# PRD — productanywhere.io **Model:** kimi-k3 · **Phase:** 2 of 3 · **Date:** 20 Aug 2026 **Inputs:** the loop brief + addendum, the seven concepts, `docs/profile/*`, `docs/strategy/*`, `productanywhere/kimi-k3/tokens.css` (Phase 1). **Status:** complete against every Phase 2 exit criterion. Items Joe must supply are collected in §11 — nothing else blocks the build. --- ## 1. Positioning **ProductAnywhere is the independent product practice of Joe da Silva** — executive product leadership, productized consulting, and active venture building, run as one ecosystem since 2014. Joe da Silva is the operator; FormFunction Digital (the Cape Town agency, to 2014) is the lineage and appears exactly once, in the history. Receipts over claims: every number on this site carries the limit of its own attribution, stated before anyone has to ask. One sentence for the fifteen-second scan: *a product leader whose ventures prove the hands, whose consulting proves the range, and whose corporate track record proves the scale.* ### Audiences and their first fifteen seconds | Audience | The job they are doing in 15s | What must be on screen | Success signal | |---|---|---|---| | **Recruiter / board member** (Head of Product → CPO, Chief Product & AI Officer searches) | "Is this person operating at the level I need — and where is the evidence?" | Headline, the colophon's *Now* row, one receipt number **with its limit**, the Hire lane | Clicks through to the receipts ledger or the Hire card; saves the profile | | **Founder / exec buyer** (broken funnel, AI that hasn't landed, GTM debt) | "Can this person fix *my* problem, what does engaging commit me to, and is the proof real?" | The assessment bar, the Consult card's fixed scope, a receipt in their vertical | Starts the assessment, or emails with a problem statement | | **Venture partner** (studios, accelerators, co-founders) | "Does this person actually build — and is there something here worth joining or funding?" | The Builder lane, Pet Basics as a live venture, the concept lab's lineage facets | Opens multiple concepts; starts a Collaborate conversation | The engagement layer speaks to visitors as **Hire · Consult · Collaborate** (their verbs). The narrative layer names Joe's lanes **The Leader · The Expert · The Builder** (his identity). Both are decided in the brief and are not reopened here. --- ## 2. Information architecture & page inventory A page earns its place only if an audience's fifteen-second job needs depth the home sections cannot hold. Five pages earn it at launch: ``` productanywhere.io ├── / Home — the hub (section order in §3) ├── /receipts The full ledger; /receipts/ per-item detail as content allows ├── /assessment The AI readiness assessment — a product surface, not a link (§5) ├── /concepts The concept lab — curated cards + lineage facets, linking to the archive └── /practice About the practice: brand architecture, history, operating principles, fact sheet ``` - **Contact is a home section, not a page.** One low-friction point; a standalone contact page adds a click and nothing else. - **/thinking (articles) is schema-ready but not a launch page.** It earns a page in delivery phase P3, when two real pieces exist. Shipping an empty index is worse than none. - **/privacy earns a page the day the assessment's email gate ships** (POPIA applies; the practice operates from South Africa). One plain-language page. - **Not pages:** CV download (available on request — the site is the evidence), testimonials wall (folded into social proof on home), services pages per lane (the lanes live on home where the reinforcement map can connect them). --- ## 3. Home page — section order, specified Chrome: the masthead (mark + `productanywhere` wordmark; Hire / Consult / Collaborate; availability chip; theme toggle) sits below the bar; the footer closes the page. Numbered sections, in order: **00 — Top assessment bar.** *"How ready is your AI, really? Take the 12-minute assessment →"* Full-width, present on every page. This is a required element and the Expert lane's quiet engine. **01 — Hero.** Eyebrow (`PRODUCTANYWHERE — PRODUCT & GROWTH PRACTICE`); three-line headline, one line emphasised; sub-line naming the practice, the operator, the base; CTA pair — primary `Engage →` (to §07), ghost `See the Receipts` (to §05). Beside it, persistent: **02 — At-a-glance status summary.** The kimi-placement/glm-copy colophon: **Now** (selectively exploring executive product leadership roles), **Also** (fractional and advisory, scoped and time-bound), **Always** (building — the ventures are the R&D lab for the leadership), **Base** (Cape Town — emerging-market depth, global reach), live status chip. This block follows the reader as the hero's companion card; it is not repeated down the page. **03 — The three lanes.