Plan-intelligence SKUs — adjacent use cases
Product context: Estym8 is built from scratch as an AI-first construction preconstruction platform. Canonical framing: AI-first positioning.
Status: Discovery map (2026-07-17).
Question answered: Beyond Estym8 (full-project estimating) and municipal plan check, what other jobs can the same plan-reading engine serve?
Related (public): ColdEyes · municipal brief · PlanM8 · A/E brief · Bodi · Bodi brief · Estym8 vs Bodi · folder workflow · AI-first positioning.
Repo: estimate-to-submittal-generator.md · paul-notes.
Shared asset vs job head
| Shared (keep building once) | Swap per SKU |
|---|---|
| Folder / multi-PDF ingest | Buyer + pricing |
| Classify process vs skip | Success metric |
| Sheet / discipline intelligence | Workflow (queue, portal, export) |
| Organic legend / schedule / plan reads | Authority (codes, specs, checklists — never invent) |
| Concerns / conflict patterns | App surface + brand |
Pattern that works: same spine, new job + buyer + success metric.
Pattern that does not: “Estym8 for X that also estimates for X,” or full-cloning the Estym8 app per vertical.
Estym8 stays the full-project estimating product (multi-trade takeoffs from plan sets and bid folders), with two depths on one organic spine: ROM / WAG (fast chase-vs-pass) and full bid-grade takeoff + estimate — see paul-notes §25. Adjacent SKUs do not sell city BOMs or invent compliance.
UX — cohesive workflow + AI assist (suite-wide)
Every suite app (Estym8, Bodi, ColdEyes, PlanM8, …) is a real job workflow with structured screens, queues, and work products. AI assists inside that experience. We do not design chatbot-first products (“ask AI about your plans” as the home). Optional scoped chat (Estee-class / Clara) is a helper on a case, never the product itself. See municipal-plan-check — UX principle.
Suite competitive position — no AI-first suite peer
Claim (honest): There is no AI-first competitor building the suite we are creating — one greenfield plan-reading spine powering Estym8 (GC takeoff/estimate) + Bodi (ICT BOD) + ColdEyes (city review workflow) + PlanM8 (A/E design QA) as cohesive workflow apps.
| True | False / do not say |
|---|---|
| No peer ships (or clearly roadmaps) this full AI-first portfolio on one organic folder-native spine | “No AI competition in construction” |
| Point tools crowd each lane (Archistar/CivCheck/Blitz municipal; takeoff AI; Procore/Autodesk platforms bolting AI on) | “Empty field for municipal plan review” |
| Strategics have breadth but are not AI-first-from-scratch suite products in our sense | “We invented AI plan review” |
How to say it: Point tools and legacy platforms compete in pieces. Nobody else is assembling an AI-first construction intelligence suite — takeoff, city permit review, design QA, ICT BOD — on one engine.
Prioritized use cases
| Priority | Use case | Buyer | Job (not estimating) | Success metric | Fit |
|---|---|---|---|---|---|
| 1 | Municipal plan check | Cities / AHJs | Completeness + locale-code concerns for permitting | Time to deficiency list; false-positive rate | Strong — see municipal-plan-check.md |
| 2 | A/E design review | Design / engineering firms | Internal QA: coordination + completeness before issue / seal | Issues caught pre-issue; review cycle time | Strong — see ae-design-review.md; Paul requirements this week |
| 3 | Applicant pre-check | GCs / A/E / developers | “Will the city bounce this?” before submit | First-pass readiness; fewer resubmittals | Same engine as #1; dual-sided GTM (Permittable already plays here) |
| 4 | Estimate → submittal | Estimators / PMs (Estym8 users) | Spec compliance / submittal packages from the takeoff | Package completeness; draft review quality | Deepest in-repo path — keep inside Estym8, not a new brand (estimate-to-submittal-generator.md) |
| 4b | ROM / WAG screening | GC / electrical estimators (high bid volume) | Chase vs pass before full bid effort (~200 opps/mo) | Time-to-ROM; band hit-rate / ranking quality (not F4c) | Inside Estym8 — one spine, runDepth: rom | full; not a new brand (paul-notes §25) |
| 5 | Owner / CM bid-package diligence | Owners, CMs, developer’s reps | Completeness, conflicts, scope gaps before award | Issues caught pre-award; cycle time to award | High — folder intelligence without BOM |
| 6 | Insurance / surety underwriting assist | Carriers, brokers | Risk flags from plans (life safety, occupancy, systems) | Underwriting time; evidence quality | Medium — concerns + citations; not pricing |
| 7 | Lender / inspector field pack | Lenders, special inspectors | What’s drawn vs what’s claimed; deficiency lists | Inspection prep time; citation-backed flags | Medium |
| 8 | Facility / as-built ops handoff | Owners, facility managers | Device/system inventory from record drawings | Inventory coverage; ops UX adoption | Longer-term — takeoff-like counts, different buyer |
Recommended sequencing
- Keep Estym8 takeoff accuracy advancing (Dogtopia / Castle / broader corpus) — spine credibility.
