# Message test — https://www.open-systems.com/ After reading your page, only 8 of 15 personas could name a reason to pick you over a similar option. - **Page tested:** https://www.open-systems.com/ - **Audience tested against:** Enterprise security executives, CISO, CIO, head of network or infrastructure, purchasing SASE Cybersecurity solutions. - **Personas:** 15 simulated - **Report:** https://grader.wynter.com/r/open-systems-autonomous-sase-ai-run-human-back-mYY4iUU > 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 | 90% | 8 without hesitation, 7 with reservations | | 2. Relevance | Can they tell what it solves, and who it's for? | 11/15 | 63% | all 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) — 11/15, 63% 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 customer stories from financial services and higher education to the proof block.** Every reference is mid-size industrial manufacturing, so regulated and education buyers cannot tell the platform runs at their scale or under their constraints. Name at least one customer in each sector with size and outcome. *effort high · impact high · tested against Proof next to the claim* **Replace "35 years of operational data no competitor can buy" with what the data produces.** The data claim is repeated but never says what it does for a buyer. State the outcome it enables, such as detection or resolution accuracy on named incident types. *effort medium · impact high · tested against Tie the feature to the outcome* **Add a Cato-to-Open Systems migration case with timeline and cutover detail.** The "Move to Open Systems from Cato Networks" link promises a switch but shows no evidence anyone completed one. Publish one migration with duration, sites moved and what broke. *effort medium · impact medium · tested against Give a reason to choose you* ### Relevance **Add a line under the hero naming the buyer role and company situation.** Nothing on the page says who it is for, so a network or security lead has to infer it from industry links buried in the nav. Name the role and the situation, such as a security team running a multi-site SASE estate without an L1 queue. *effort low · impact high · tested against Name the audience* **Replace the garbled hero line "AI just Reset Th%-/\\#\\" with a plain statement of the job.** Readers treat the broken characters as a rendering fault and scroll past the most valuable space on the page. Use the space to say what the product does for whom. *effort low · impact high · tested against Lead with the use case* **Define "Autonomous SASE" in one sentence directly beneath the H1.** The category label carries the whole headline but a first-time reader cannot tell what the AI actually operates or decides. Say plainly that it runs L1 and L2 network security operations with Level-3 engineers on critical calls. *effort low · impact medium · tested against Plain language* ### Value **Add scale context beside the 41-second remediation figure: incident count and period.** A single fast resolution reads as a cherry-picked best case with no way to judge typical performance. Publish the median alongside it and the number of incidents it was drawn from. *effort medium · impact high · tested against Proof next to the claim* ### Clarity **Add one-line job descriptions to the agent list beyond the role labels.** "Lucy, Hermes, Argus" tells a reader nothing about what each agent decides or changes in their environment. Give each a single sentence naming the task it performs and what a human would otherwise do. *effort medium · impact medium · tested against Tie the feature to the outcome* ### Brand alignment (side metric) **Cut "world's best trained" from the H1 and state a measured result instead.** The superlative invites doubt rather than trust and the competitor callout reads defensive. Replace it with a number, such as median time to remediate across incidents last year. *effort low · impact medium · tested against Specifics beat superlatives* **Add one line next to "A Swiss Post company" stating what the ownership means.** Swiss Post ownership and the founding heritage appear as badges with nothing behind them. Say what buyers get from it, such as Swiss data residency and majority state ownership since a given year. *effort low · impact medium · tested against Proof next to the claim* --- ## 03 · What the personas said ### The hero line is broken and says nothing concrete Respondents read the garbled hero copy as a rendering error or copy-paste artefact and had to scroll past it to find substance. Two flagged it directly. > The garbled hero line "AI just Reset Th%-/\#\" actually broke my read — looked like a rendering error or a copy-paste artefact > > — CIO, Financial Services, 1001-5000 > phrases like "AI just reset the race" and the garbled "Th%-/\\#\\" string right after "AI just Reset" — read like broken formatting or a template error, not a claim I could evaluate, so I had to scroll past the whole top section > > — Chief Information Security Officer, Insurance, 5000+ ### Core terms and AI agents are named before they are explained Respondents hit key terms used without definition, multiple named AI agents listed with no stated advantage, and an 'operational data' claim repeated without metrics or benchmarks. Four points cover this. > terms like "Autonomous SASE" and "AIOps" get used before they're defined, so I'm guessing at category boundaries from the competitor bar rather than a clean definition sentence > > — CIO, Financial Services, 1001-5000 > naming ten agents (Lucy, Hermes, Argus, Lex, Atlas...) without a single line explaining why ten named agents beat one good automation engine > > — Head of Infrastructure, Manufacturing, 1001-5000 > "35 years of operational data" is repeated so often as a floating claim, never tied to a metric, that it reads more like a brand tagline than a technical differentiator > > — Head of Infrastructure, Manufacturing, 1001-5000 ### The 41-second remediation example reads as cherry-picked Respondents found the headline incident resolution promising but suspected it was unrepresentative and wanted scale validation. Four points raise this. > the 41-second remediation number (is that cherry-picked or typical?) > > — Head of Infrastructure, Manufacturing, 1001-5000 > the "41s resolved, replicated L3 workflow" example is the kind of number I'd want repeated at scale, not as a demo > > — Chief Information Officer, Financial Services, 5000+ > But that's a cherry-picked example, not a baseline — I have no idea what my current MTTR is on comparable incidents with Zscaler, and they