# Message test — https://www.craftflow.com/ After reading your page, only 10 of 15 personas could name what kind of product this is, unprompted. - **Page tested:** https://www.craftflow.com/ - **Audience tested against:** home improvement leaders of companiges doing $100m revenue - **Personas:** 15 simulated - **Report:** https://grader.wynter.com/r/craft-ai-sales-cx-platform-for-home-improvemen-RcallBE > 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? | 10/15 | 78% | all with reservations | | 2. Relevance | Can they tell what it solves, and who it's for? | 15/15 | 88% | 7 without hesitation, 8 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) — 13/15, 70% 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. --- ## 02 · What to change, layer by layer Ordered worst-first. Specific edits, not a restatement of the score. ### Clarity **Explain what "Score 86" and the coaching prompt are built from.** Call scoring and real-time coaching appear as outputs with no stated inputs, so readers can't tell what the model learns from or how it weights a call. Add a line under Coaching & QA naming the inputs and who sets the scorecard. *effort medium · impact high · tested against Tie the feature to the outcome* **Replace "AI agents for your contact center, sales team, follow up" with one job.** The subhead lists four surfaces at once, so the five modules read as a bundle rather than one product. State the single job — answer, qualify and book every inbound home improvement lead — and let the modules follow. *effort low · impact high · tested against Lead with the use case* **Cut "one context layer" from the journey section subhead.** "Every stage makes the next one smarter" on "one context layer" is an undefined internal term doing the work of an explanation. Say concretely that the in-home conversation is what the follow-up agent uses. *effort low · impact medium · tested against Plain language* ### Differentiation **Say what Craft knows that a horizontal AI vendor does not.** Every module description — voice agents, coaching, CRM sync — could run on a generic sales AI page. Anchor the difference in home improvement specifics: in-home appointments, financing objections, design consultant dispatch. *effort medium · impact high · tested against Concrete over abstract* **Give Handy Bros' "5 → 1" tile a sentence naming the tools replaced.** Consolidation is the one claim on the page a point vendor cannot match, but it sits as a bare number in a logo strip. Name the categories collapsed and what that cost the customer before. *effort low · impact high · tested against Give a reason to choose you* **Move the Brookstone windows-and-doors proof up beside the headline.** The headline says "home improvement teams" while the strongest sub-vertical evidence sits buried in a logo row. Naming windows and doors, roofing and flooring early lets each operator see themselves. *effort low · impact medium · tested against Name the audience* ### Value **Add baseline, timeframe and company size to each percentage tile.** "+31.4% revenue per opportunity" and "+43.9% NSLI increase" float without a starting point or period, so they read as unverifiable. Give each one a before figure, a measurement window and the customer's location count. *effort medium · impact high · tested against Proof next to the claim* ### Brand alignment (side metric) **State the SOC 2 Type II completion date rather than "in progress".** An in-flight audit next to enterprise claims reads as a maturity gap and hands the deal to a compliant competitor. Name the target date and the controls already live. *effort low · impact medium · tested against Answer the live objection* --- ## 03 · What is working ### The headline and subhead name the buyer and the problem without inference Respondents said the headline, subline, customer logos and mockups immediately identified multi-location home improvement operators and the missed-call/missed-lead problem. Several noted this was clear on first read, not generic SMB messaging. > headline says "Home improvement teams win more with Craft" and the subline spells it out: "AI agents for your contact center, sales team, follow up and full customer lifecycle." The three-stage breakdown (Contact Center / Sales / Operations) confirms the buyer is a home improvement operator dealing with missed calls, unscored leads, and rep coaching gaps. > > — Operations Director, Roofing, 1001-5000 > the logo wall (Wilson, LevelTech, Express Flooring, Handy Bros, Brookstone) confirms it's built specifically for home improvement operators > > — VP of Operations, Flooring, 201-500 > the H1 practically spells it out: "Home improvement teams win more with Craft" and "AI agents for your contact center, sales team, follow up and full customer lifecycle." I didn't have to hunt for who this is for; the logos (Wilson, Express Flooring, Handy Bros, Brookstone) > > — Contact Center Manager, Windows and Doors, 501-1000 > the headline "Home improvement teams win more with Craft" plus the very first line "AI agents for your contact center, sales team, follow up and full customer lifecycle" tells me both the audience and the problem in the first two seconds, and then the "Speed to Lead / New lead / Reply 8s" mock-up right under it makes the voicemail/missed-lead pain