P
Perch · Roadmap Reconciliation & Plan

Roadmap Reconciliation & Plan

My read of the Perch Centaur Roadmap cross-checked against the codebase — the working artifact for our side-by-side discussion. Preliminary: codebase presence ≠ production-ready. Confirm before I build.

Headline finding
The codebase is materially ahead of — and broader than — the roadmap.

A lot of what's scheduled for later phases already exists at least in part, and several shipped capabilities (MEDDPICC engine, field-ops, Company Brain) aren't in the roadmap at all.

01 What the roadmap is From the document

A bootstrapped, founder-capital-minimized sprint to $100M ARR in 52 weeks, four phases. Phase names, week boundaries, month mapping, and goals are all verbatim from your roadmap — not my labels.

PhaseWeeksGoalCore buildMotion
1 · Elite Wedge1–12$1M ARRAgentic Pipeline + Glyph Traceability; full Division/Section coveragePaid beta subs ($20–50k ACV) via $300M partner
2 · Enterprise Pivot13–26$15M ARRGC Command Center + Deep Diff engineLock $2B/$3B GCs on MSAs
3 · Viral Flywheel27–39$50M ARRSub-Contractor Portal"The Mandate" — GCs force subs onto Perch
4 · Industry Standard40–52$100M ARRLine-item costing (RSMeans) + RFI automation → Pre-Con OS; SOC2 IIAI-SDR on ENR Top 400

02 Preliminary reconciliation Needs your confirm

Roadmap capability ↔ codebase evidence ↔ my read. Status tags: Done In progress Not yet Beyond plan

Roadmap capabilityCodebase evidenceMy read
Doc upload → SOW generationdoc pipeline, "scope of work generation"Done
Glyph-level traceability / reconstructiondoc_server reconstruction stageDone
Entity extraction / spaCy labelingextraction pipeline + testsDone
Addenda diffing (basic)addenda refsBasic done · "Deep Diff" = the upgrade
Hallucination tracking / V3 harness / tunerharness dashboard, V3 bench, eval suitesDone / partial
Division/Section (trade) coverageprecon-fs; tuning masonry/metals/cabinetry/electricalIn progress
Analytics (KPIs, usage, link opens)analytics pagesPartial
Lead gen (ITB shadow cc/bcc, daily cron)lead-gen-preferences, lead-gen pages, ITB refsPartial
GC Command Center / Deep Diff (Ph2)business-intelligence, addendaPartial / early
Sub Portal + bid submission (Ph3)bid-verification, multi-GC scenariosPartial / unclear
RSMeans line-item costing / RFI (Ph4)marketing refs only; some RFI refsMostly not yet
SOC2 Type II (Ph4)claimed in pricing FAQClaim vs. actual
AI-SDR (Ph4)BD-rep agent existsPartial
MEDDPICC sales pipelinefull sales-orchestrator, qualification scoring, MAP stepsBeyond plan · not in roadmap
Field-ops, network, tribal knowledge, legal playbooks, Company Brain, agent guardrailsshipped modules/pagesBeyond plan

03 Two strategic flags Your eyes needed

⚑ 1 · Scope has outgrown the roadmap

The product is no longer just a "Pre-Con OS" sprint — field-ops, the Company Brain, and the MEDDPICC engine are live. The plan should probably acknowledge that.

⚑ 2 · Sequence tension

The roadmap is sub-first (beta subs → GC pivot → mandate). The playbook we just built pivoted to GC-first ("head of the beast"). These need reconciling — either the roadmap updates to GC-first, or we frame sub-first as the bootstrap/data motion and GC-first as the scale motion.

04 What I'd build On your go

  • Full reconciliation pass — line-by-line roadmap items tagged Done / In-progress / Remaining / Beyond-plan, grounded in the codebase, with a re-baselined "we're really at ~Week N" marker.
  • Updated 52-week plan — revised content reflecting what's done, re-sequenced.
  • Interactive HTML timeline artifact — the 4 phases × 52 weeks, color-coded by status, "you are here" marker, in the playbook's design language (lives in perch-cro).
  • GTM & Sales Enablement plan — weaves the MEDDPICC playbook into the roadmap: which plays run in which phase, how sub-wedge → GC-mandate maps to the GC-first thesis, and the enablement cadence.

05 Questions before I start Answer or "use your judgment"

🔨 Decide these and I'll build
  1. Format for the updated plan — keep .docx I regenerate, convert to an HTML artifact in perch-cro (matching the playbook), or both?
  2. Sequence reconciliation — make the roadmap GC-first to match the playbook, or keep sub-first-bootstrap → GC-scale as two distinct motions?
  3. Source of truth for "done" — propose the done-list from the doc's annotations + my codebase reconciliation for you to approve, or will you hand me an authoritative done-list?
  4. MEDDPICC / GTM placement — a new section inside the roadmap doc, or a standalone "GTM & Sales Enablement" artifact that links roadmap ↔ playbook?
  5. Real names — OK to reference actual partners (Built, Zwick, Westland, Big-D, the $300M/$2B/$3B partners) in these internal CRO artifacts?
Perch · CRO workspace · roadmap-reconciliation Preliminary — confirm before external use