# Message test — https://kylon-main-landing-mockup.lesha27109313.chatgpt.site/ After reading your page, only 2 of 15 personas could name a reason to pick you over a similar option. - **Page tested:** https://kylon-main-landing-mockup.lesha27109313.chatgpt.site/ - **Audience tested against:** C-level executives at series A start ups - **Personas:** 15 simulated - **Report:** https://grader.wynter.com/r/kylon-your-people-and-ai-finally-on-the-same-t-NcHMdps > 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 | 78% | all with reservations | | 2. Relevance | Can they tell what it solves, and who it's for? | 10/15 | 59% | all with reservations | | 3. Value | Do they actually want it? | 9/15 | 56% | all with reservations | | 4. Differentiation | Is there a reason to pick you over the alternatives? | 2/15 | 30% | all with reservations | **Brand alignment** (a side metric, not one of the four layers) — 2/15, 31% 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 **Replace "Real execution" with a line naming what Kylon does that a Zapier-plus-Slack setup cannot.** A buyer already running Zapier and Slack cannot tell what changes here. Say what Kylon does those tools cannot, such as carrying approval, permissions and company context across multi-step work in one place. *effort medium · impact high · tested against Give a reason to choose you* **Under the approval screenshot, add a line describing how approval is enforced.** The "Human approval required" badge is a picture any competitor could show. Add a sentence saying what blocks an agent from sending, who approves, and what is written to the record. *effort low · impact high · tested against Proof next to the claim* **Answer the agent-versus-builder question in the hero instead of leaving it to the FAQ.** The page asks "Is Kylon an AI agent or an agent builder?" in the FAQ, which is where a reader notices nobody has said. State it in the subhead so the category is settled before they scroll. *effort low · impact high · tested against Lead with the use case* ### Value **Add a before-and-after figure under "Why Kylon" tied to one named workflow.** "Agents handle the repeatable work that slows everyone down" has no number behind it, so there is nothing to justify a switch. Give hours saved or errors cut on a specific workflow, and say whose. *effort medium · impact high · tested against Proof next to the claim* **Replace the "What does setup look like" FAQ answer with time-to-first-workflow on the page.** Buyers stall on setup cost and integration before booking anything. State how long the first workflow takes, who does the work, and which systems connect out of the box. *effort medium · impact high · tested against Answer the live objection* **Name the systems Kylon connects to beside "Connect the context."** "The right sources, tools, permissions, and company knowledge" never names a single system, so a buyer cannot tell if their CRM or inbox is covered. List the actual integrations. *effort low · impact high · tested against Concrete over abstract* ### Relevance **Name the role and team in the hero subhead under "The workspace for human + AI teams".** The audience is only guessable from the Guest Care example and the demo button. Say who this is for, such as ops leads running revenue and customer workflows on small teams. *effort low · impact high · tested against Name the audience* **Label the workspace demo card with the team and company type it shows.** "Company Ops" with a Guest Care agent leaves readers guessing whether this is hospitality, support, or something else. Caption it with the industry and team size so buyers can see themselves. *effort low · impact medium · tested against Name the audience* ### Brand alignment (side metric) **Add named customers or logos beside the "A shared workspace" stat band.** Nothing on the page shows another company using Kylon, so the stat band reads as aspiration. Put two or three customer names, or the teams running it, next to those numbers. *effort medium · impact high · tested against Proof next to the claim* **Rewrite "Clear permissions" with the specifics a security reviewer asks for.** "Decide what every agent can see, do, and send" is warm but tells an IT or security reviewer nothing about scopes, audit logs, or data residency. Name the controls and where the record lives. *effort medium · impact high · tested against Concrete over abstract* **Replace "100%" in the stat band with a measured figure or delete it.** "100% permissioned and reviewable" is unsourced and reads as marketing filler next to two blank stats. Use a real number from a customer or cut the band. *effort low · impact medium · tested against Specifics beat superlatives* --- ## 03 · What is working ### The hero line and example card land the problem within seconds Three respondents said the hero line and example card made the problem and solution legible immediately, even where the audience remained implicit. > the hero line "Turn your AI tools into coordinated work" plus the example card showing a "Guest Care agent" drafting an offer that needs human approval made it click within the first screen > > — Chief Technology Officer, Software and SaaS, 51-200 > the hero line "Turn your AI tools into coordinated work. Give every agent a role, the right context, and a place to work with your team" tells me the problem (scattered, ungoverned AI tools) and the fix (a shared, permissioned workspace) within a few seconds. > > — Chief Technology Officer, Software and SaaS, 51-200 > The examples (Guest Care agent, Revenue Ops forecast, Inbox Agent) did the heavy lifting of making it concrete — that's where I actually understood the problem, not