- Prove municipal (#1) with a design-partner city — completeness + citation comment log + back-check (Paul: highest city value).
- A/E design review / pre-check (#2 + #3) with Paul — same rule packs; firm QA + “run city-style checks before submit.”
- Feasibility (parcel/zoning) when public-data path is clear — early designer/developer wedge.
- Keep submittal (#4) in Estym8 — do not spin a second company for it.
- ROM / WAG (#4b) in Estym8 — same spine, thinner depth; build after full-path organic trust allows bandwidth (not a parallel pipeline).
- Owner / CM diligence (#5) if municipal GTM is slow.
- Revit/BIM model audit — time-boxed spike before promising GA.
- Insurance / lender / facility (#6–8) after concern + citation UX is proven.
- Expansion board (addenda delta, shop vs issued, bid-leveling, packs) — modes on existing products, not new brands —
SUITE_EXPANSION_BOARD.md.
Portfolio (investors — refreshed 2026-07-24; ROM note 2026-07-31)
| Product | Working brand | Buyer | Job |
|---|---|---|---|
| Estym8 | Estym8 | GC / estimators | Full-project takeoff + estimate (+ submittal docs) — raise anchor; ROM/WAG = fast depth on same spine |
| Bodi | Bodi | ICT / LV (+ project-brain framing) | Project brain → BOD (sources/claims) — landing /bodi, host getbodi.ai |
| Municipal plan check | ColdEyes | Cities / AHJs (+ signed-in submitters) | Upload → cold-eyes draft → staff approve/deny/edit → notify; back-check + audit trail; staff keep the stamp (not city BOM) |
| A/E design review / pre-check | PlanM8 | Design / engineering firms | Internal QA + “city-style” pre-submit checks |
| Feasibility (discovery) | SiteM8 (sitem8.ai) | Architects / developers | Parcel/plat/zoning → buildable envelope — landing /sitem8 |
| Model audit (future) | Shared spine | All SKUs | Revit/BIM relational extract + PDF↔model consistency |
Same plan-reading spine; distinct brands/surfaces. Paul notes (2026-07-24): full suite map + shipped-vs-aspirational tags → paul-notes-construction-ai-suite-2026-07-24.md.
SKU sketches (short)
1. Municipal plan check (ColdEyes)
Signed-in upload (applicant or staff) → Submitted → draft review (completeness + citation findings) → notify the appropriate staff or applicant → humans act (approve / deny / return / edit / issue) → notify again. City SoR integration later.
Not city estimation; never automatic outcomes. Detail: municipal-plan-check.md · landing /coldeyes.
2. Applicant surface (same product)
Submitter upload, status lookup, and notifications are part of ColdEyes (dual-sided), not a separate brand. Pre-submit “readiness” polish can deepen later on the same rules.
3. Estimate-to-submittal (Estym8)
Already on the Estym8 roadmap — catalog, packages, compliance matrix, draft spec AI.
Do not rebrand as a separate municipal product.
3b. ROM / WAG screening (Estym8 depth mode)
Job: Rough Order of Magnitude / WAG price (and coarse organic drivers) so estimators can decide which bid opportunities deserve a full detailed estimate.
Architecture: One organic folder spine; runDepth: rom | full. Shared ingest/classify/legend/schedule/scope; ROM early-exits with thin counting + indicative $; full path = current consensus takeoff + estimate. Promote-to-full reuses artifacts.
Not a new brand or a second processing codebase. Detail: paul-notes §25.
Engine: MVP shipped (runDepth + ROM early-exit + labeled romJson). Open: promote-to-full, indicative $ band, ROM eval.
4. Owner / CM diligence
Ingest ITB / bid folder → missing disciplines, revision conflicts, scope ambiguity, coordination risks → diligence memo for award.
No BOM required; concerns + package completeness are the deliverable.
5. Insurance / surety assist
Ingest drawings → structured risk flags (egress, fire, occupancy, major systems) with sheet evidence.
Human underwriter stays in the loop; product must not claim “approved for coverage.”
6. Lender / inspector pack
Lightweight deficiency + evidence export for draw / special inspection prep.
Often a thin export mode on #4 or #1 rather than a full new app.
7. Facility / as-built ops
Record-drawing inventory (devices, panels, equipment tags) for ops/handoff.