concrete > > — VP of Operations, Windows and Doors, 201-500 > the logos strip (Wilson, Express Flooring, Brookstone) confirms it's aimed at multi-location home improvement operators like us, not a generic SMB tool. > > — Contact Center Manager, General Contracting, 501-1000 ### Rep productivity is the value respondents believed and could restate Four respondents connected lead scoring and real-time coaching to rep productivity, idle time, and appointment waste, with one saying eliminated voicemail and missed-lead losses would be verifiable within a quarter. > Craft scores every lead and routes the right rep to the right opportunity, learned from closed-loop outcomes > > — VP of Operations, Flooring, 201-500 > Honestly, it's fewer leads lost to voicemail showing up in the pipeline reports within a quarter - not a demo stat, an actual drop in missed-call losses my ops team can point to. > > — Sales Operations Manager, General Contracting, 1001-5000 ### Concrete output beats feature lists — the coaching example and named customers landed Respondents singled out the coaching example with the financing prompt and named customer proof with hard numbers as what drove credibility, contrasting both against generic claims and feature lists. > The named customers - Wilson, LevelTech, Express Flooring, Brookstone - with specific numbers like "+20% close rate" and "+31.4% revenue per opportunity" stuck with me more than the feature list did, that's what would make me actually take a call. > > — Sales Operations Manager, General Contracting, 1001-5000 > the "AI coach" example (price-sensitive buyer, lead with $142/mo financing) shows what good coaching output actually looks like rather than just claiming it > > — Contact Center Manager, Windows and Doors, 501-1000 --- ## 04 · What the personas said ### How the AI models are trained and weighted is never explained Respondents said umbrella terms obscured the product without concrete mockups, that scoring and coaching model training and weighting were unclear, and that the page assumes buyers want slick UI over methodology and algorithm transparency. > it assumes I'll be sold by dashboards, dollar signs, and percentage lifts without asking how those numbers were computed, which tells me their normal buyer is more impressed by slick UI mockups than by a VP who wants methodology. > > — Operations Manager, Roofing, 501-1000 > phrases like "full customer lifecycle" and "Craft Intelligence" are vague umbrella terms that don't tell me what's actually happening, so I had to lean on the concrete mockups (Speed to Lead, AI Voice Agent, Live Scheduling) to actually figure out what the product does. > > — VP of Operations, Windows and Doors, 201-500 > nowhere does it say what data trains it or how the scoring weights get set, so I'm left guessing whether that's a real model or just if-then rules dressed up as AI > > — Sales Operations Manager, Home Improvement, 1001-5000 ### The breadth of the platform reads as bundling rather than a coherent product Respondents described five tools spanning contact center, sales, and follow-up, with one calling it an overloaded pitch wrapping voice agents and scheduling around a core coaching/QA piece and another reading it as optimizing for expansion revenue. > It's an AI layer bolted onto the contact center and sales process for home improvement companies — answering calls, booking appointments, coaching reps mid-sale, and feeding follow-up and CRM data. > > — Operations Director, Roofing, 1001-5000 > the pieces I actually care about (coaching, call scoring, objection tracking) are in there, but it's wrapped in a much bigger pitch about voice agents and scheduling that I don't need > > — Contact Center Manager, Windows and Doors, 501-1000 > the page still tries to sell me the whole stack (voice agents, scheduling, follow-up) when I'd rather they let me buy just the coaching/QA slice, which reads more like a startup optimizing for expansion revenue than one that's fully tuned to a narrow enterprise buyer like me > > — Contact Center Manager, Windows and Doors, 501-1000 ### Home improvement is named but sub-verticals are not One respondent wanted the windows-and-doors vertical named early and a case study from a comparably sized company, indicating the vertical framing is not specific enough for every reader to see themselves. > I'd need my own vertical named outright early - windows and doors, not just "home improvement" broadly - plus a case study from an operator my size (200-500 employees, multi-location) > > — VP of Operations, Windows and Doors, 201-500 ### Every percentage lift is unusable because the page gives no baseline, sample size… Six respondents attacked the case study statistics for missing attribution, baseline, timeframe, and company size. Several said unverified claims would rule the product out or require verified before/after metrics from named accounts before they'd believe… > the page leans on numbers like +31.4% revenue per opportunity and +43.9% NSLI with zero methodology, and a SOC 2 audit that's only "in progress." > > — Operations Director, Roofing, 1001-5000 > Express Flooring +20% close rate) don't say what these companies looked like before, over what time period > > — VP of Operations, Flooring, 201-500 > The percentage lifts (+31.4% revenue per opportunity, +43.9% NSLI) are cited with customer names but no baseline, sample size, or timeframe, so I'd treat those as marketing until I saw the actual case study data. > > — Director of Sales Operations, Home Improvement, 201-500 > The logos and case study percentages (+20% close rate, 5→1 tools) mean nothing without methodology behind them > > — Operations Manager, Roofing, 501-1000 > A verified before/after answer-rate and speed-to-lead number at a named account during their actual peak season - something like 'voicemail rate dropped from X% to Y% at Wilson in Q3' with dates and call volume attached, not a static dashboard screenshot. > > — Director of Sales Operations, Home Improvement, 201-500 ### The in-progress SOC 2 Type II audit undercuts the page's confidence and would lose deals Five respondents flagged the incomplete SOC 2 as a maturity signal, reading the vendor as Series B/C or mid-scale SaaS pretending to enterprise readiness. One said it would lose outright to a compliant competitor. > The "SOC 2 Type II audit in progress" is the tell - a mature vendor would already have the cert > > — VP of Operations, Flooring, 201-500 > I picture a mid-stage B2B SaaS vendor, maybe 5-10 years old, that started with one home improvement niche (windows/roofing/flooring — the demo screens are all window replacement) and is now pushing "enterprise" language to sell up-market to multi-location operators like us. The "SOC 2 Type II audit in progress" and "dedicated AI engineer on every enterprise account" lines tell me they're mid-scale, not a giant incumbent — they're trying to look enterprise-ready but aren't fully there yet. > > — Operations Director, Roofing, 1001-5000 > for a 500-1000 person operation that's not compliant yet, and if a competitor already has that cert finished, that's an easy tiebreaker against Craft > > — Contact Center Manager, Windows and Doors, 501-1000 > the "SOC 2 Type II audit in progress" line gives that away, a mature enterprise vendor would already have it done, not "in progress." > > — Director of Sales Operations, Home Improvement, 201-500 > the SOC 2 "in progress" caveat and the vague "5 → 1" stat make me think they're still earlier-stage than the confident headline suggests. > > — Operations Director, Flooring, 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 proof section is the page's biggest liability, not its asset** *(high)* Four respondents rejected every percentage lift for missing baseline, timeframe, attribution and company size, and several said unverified claims would rule the product out. Only two cited named-customer numbers as credible. - **The page's credibility collapses at the exact moment a buyer starts diligence** *(high)* Five respondents read the in-progress SOC 2 as evidence of a vendor pretending to enterprise readiness, one saying it loses outright to a compliant competitor, while four more demanded verified before/after metrics before believing anything. - **Strong first-read clarity is spent on a product that stops making sense the moment it is examined** *(high)* Six respondents named the buyer and problem on first read, but three described five tools across contact center, sales and follow-up as an overloaded bundle optimized for expansion revenue, and three said model training and weighting were never explained. - **The page sells UI where buyers are pricing methodology** *(medium)* Three respondents said umbrella terms obscured the product, scoring and coaching model weighting was unclear, and the page assumes buyers want slick UI over algorithm transparency — leaving the core AI claim unsupported. - **Vertical targeting stops one level too high to close the deal** *(medium)* The multi-location home improvement framing registered with six respondents, but a respondent still needed the windows-and-doors sub-vertical named early and a case study from a comparably sized company. - **The one value story that landed is narrow enough to be replaced by a point solution** *(medium)* Three respondents restated rep productivity — idle time, appointment waste, eliminated voicemail — while three others read the surrounding voice agents and scheduling as bundling wrapped around a core coaching/QA piece. --- ## 06 · Who answered | # | Role | Industry | Company size | | --- | --- | --- | --- | | 1 | Operations Director | Roofing | 1001-5000 | | 2 | VP of Operations | Flooring | 201-500 | | 3 | Contact Center Manager | Windows and Doors | 501-1000 | | 4 | Sales Operations Manager | General Contracting | 1001-5000 | | 5 | Director of Sales Operations | Home Improvement | 201-500 | | 6 | Operations Manager | Roofing | 501-1000 | | 7 | Operations Director | Flooring | 1001-5000 | | 8 | VP of Operations | Windows and Doors | 201-500 | | 9 | Contact Center Manager | General Contracting | 501-1000 | | 10 | Sales Operations Manager | Home Improvement | 1001-5000 | | 11 | Director of Sales Operations | Roofing | 201-500 | | 12 | Operations Manager | Flooring | 501-1000 | | 13 | Operations Director | Windows and Doors | 1001-5000 | | 14 | VP of Operations | General Contracting | 201-500 | | 15 | Contact Center Manager | Home Improvement | 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 2026-08-31, then deleted along with the personas and their answers.