from the headline copy. But the intended reader is never spelled out explicitly — there's no "built for ops leaders at 50-person firms" type line — I had to infer it's someone like me > > — Chief Operating Officer, Professional Services, 11-50 --- ## 04 · What the personas said ### Core terms — role, context, tools, workspace, permissions, real execution — are never… Six respondents flagged the page's central vocabulary as abstract, saying they could not map the terms to actual features, data flow, or technical behaviour. > I can't tell if "role" means an org chart position, a permission set, or a prompt template, and that ambiguity is what makes me unsure this is more than a skin on existing permissioning. > > — Chief Technology Officer, Software and SaaS, 51-200 > Phrases like "permissioned and reviewable," "real execution," and "one company memory" are labels without definitions — I don't know what counts as a permission, what "real execution" means technically versus a draft, or what sources feed "company memory." > > — Chief Executive Officer, Technology, 1-10 > I couldn't tell you what's actually under the hood mechanically, or how it's different from just giving Zapier or a custom GPT the same permissions, because the page never specifies that. > > — Chief Operating Officer, Professional Services, 11-50 > generic nouns like "context," "tools," and "company knowledge" are never defined — context could mean a Slack history or a full data warehouse, and I don't know which > > — Chief Technology Officer, Software and SaaS, 51-200 > Words like "workspace," "role," "context," and "coordinated work" are all abstractions with no definition > > — Chief Executive Officer, Technology, 1-10 ### The page does not resolve whether this is an agent or an agent builder Two respondents could not place the product in a category, and one called the unresolved positioning a buyer-confusion red flag. > the FAQ literally asking "Is Kylon an AI agent or an agent builder?" told me the vendor itself hasn't settled the category, so I couldn't either > > — Chief Technology Officer, Software and SaaS, 51-200 > if they're putting that as an FAQ item, it tells me buyers are confused about category fit, and I'd rather not be the one explaining it internally before I've even seen it work > > — Chief Technology Officer, Software and SaaS, 51-200 ### The buyer is never named and has to be inferred from examples and the CTA Five respondents said the target audience was only deducible from the workflow example, demo, or CTA, never stated outright on the page. > The "who" is inferred, not stated — there's no "built for CTOs/ops leaders" line anywhere, I pieced it together from the example (Guest Care agent, Revenue Ops, renewal forecasts) > > — Chief Technology Officer, Software and SaaS, 51-200 > the problem - fragmented AI tools needing coordination plus oversight - became clear only after I'd read the demo mockup, not from the headline > > — Chief Operating Officer, Professional Services, 11-50 > I had to infer "small/mid company with cross-functional ops roles" from the example (Guest Care, Revenue Ops) rather than being told directly. > > — Chief Executive Officer, Technology, 1-10 > the intended reader is never explicitly named—there's no "built for ops leaders" or "for CTOs at mid-market companies" line > > — Chief Technology Officer, Software and SaaS, 51-200 > The examples (Guest Care agent, Revenue Ops forecast, Inbox Agent) did the heavy lifting of making it concrete — that's where I actually understood the problem, not from the headline copy. But the intended reader is never spelled out explicitly — there's no "built for ops leaders at 50-person firms" type line — I had to infer it's someone like me > > — Chief Operating Officer, Professional Services, 11-50 ### There are no customer names, metrics, or case studies anywhere on the page Six respondents said the absence of named customers, before/after numbers, or quantified time and error savings left the workload-reduction claim unvalidated and unable to justify switching. > there's no number on hours saved, error rate, or approval turnaround, and I already run similar workflows through Zapier plus a human check - so the delta they're promising over what I have is unproven > > — Chief Operating Officer, Professional Services, 11-50 > there's no customer name, no before/after numbers, nothing proving this isn't just a demo screenshot. > > — Chief Operating Officer, Professional Services, 11-50 > I'd want them to first show me a concrete before/after (hours saved, errors caught) versus my current stack, because right now it reads as "Zapier plus a sign-off step" with better branding > > — Chief Executive Officer, Technology, 1-10 > I'd need a named customer our size with a number attached — 'Company X, 180 people, runs 500 approvals a week through Kylon' — plus something on security/audit trail (SOC2, data residency for EU since we're an EU shop) > > — Chief Technology Officer, Software and SaaS, 51-200 > what's still fuzzy is the actual time/cost saved versus just using the underlying tools directly > > — Chief Technology Officer, Software and SaaS, 51-200 ### Differentiation claims carry no mechanism or comparison, and the approval flow reads as… Four respondents said the page never explains what is new versus an existing Zapier/Slack setup, offers no competitive comparison, and shows the approval flow only as a screenshot any competitor could claim. > the one thing that could differentiate it, "100% permissioned and reviewable," is just a stat with no mechanism attached > > — Chief Operating Officer, Professional Services, 11-50 > I'd want to know what's actually new