** Numbered lane rows (facet, name, duty, arrow): **01 The Leader** — executive product leadership, fractional to full time; **02 The Expert** — productized consulting, fixed scope, named deliverables; **03 The Builder** — active venture building, Pet Basics in market. Beneath the rows, the reinforcement map renders its six directed edges in lane hues (Leader→Expert validates authority; Leader→Builder grounds venture strategy; Expert→Leader proves adaptability; Expert→Builder surfaces ventures; Builder→Leader proves hands-on credibility; Builder→Expert proves implementation capability). **04 — Social proof.** Headed by its own honesty rule: **"Named where public, anonymised where not."** Available today: Absa and Clickatell (public record), Camvas (dated deck, shareable on request). Explicitly stated on the page: several engagements sit under NDA; attributed references are available on request; client logos ship only with written permission. (Asset gap owned by Joe — see §11.) **05 — The receipts ledger.** Section header: **"The receipts. Every number here tells you what it isn't."** Four featured receipts (§4, §6), each `number → claim → limit → proof slug → confidence`, facet at lane hue × confidence step. Link to `/receipts` for the full ledger. **06 — The engagement triad, lightest-first.** **01 Consult** — scoped engagement (30-day growth pilots, governed AI implementation, GTM strategy; 2–6 weeks, fixed scope, named artefacts). **02 Collaborate** — build together (venture studios, accelerators, founders; custom agentic builds, zero-to-one; 3–12 months, partnership). **03 Hire** — executive mandate (fractional, interim, permanent; Head of Product / VP Product / CPO / Chief Product & AI Officer; 1 day/wk → full time). Closing line: *"Most people arrive at 01 and never need the rest. That is a good outcome."* Fractional and interim live under Hire, never Consult. **07 — The concept lab.** Filterable (All / Leader / Expert / Builder); lineage renders in the facet (v1→01, v2→02, v3→03, shipped→04). Cards link to the archive URLs on `portfolio.formfunction.co.za` — the lab is a curated view, the archive stays put (§7). **08 — "Not a fit."** Three reasons *not* to engage, each with a do-this-instead (copy in §6). This section is a conversion device for the right visitor, not a hedge. **09 — Contact.** *"Start with a problem, not a CV."* One email address, one sentence about what a good first message contains, no form. **10 — Footer.** Wordmark, the lineage line (*"ProductAnywhere is the practice of Joe da Silva. FormFunction Digital is the antecedent."*), mono colophon, theme note. --- ## 4. Content model The schema the saturation ladder renders. `lane` ∈ {leader, expert, builder} → hue. `maturity` ∈ {01, 02, 03, 04} → saturation step. Nothing else may set a facet's colour. **receipt** — `slug` · `figure` (display number, e.g. "142M+") · `claim` (one sentence) · `baseline` (what was true before) · `delta` (what changed) · **`limit`** (what Joe did *not* do — required, non-nullable) · `proof_slug` (`proof/absa-142m-faq-deflection`) · `proof_type` (public record | dated artefact on request | published self-audit) · `confidence` (low→01, medium→02, medium-high→03, high→04) · `lane` · `stage` · `vertical` · `year`. Launch inventory: the four receipts from the opus-5 concept (Absa 142M+, Camvas Feb-2022 deck, Clickatell 224% — flagged *downstream evidence, not a claim*, superdev C+ self-audit), plus Pet Basics as a Builder receipt with no invented metrics (group-buy validation → turnkey platform, external investment; figures added only when Joe supplies them). **lane** — `id` · `name` (The Leader…) · `formal_service` (Executive Product Leadership…) · `hue_tokens` (leader-1…4) · `duty` (one line) · `reinforcement_edges` (six, directed). **engagement** — `id` (01/02/03, ascending commitment) · `name` (Consult / Collaborate / Hire) · `lane_affinity` · `duration_band` · `scope_rule` (fixed scope before price) · `deliverables` · `cta` · `closing_line` on the last card. **concept** — `title` · `model` · `lineage` (v1–v3 → steps 01–03; shipped → 04) · `lane_tags[]` · `archive_url` · `status` (lab | shipped). **article** — `slug` · `title` · `date` · `lane` · `stage` · `abstract` · `receipts_referenced[]`. Schema ships now; the index page waits for two real pieces. --- ## 5. The 12-minute AI readiness assessment — product surface **Entry points:** the top bar (every page), the Consult card, the receipts footer line, and the direct `/assessment` URL. It is the Expert lane's proof of method, given away free at the top of the funnel. **Conversational shape.** A conversation, not a form — about fourteen turns, no radio buttons. It adapts to answers and says so when an answer changes the diagnosis (*"I'm scoring you against NIST AI RMF GOVERN-1.2 — the levels are mine, not NIST's. You can check the subcategory yourself."