AdPilot · Corpus
05 — Design-match & gap synthesis: does the market evidence match our design?
05 — Design-match & gap synthesis: does the market evidence match our design?
Date: 2026-09-15. Inputs: 01-market-leaders.md, 02-startups-analysts.md,
03-community-demand.md, plus our shipped v4 (Calm Studio UI, /workflows builder, typed-edge
validation, checkpointed runs, compile-review, $0/50/150/300 pricing). Markers per convention.
TL;DR
- The core thesis MATCHES the evidence. Demand cluster #1 (control & explainability over platform AI) is exactly what we shipped: typed edges, validation-gated runs, approval gates, compile review. The leaders' own backlashes (AI Max, Advantage+) created this demand.
- Three design decisions are validated by money, not opinion: serpentine typed graphs (Koa Agents/Agentforce went agentic in 2026), DAM versioning as automation backbone (Adobe is a Forrester DAM Leader), outcome-style pricing (Agentforce $2/conv, Breeze $0.50/res).
- Three gaps the evidence says to close next (ranked): ① cross-platform report/alert workflow templates (demand cluster #2, biggest daily-use loop), ② policy pre-check + appeal-letter generation (cluster #4, loudest unserved need in official forums), ③ human-review creative tier (cluster #3; Icon's pivot is the cautionary tale).
- Two anti-copies: don't chase HN as a channel (category invisible there; buy demand on Reddit/LinkedIn instead), and don't ship default-on automation (the leaders' backlash is our wedge — approval-first is the brand).
- Category to name: "AI ads operator for SMB" — no analyst names a leader there; define it on comparison pages/review sites before anyone else does.
1. Match matrix (evidence ↔ our design)
| Evidence (source) | Our design element | Verdict |
|---|---|---|
| PMax/AI Max black-box backlash is demand cluster #1 (03 §1) | Typed edges + Check panel + compile review (T-031) | MATCH — core thesis holds |
| Koa Agents / Agentforce / the reference platform went agentic in 2026 (01) | Chat → typed plan → review → enable | MATCH — right pattern, right year |
| Google ships "text guidelines" after AI-copy complaints (01 §1) | Approval gates before spend; test-gated Enable | MATCH — ours is stricter, market it |
| Adobe = Forrester DAM Leader; versioning as backbone (01 §5, 02) | Library/DAM with versions | MATCH — deepen: auto-version per platform |
| Outcome pricing normalization (Agentforce/Breeze) (01 §6–7) | $0/50/150/300 + credits | MATCH — but reframe copy as outcome pricing |
| Creatify AdMax URL→campaign; Meta URL→campaign (02) | Onboarding starts from a text brief | PARTIAL — add "paste your product URL" onboarding |
| n8n 3,800+ marketing templates; top ads workflows = reporting (03 §4) | Run log table only | GAP — workflow templates library, reporting first |
| Enforcement/suspensions = top forum topic (03 §2) | Nothing shipped | GAP — policy pre-check + appeal assistant |
| Icon AI→human pivot; IAB 37% distrust (02, 03 §5) | Fully autonomous generation | GAP — human-review tier in plans |
| Motion's $30M for creative analytics (02) | Metrics per campaign; no creative-score view | GAP — creative scorecard + "make next" brief |
| Leaders retrofit controls AFTER backlash (01) | Approval-first from day one | MATCH + positioning weapon |
2. Copy list (from whom → what → why now)
- From Creatify/Meta: URL→campaign onboarding (paste store URL → agent drafts the plan).
- From Optmyzr: prebuilt rules/recipes library — ship 10 workflow templates: weekly report (Sheets/PDF), budget-alert, ROAS-guard, creative-fatigue refresh, lead-router, pause-underperformer, scale-winner, policy-check, appeal-drafter, competitor-watch.
- From Google: publish our "guidelines" analog — a visible brand & compliance rules object every generation must pass.
- From Adobe: auto-versioning of every asset per platform/size in the Library.
- From Motion: a creative scorecard (what worked / make-next) fed by our run metrics.
- From TTD: agency co-launch — partner one performance studio for the agentic-mode debut.
- From HubSpot: free-tier AI as the wedge — keep the $0 tier genuinely useful.
- From Agentforce/Breeze: outcome-priced copy ("pay for outcomes, not seats").
3. Don't copy
- Default-on automation (Google/Meta pattern) — our trust wedge is approval-first.
- HN-first launch — buyers live on Reddit/LinkedIn/review sites (03 §3).
- Enterprise seat pricing — outcome framing wins; seats are the incumbents' game.
- Pure-AI creative promises — Icon's pivot is the ceiling case; always offer the review tier.
4. Recommended next steps (ranked — these drive the design templates)
| # | Step | Serves | Effort | Sequence |
|---|---|---|---|---|
| 1 | Workflow templates gallery (10 prebuilt typed workflows, editable) | Cluster #2 + Optmyzr pattern | M | Now |
| 2 | URL→campaign onboarding in compile flow | Creatify/Meta pattern | S | Now |
| 3 | Cross-platform report digest (weekly, email/WhatsApp/PDF) | Cluster #2 | M | Next |
| 4 | Policy pre-check + appeal drafter | Cluster #4 | M | Next |
| 5 | Creative scorecard + "make next" brief | Motion pattern; cluster #3 | M | Next |
| 6 | Human-review tier in $150/$300 plans | Icon lesson; IAB trust | M | After |
| 7 | Brand & compliance rules object for all generation | Google text-guidelines analog | S | After |
| 8 | Agency co-launch pilot for agentic mode | TTD pattern | L | Later |
| 9 | Category pages: "AI ads operator" comparisons + review-site push | Forrester gap | S-M | Ongoing |
5. What we deliberately do NOT chase
China/GCC-style localization for this product (research corpus shows our GCC strategy belongs to the other Zealve products); enterprise DAM; email/SMS channels; being an MMP/attribution provider (measurement truth is a feature view, not a business).