here versus Zapier-plus-Slack-plus-a-human-in-the-loop, because the mechanism described — "connect the context," "approve the work" — sounds a lot like what I already cobble together. > > — Chief Executive Officer, Technology, 1-10 > The "Human approval required" gate on the dashboard mock-up is the one concrete thing that'd tip me toward this over a competitor — it directly speaks to the thing that burned me before (a tool sending something dumb without sign-off). But it's a screenshot label, not proof > > — Chief Executive Officer, Technology, 1-10 ### The missing proof layer and generic warmth undercut credibility with security-conscious… Two respondents read the absence of customer and integration evidence as a founder-stage signal, and one said the warm, generic tone was not aimed at a security-conscious operator. > What's missing that a more mature vendor would have is any proof layer—named customers, case studies, an integration list—so it feels like a founder-led team that's nailed the narrative and the UI mockup but hasn't yet built out the trust signals a company my size needs > > — Chief Executive Officer, Technology, 1-10 > The tone is clean and confident but generic — "Finally on the same team," "Good morning, team" — it reads like it was written to feel warm and inclusive to a broad swath of ops leads rather than specifically to a CTO who's been burned before. > > — Chief Technology Officer, Software and SaaS, 51-200 ### Commitment is blocked pending setup cost, CRM integration, and a demo on the buyer's own… Three respondents said they would need proof at scale, a live demo with their own data flows, or answers on setup and integration before any budget decision. > nothing on the page proves agents reliably work across messy real CRM/inbox data at our scale, and I'd need a case study or a technical trial showing error rates and rollback before I'd believe it holds up. > > — Chief Technology Officer, Software and SaaS, 51-200 > I'd want them to show me the permissioning model and the "record of what happened" in a live account with our kind of data flows (inbox + CRM + docs) before I'd put real budget behind it > > — Chief Technology Officer, Software and SaaS, 51-200 > the FAQ literally admits the open questions I'd need answered first ("Is Kylon an AI agent or an agent builder?", "What does setup look like for our team?"), and until I know setup cost and whether it plugs into our actual CRM/inbox without a rebuild, I can't tell if this saves time or just adds another system to manage > > — Chief Operating Officer, Professional Services, 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. - **The page asks for a buying decision while withholding every input a buyer needs to make one.** *(high)* Six respondents found no customers, metrics, or case studies, and three said they could not commit without setup cost, CRM integration answers, and a demo on their own data. The ask outruns the evidence. - **Abstract vocabulary and an unnamed category mean the page never establishes what is actually being sold.** *(high)* Six respondents could not map role, context, tools, workspace, permissions, or real execution to features or technical behaviour, and respondents could not tell whether this is an agent or an agent builder. - **Without a mechanism, the product reads as a reskin of tools buyers already own.** *(high)* Four respondents said nothing explains what is new versus an existing Zapier/Slack setup, and the approval flow appears only as a screenshot any competitor could claim. Abstract core terms give the claim nothing to stand on. - **Negative reactions dominate across every dimension tested, so this is a rewrite rather than a tune-up.** *(high)* Six of the eight themes are negative and span clarity, relevance, value, differentiation, and brand alignment; the only positive covers a hero line three respondents liked. No single fix addresses that spread. - **The strong hero is wasted because nothing downstream converts that comprehension into belief.** *(medium)* Three respondents said the hero and example card land the problem in seconds, but six found no proof and four found no mechanism. Fast comprehension followed by an empty page produces faster disqualification. - **Making buyers infer their own relevance hands the qualification work to the reader and loses the ones who do not bother.** *(medium)* Five respondents said the audience was deducible only from the workflow example, demo, or CTA, never stated. The page relies on the single example card to do targeting it never does explicitly. --- ## 06 · Who answered | # | Role | Industry | Company size | | --- | --- | --- | --- | | 1 | Chief Technology Officer | Software and SaaS | 51-200 | | 2 | Chief Executive Officer | Technology | 1-10 | | 3 | Chief Operating Officer | Professional Services | 11-50 | | 4 | Chief Technology Officer | Software and SaaS | 51-200 | | 5 | Chief Executive Officer | Technology | 1-10 | | 6 | Chief Operating Officer | Professional Services | 11-50 | | 7 | Chief Technology Officer | Software and SaaS | 51-200 | | 8 | Chief Executive Officer | Technology | 1-10 | | 9 | Chief Operating Officer | Professional Services | 11-50 | | 10 | Chief Technology Officer | Software and SaaS | 51-200 | | 11 | Chief Executive Officer | Technology | 1-10 | | 12 | Chief Operating Officer | Professional Services | 11-50 | | 13 | Chief Technology Officer | Software and SaaS | 51-200 | | 14 | Chief Executive Officer | Technology | 1-10 | | 15 | Chief Operating Officer | Professional Services | 11-50 | --- ## 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 2026-10-09, then deleted along with the personas and their answers.