*). The assistant reads the practice's evidence base and leads with attribution fields; it is labelled as an analyst working from a published rubric, never as Joe, and never as a chatbot gimmick. **Free vs gated.** Free, before anything is asked: the score, the diagnosis ("the read"), and the Monday-morning plan. Gated behind an email address: the full roadmap and the PDF. That is the only data the launch site collects. **Scoring — the published rubric (this *is* the publication):** - Every question maps to a **named subcategory of the NIST AI Risk Management Framework 1.0** — four functions, nineteen categories, seventy-two subcategories, all public. - Scoring per subcategory is **0 / 1 / 2**; a dimension score is the **mean × 2.5**, giving 0–10 per dimension. - Maturity **levels 1–5 are the practice's own invention, stated as such on the page.** NIST is explicitly "outcome-focused and non-prescriptive" and defines no maturity levels; anyone selling a "Level 3 against NIST" invented that ladder too. The difference is that this one is published. - On the governance dimension, the relevant **ISO/IEC 42001 Annex A control areas are named** — the standard is not reproduced (it is paywalled, ~CHF 225), precisely because an unverifiable score is not a score. - **Peer benchmarking is refused.** No public dataset supports a percentile and there is no respondent population yet. When the data exists, the practice publishes the *n*. The page says exactly that. --- ## 6. Copy deck — home page (actual sentences) **Bar.** How ready is your AI, really? **Take the 12-minute assessment →** **Hero.** Eyebrow: `PRODUCTANYWHERE — PRODUCT & GROWTH PRACTICE` Headline: **Scale products. / Build ventures. / *Govern the AI.*** Sub: ProductAnywhere is the practice of Joe da Silva — executive product leadership, productized consulting, and active venture building, run as one ecosystem. Based in Cape Town, operating globally, since 2014. CTAs: `Engage →` · `See the Receipts` **Colophon.** Now — Selectively exploring executive product leadership roles. *CPO / VP Product / Head of Product.* Also — Fractional and advisory engagements, scoped and time-bound. *Fixed scope, named deliverables.* Always — Building. The ventures are the R&D lab for the leadership. *Pet Basics, in market.* Base — Cape Town. Emerging-market depth, global reach. *Remote-first.* **Lanes.** 01 **The Leader** — Executive Product Leadership. *I don't just manage products; I scale them — org design, roadmap ownership, and the board conversation that goes with it.* 02 **The Expert** — Productized Consulting. *Senior product and AI expertise without the full-time hire: scoped sprints with named deliverables, never open-ended retainers.* 03 **The Builder** — Active Venture Building. *Pet Basics is live proof that I understand the How as well as the Why — from group-buy validation to a turnkey, externally backed platform.* **Social proof header.** Named where public, anonymised where not. *Several engagements sit under NDA. Attributed references are available on request, and client logos appear here only with written permission — which is exactly how your name would be treated.* **Receipts header.** The receipts. *Every number here tells you what it isn't.* (The four launch receipts render verbatim from the source material, limits included: Absa 142M+ FAQ interactions — "I did not build the model, and I had left before this was measured"; Camvas Feb 2022 — "the deck is dated and I'll send it, judge the dates"; Clickatell 224% — "downstream evidence, not a claim"; superdev C+ — "this one tells you how I audit.") **Engage.** *One subject. Three ways in. Ordered by depth of commitment, lightest first. Every engagement has a fixed scope and a fixed duration before it has a price.* — then the three cards per §3.06, closing: *Most people arrive at 01 and never need the rest. That is a good outcome.* **Not a fit.** 1. *You need AI added to a product that isn't working yet.* AI will make a bad product fail faster and more expensively. **Instead:** fix activation first — a good PM costs less than I do. 2. *Your leadership hasn't agreed that this matters.* Transformation without executive buy-in becomes an expensive internal argument. **Instead:** get one sponsor with budget authority to say it out loud, then call me. 3. *You want someone permanent, in the room, five days a week.* Fractional is the wrong shape and I'll tell you so in the first call. **Instead:** hire — and if the role is right, that's engagement 03. **Contact.** Start with a problem, not a CV. *Three sentences are enough: what's broken, what you've tried, what "fixed" is worth. — hello@productanywhere.io* (address recommendation, §11). --- ## 7. Tech approach, hosting, migration **Build.