don't give me one to compare against. > > — CIO, Healthcare, 1001-5000 ### MTTR and operational claims lack audited or comparative numbers Respondents wanted published, audited MTTR figures across real incidents and a baseline comparison against Zscaler, plus before-after operational metrics. Three points. > the page gives me one console mockup, not an audited number: I don't know what "replicated" means in practice, how many incident types this covers versus just the demo scenario, or what the failure rate looks like when Lucy's root cause is wrong > > — Chief Information Security Officer, Insurance, 5000+ > But that's a cherry-picked example, not a baseline — I have no idea what my current MTTR is on comparable incidents with Zscaler, and they don't give me one to compare against. > > — CIO, Healthcare, 1001-5000 ### Proof is manufacturing-only, so financial services and higher-ed buyers see themselves… Respondents noted customer logos are mid-size industrial manufacturing with no financial services at scale and no higher-ed references, leaving their own sector unvalidated. Three points. > I'd need reference customers of my size in financial services specifically, not just manufacturing names like Kelvion and CLAAS > > — CIO, Financial Services, 1001-5000 > the customer logos (Kelvion, CLAAS, KEMET) are mid-size manufacturing, not financial services at 5000+ employees > > — Chief Information Officer, Financial Services, 5000+ > none of those proof points are from higher ed - I'd need a customer story that looks like my environment (campus + remote, mixed device population) before I'd trust the "AI resolved it in 41s" demo translates to my world > > — CISO, Higher Education, 1001-5000 ### The specific migration case a Cato incumbent would need is absent One respondent looked for a Cato-to-Open Systems migration case study with technical rigour and found none. > every migration promise ("Move to Open Systems from Cato Networks... Sovereign, not just cloud") is one line with zero detail on how a live Cato SASE deployment actually gets cut over without downtime > > — Head of Infrastructure, Manufacturing, 1001-5000 ### Tone overcompensates when defending AI, and heritage claims go unsubstantiated Respondents saw the voice swing between confident and defensive on AI scepticism, and flagged Swiss Post ownership and founding claims as needing substantiation. Two points. > the constant hammering of "no L1, no L2, no queue" reads like they know CIOs my age are skeptical of AI hype and are overcompensating a bit > > — CIO, Financial Services, 1001-5000 > I'd still want the Swiss Post ownership and the "35 years" claim substantiated with an actual founding date and headcount before I take the pitch at face value > > — CIO, Financial Services, 1001-5000 ### The product category reads as SASE plus an AI-ops layer Respondents could restate the offer as network security combined with AI-powered L1/L2 automation, positioned above Cato, Zscaler and Palo Alto. The operating model tiers were concrete enough to follow. > the three operating model tiers (Self-serve Platform, AIOps, Mission Control) are laid out clearly enough to be concrete, not just marketing fog > > — CIO, Financial Services, 1001-5000 > SASE platform - network and security combined, with AI agents doing L1/L2 ops instead of humans. > > — Head of Network, Healthcare, 5000+ > SASE is the category, the AI/human-backed ops model is their pitch for why it's better than Cato, Zscaler, or Palo Alto. > > — Head of Infrastructure, Manufacturing, 1001-5000 --- ## 04 · The hardest read An adversarial pass over the findings. Every claim below was checked against the panel's own answers; unsupported ones were dropped. - **Every substantive theme on the page is negative — the single non-negative read is a bare category restatement.** *(high)* Seven of eight themes are negative, spanning clarity, value, differentiation and brand alignment; the only neutral theme records that people could restate the offer as SASE plus AI-ops, which is comprehension, not persuasion. - **The page's headline proof point actively damages credibility rather than building it.** *(high)* The 41-second remediation example drew four points of suspicion about cherry-picking, and three more demanded audited MTTR and baseline comparisons against Zscaler. The flagship number invited interrogation instead of belief. - **The customer evidence disqualifies the page for anyone outside mid-size industrial manufacturing.** *(high)* Logos are manufacturing-only, with no financial services at scale and no higher-ed references, and the Cato-to-Open Systems migration case a displacement buyer needs is absent. Three sectors left unvalidated. - **Comprehension stops at the category line — nothing below it survives scrutiny.** *(high)* Three respondents restated the offer correctly as SASE plus L1/L2 automation, yet key terms and named AI agents appear undefined with no stated advantage, and 'operational data' is repeated without metrics. - **The hero fails at the most basic level: it renders as an error and is skipped.** *(high)* Respondents read the garbled hero line as a rendering artefact or copy-paste mistake and scrolled past it to find substance. The highest-value real estate on the page contributes nothing. - **Positioning above Cato, Zscaler and Palo Alto is asserted and then left unproven.** *(high)* Respondents correctly read the competitive frame, but then asked for a baseline comparison against Zscaler and a Cato migration case study and found neither. The claim is made without the evidence it requires. --- ## 05 · Who answered | # | Role | Industry | Company size | | --- | --- | --- | --- | | 1 | CIO | Financial Services | 1001-5000 | | 2 | Head of Network | Healthcare | 5000+ | | 3 | Head of Infrastructure | Manufacturing | 1001-5000 | | 4 | Chief Information Security Officer | Insurance | 5000+ | | 5 | CISO | Higher Education | 1001-5000 | | 6 | Chief Information Officer | Financial Services | 5000+ | | 7 | CIO | Healthcare | 1001-5000 | | 8 | Head of Network | Manufacturing | 5000+ | | 9 | Head of Infrastructure | Insurance | 1001-5000 | | 10 | Chief Information Security Officer | Higher Education | 5000+ | | 11 | CISO | Financial Services | 1001-5000 | | 12 | Chief Information Officer | Healthcare | 5000+ | | 13 | CIO | Manufacturing | 1001-5000 | | 14 | Head of Network | Insurance | 5000+ | | 15 | Head of Infrastructure | Higher Education | 1001-5000 | --- ## 06 · 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 2026-09-16, then deleted along with the personas and their answers.