Closer to takeoff counts than permitting; lowest near-term priority vs #1–4.
Lines of revenue & ancillary products
Strategy SoT (meters, services, packs): ../SUITE_LINES_OF_REVENUE.md.
Ranked similar / supporting / ancillary + how to monetize and market: ../SUITE_EXPANSION_BOARD.md (public /docs/suite-expansion).
ColdEyes competitive bands: ../ops/coldeyes-pricing-competitive.md.
Honest suite gaps to keep visible: owner/CM diligence ownership (expansion S1), outsourced plan-check ICP (S3), discipline-pack pricing (P4), implementation services rate card (A1), shared citation primitives (S5), insurance partner path (A4) — detail in those two docs.
What not to build
| Anti-pattern | Why |
|---|---|
| City cost estimating / BOMs | Wrong job; confuses Estym8 and municipal GTM |
| Automatic permit approval | Liability; cities will not trust a black box |
| Full Estym8 clone per vertical | Maintenance tax; share libraries, not the app |
| Separate ROM pipeline / “second AI” | Drifts from full path; invites Prime Directive shortcuts; kills promote-to-full reuse |
| Human BOM–derived compliance | Violates Prime Directive culture; invents “truth” |
| Claiming empty-field AI plan review (single lane) | CodeComply, Blitz, Permittable, Archistar, etc. already compete |
| Claiming no AI-first suite peer | OK — see Suite competitive position above |
Brand / engine relationship
| Product | Brand | Engine credit |
|---|---|---|
| Estym8 | Estym8.ai | Primary — estimating |
| Bodi | Bodi (portfolio) | ICT/BOD — complementary; see BODI_ESTYM8_INTEGRATION.md |
| Municipal + applicant | ColdEyes (cold eyes = story/tagline; was ClearToPermit) | “Powered by the same plan-reading engine as Estym8” in partner materials |
| A/E design review | PlanM8 (was CheckSet / ClearToIssue) | Same spine; design-firm GTM — module fallback if one-login preferred |
| Owner / CM diligence | TBD — could share municipal brand or Estym8 add-on | Decide after #1 pilot |
| Submittal | Estym8 | In-product |
Open decisions
Track execution in
../UNIFIED_BACKLOG.mdSuite discovery — not a second queue.
- Confirm sequence: municipal ↔ A/E design review discovery in parallel → applicant → owner diligence
- Whether applicant and city share one domain
- A/E design review: separate brand vs Estym8 “Design Review” module (Paul)
- Whether owner diligence is a third brand or Estym8 “Diligence” SKU
- Insurance/lender: partner vs build after municipal proof
- Investor pack refresh (2026-07-22) — overview / staged brief / pitch include ColdEyes + ClearToIssue; Bodi still secondary
- Paul extensive notes ingested (2026-07-24) — suite modules, back-check, parcel/feasibility, Revit, comment-log schema; see paul-notes…
- Central backlog SoT (2026-07-24) — UNIFIED Now/Next/Later/Suite/Ops
- Paul confirms PlanM8 vs Estym8 module + v1 checklist priority (title-block/index first?)
- Public PlanM8 landing
/planm8+planm8.ai(no longer a ColdEyes alias) - Revit/APS spike scheduled — bar locked 2026-09-01 (Paul): trace + not-judged; CTI24 / S8 / PM8-20
- ROM / WAG depth mode (2026-07-31) — one Estym8 spine,
ROM|FULL; MVP engine + UI — paul-notes §25
Document history
| Date | Note |
|---|---|
| 2026-07-17 | Initial map from municipal discovery conversation — prioritized SKUs sharing Estym8 plan-reading spine |
| 2026-07-20 | A/E design review (#2) for Paul requirements; portfolio note (Estym8 + Bodi + municipal + A/E); investor refresh deferred checklist |
| 2026-08-28 | Public brand PlanM8 landing /planm8; planm8.ai is A/E, not ColdEyes |
| 2026-08-05 | Public cutover: ColdEyes landing /coldeyes (+ CheckSet name); cold eyes = tagline |
| 2026-07-22 | Working brands ColdEyes + ClearToIssue; investor materials refreshed |
| 2026-07-24 | Paul notes synthesis; suite modules (back-check, feasibility, Revit); sequencing updated |
| 2026-07-31 | ROM/WAG (#4b / §3b) — inside Estym8, one spine two depths; engine deferred |
| 2026-09-01 | Paul Revit × code-check signal — spike bar, not Now; PlanM8 waitlist stays PDF |
| 2026-09-01 | Expansion board ranked (similar / supporting / ancillary) + GTM — SUITE_EXPANSION_BOARD.md |