** Static, self-contained HTML with adjacent `tokens.css` and minimal inline JS — the same no-build-step rule as the concepts repo, so every artefact stays independently openable. Fonts: Google Fonts with preconnect and real fallback stacks. The no-hex-outside-tokens grep is the build gate. **Hosting.** The existing Hetzner box already serves the portfolio via nginx in Docker; the new site is a second static vhost (`productanywhere.io`) on the same box, deployed as a git checkout volume-mounted read-only — the pattern proven this week by the concepts gallery. Analytics: self-hosted, cookieless (Umami or Plausible in the existing compose network), so no consent banner is needed for launch analytics. **Migration.** Nothing moves, so nothing breaks: - `portfolio.formfunction.co.za` **remains the concepts archive of record.** The twelve concept files (plus JobEasy, brand mark, and both models' loop artefacts) keep their exact URLs. No redirects, no 410s — preservation by non-intervention. - The gallery gains one banner line: *"The practice this gallery informed now lives at productanywhere.io."* - `productanywhere.io/concepts` is a curated lab linking *out* to the archive; it does not copy files, so there is exactly one canonical copy of each concept. - Email moves to a `productanywhere.io` address (recommendation: `hello@`, keeping the convention); `hello@formfunction.co.za` forwards during transition. FormFunction Digital appears once, in `/practice`, as lineage. - Canonical URLs, OG cards and the masthead mark all land on `productanywhere.io` at cutover (delivery phase P4). --- ## 8. Analytics & success metrics Directional, reviewed monthly — this is a credibility surface, not a growth dashboard. | Audience | Events | Success looks like | |---|---|---| | Recruiter / board | receipts-ledger entry from home; proof-slug views; Hire card click | ≥40% of name-search visits reach the ledger; Hire clickthrough from hero or engage | | Founder / exec buyer | assessment-bar click; assessment start → completion; Consult card click; contact email | Assessment completion is the primary conversion; Consult clicks second | | Venture partner | concept-lab filter use; multi-concept depth; Collaborate card click | Filtered exploration of ≥2 concepts; Collaborate conversations started | Refused metrics: no peer-comparison claims in any report, no vanity counts on the site itself, no tracking beyond the self-hosted, cookieless default. --- ## 9. Non-goals - No blog engine, no CMS, no framework, no build step. - No client logos or testimonials published without written permission; no invented metrics, ever — a receipt without substantiation ships as prose or doesn't ship. - No peer benchmark percentiles in the assessment until a respondent population exists. - No chatbot performing "Joe". The assessment is a labelled analyst reading a published rubric. - No dark-by-default, no animation budgets, no popups, no newsletter gate. - No agency cosplay: this is a practice of one, and the copy never implies a staff that doesn't exist. - The site does not try to close anyone. It presents evidence and three doors, lightest first. --- ## 10. Phased delivery plan - **P0 — Design system. Done** (this loop's Phase 1). *Definition of done:* tokens.css + style guide live in both themes; contrast tables published; grep gate passes. - **P1 — PRD. Done** (this document). *Definition of done:* every exit criterion in the brief met; §11 owner list handed to Joe. - **P2 — Home page build.** All §3 sections with §6 copy, built strictly on tokens.css. *Definition of done:* the brief's Phase 3 exit criteria — no-hex grep passes; light/dark correct with persisted toggle; 375/768/1280 clean with no body horizontal scroll; semantic, keyboard- navigable, AA against the Phase 1 tables; meta/OG/favicon/title; no render-blocking third-party JS. - **P3 — Receipts ledger + assessment MVP + /practice.** Ledger from the §4 schema; assessment as the specified conversational surface on the evidence base; privacy page when the email gate ships; first two articles unlock /thinking. *Definition of done:* rubric published on-page; free score works end-to-end; email gate is the only data collected. - **P4 — Cutover.** DNS + canonical vhost for productanywhere.io, gallery banner, email move, OG cards. *Definition of done:* all old URLs still serve; new canonical resolves; nothing redirected because nothing moved. ## 11. What Joe must supply (the only blockers) 1. Social-proof assets: written permission for any client logo; one or two attributed references if he wants them published; confirmation that Absa/Clickatell naming stays as the concepts have it. 2. Receipt figures for Pet Basics, if any are publishable — otherwise it ships as prose. 3. Email decision: `hello@productanywhere.io` (recommended) and the FormFunction forward. 4. DNS access for the P4 cutover.