demo claim until I see it on a messy 45-minute call with crosstalk and three different trackers in play. > > — Operations Lead, Product Management, 1001-5000 > One real, verifiable case where a HubSpot deal field updates correctly straight off a call transcript, with no manual cleanup after — that single mechanism working as claimed is what turns this from marketing copy into something I'd actually pilot > > — Team Lead, Software Development, 51-200 ### No named customers, logos or retention data — the most-cited blocker Six respondents said the absence of named reference customers, logos, retention figures and before/after numbers blocks advancement. Several specified what would unblock: a named Jira customer, or a 200+ person company past month three. > What would rule it out, or at least stall it, is that every proof point is generic: no named customer logos, no "Acme Corp cut X hours of admin," just a fictional-sounding "Acme" in the demo screenshots > > — Operations Manager, Technology Services, 1001-5000 > I'd need to see it hold up on a messy 90-minute cross-team call with crosstalk and jargon, and I'd want a named reference customer our size using it with Jira specifically — the page lists Linear/Jira/GitHub integrations but every example screenshot is Linear, which makes me wonder how solid the Jira path actually is. > > — Director of Operations, Technology Services, 501-1000 > A named 200+ person company on record saying they still use it after three months and that decisions logged in the graph actually got pulled up and acted on in a later meeting > > — Senior Operations Manager, Technology Services, 201-500 > there's no named customer anywhere on this page, no logo, no "Company X runs this across 400 engineers on Jira" — every proof point is a generic mocked-up screenshot (and it's always Linear, never Jira, in the examples), so on the thing that matters most to me — who else like me already trusts this — the page gives me nothing, and that absence is the single biggest reason I'd hesitate to pick it over a competitor > > — Director of Operations, Technology Services, 501-1000 > A named reference customer with a quote about hours saved or decisions not re-litigated — something I can call and verify > > — Operations Manager, Engineering, 11-50 > the page gives me zero actual figures, just the mocked timeline graphic — before I take a meeting I'd want one customer reference with hard before/after metrics > > — Head of Operations, Product Management, 501-1000 ### The page reads as founder-to-founder and excludes ops and manager roles Seven respondents said the tone and personas target AI-native startup builders, developers and individual contributors, leaving ops managers, compliance-focused orgs and cross-functional leaders outside the frame. > The tone is confident and a bit cheeky ("Meetings that ship," "your IT department won't freak out") which is fun copy, but it's clearly tuned for a scrappier buyer persona than mine > > — Operations Manager, Technology Services, 1001-5000 > there's no "ops," "internal comms," or "cross-functional" persona bucket, so I'd read this as a tool built by a dev-tool-adjacent team for similar teams > > — Senior Operations Manager, Engineering, 51-200 > my role (ops manager tracking institutional memory across meetings) isn't one of the named buckets > > — Senior Operations Manager, Engineering, 51-200 > "For AI-native teams that build products, not PowerPoints" is a jab at traditional corporates, which tells me they're chasing startups and scale-ups, not a 501-1000 person ops org like mine > > — Head of Operations, Product Management, 501-1000 > The tone is sharp and confident — "Tana doesn't take notes" as an opener is a good hook — but it's written for someone building product or closing deals, not running ops; there's no "Director of Ops" or cross-functional admin persona card, so I'm reading it as an adjacent buyer rather than the intended one > > — Director of Operations, Engineering, 1001-5000 > Tone's for builders, not directors like me. > > — Team Lead, Product Management, 11-50 > the tone leans more "founder pitching founders" than "vendor pitching an ops manager," it's confident and a bit breathless > > — Operations Manager, Technology Services, 51-200 ### The role-based pitch sections read as filler against the technical copy One respondent found the generic role pitches inconsistent with the specificity of the rest of the page. > the role-based mini-pitches ("Run more deals without losing the thread") feel more like generic SaaS marketing copy than something aimed at a skeptical ops manager > > — Operations Manager, Engineering, 11-50 --- ## 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. - **Clarity is the page's only real win, and it buys nothing — people understand the product and still refuse to move.** *(high)* Seven restated the mechanism correctly and five found the hero legible, yet six cited missing logos and retention data as an advancement blocker and five withheld belief pending a pilot. Comprehension is not the bottleneck; proof is. - **The page asks for belief it never earns: every quantitative claim on it is unsourced.** *(high)* Five respondents dismissed the mocked timeline and called out the 30,000 hours stat as unsourced, and six demanded before/after numbers, named customers and retention figures. Nothing on the page survives a request for evidence. - **Audience targeting is self-defeating: the page names roles it then writes past.** *(high)* Five said the role-segmented blocks landed immediately, but seven said the tone speaks founder-to-founder and excludes ops, compliance and cross-functional leaders — and one called the role pitches filler against the technical copy. - **Buyers told you exactly what would unblock them and the page supplies none of it.** *(high)* Respondents specified a named Jira customer, a 200+ person company past month three, and a live pilot on unscripted calls. Six and five respondents respectively named these gaps — the highest-cited themes on the page. - **The strongest objection-handling moment on the page is invisible to almost everyone.** *(medium)* Only one respondent surfaced propose-then-accept as answering the fear of an agent with write access, despite seven correctly describing a product that files into Linear, Jira and HubSpot. The safeguard is buried under the capability. - **Undefined jargon undercuts the page's one working asset.** *(low)* The act-vs-transcribe distinction and named integrations carried comprehension for seven respondents, but 'knowledge graph' was flagged as undefined and structurally unexplained — friction introduced where the page was otherwise clear. --- ## 06 · Who answered | # | Role | Industry | Company size | | --- | --- | --- | --- | | 1 | Operations Manager | Technology Services | 1001-5000 | | 2 | Team Lead | Product Management | 11-50 | | 3 | Senior Operations Manager | Engineering | 51-200 | | 4 | Head of Operations | Software Development | 201-500 | | 5 | Director of Operations | Technology Services | 501-1000 | | 6 | Operations Lead | Product Management | 1001-5000 | | 7 | Operations Manager | Engineering | 11-50 | | 8 | Team Lead | Software Development | 51-200 | | 9 | Senior Operations Manager | Technology Services | 201-500 | | 10 | Head of Operations | Product Management | 501-1000 | | 11 | Director of Operations | Engineering | 1001-5000 | | 12 | Operations Lead | Software Development | 11-50 | | 13 | Operations Manager | Technology Services | 51-200 | | 14 | Team Lead | Product Management | 201-500 | | 15 | Senior Operations Manager | Engineering | 501-1000 | --- ## 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.