The Perch Centaur Timeline

52 weeks · Jul 1 2026 → Jun 30 2027. Not a vague roadmap — an operating blueprint. Four phases (your roadmap's own), recalendared to the MVP launch. Each phase carries three swimlanes — Product, PLG, and SLG + MEDDPICC — so we see exactly when the sales engine turns on relative to what ships. Toggle the motion: GC-first (preferred) or Sub-first. Full acronym glossary at the bottom.

🔥

Sell from the MVP, not the finish line.

SLG + MEDDPICC start Week 1 (Jul 1, 2026) — prospecting, discovery, qualification, and lighthouse betas run while the product matures. We never wait for "done." PLG and SLG ramp together as capability lands.

◉ You are here

Late Jun 2026 — pre-launch. Most of Phase 1's product is already shipped or in flight (we're ahead of the original plan). MVP target: Jul 1, 2026 (assumption — confirm). The clock below starts there.

◆ Running from Day 1 — not a later phase

Per Jeremy + Brock: the sub-portal, the relationship graph, and the "Mandate" pull-through are already built and being refined — they activate at launch, not in Q3. We run these every day from Jul 1. (Q3 below shows only what genuinely must be newly built.)

Product — already live, refining

Token-based sub-contractor bid portal (landing, claim/signup, Scope · Documents · AI-Proposal · Bid-Summary workspace)
Bid invitations + submissions (GC invites subs by division; status tracking) — built, not fully tested yet · QA before we lean on it
Relationship data: CRM companies/contacts, encounters, Deal Board, Forecast, Account-360
Doc knowledge graph (Neo4j: documents, sections, entities, coordination)

Motion — low-hanging fruit, do daily

The Mandate (light): recruit subs through the GCs we land — every GC bid invite onboards its subs
SDRs work subs surfaced inside GC bid funnels — warmest leads we have
Build the pipe: contacts, leads, and SDR outreach from Jul 1 (see Sales Engine)
Dogfood + referrals via Chris @ Built Exteriors (friends-and-family)

The Elite Wedge

Q3 2026 · Jul–Sep · Weeks 1–12

$1M ARR

Phase goal

TargetsSMB + warm networkBuilt dogfood + referrals (Chris)First cold logo = the milestone

Productwhat ships

MVP launch (Jul 1) — agentic doc pipeline → SOW generation shipped

Glyph-level traceability / reconstruction shipped

Entity extraction + spaCy labeling shipped

Full division / section trade coverage — completing & tuning partial

What it means: every uploaded plan/spec page is auto-tagged to its CSI MasterFormat Division + Section and to one of 34 construction trades — then synthesized into per-division scope (work items, submittals, coordination, materials).
Built today: spec-affinity vocab for Divisions 00–14, 21–28, 31–35 — which are the active building-construction divisions; 34 trade pattern packs (masonry, metals, millwork/casework, electrical, mechanical, concrete…); Neo4j Section nodes; V3 division-scope synthesis feeding the job scope view + bid packages.
✅ Correction (Jeremy was right): Divisions 15–20, 24, 29, 30 are reserved / unused in modern MasterFormat — not gaps. The 2004 expansion retired old Div 15 (Mechanical) & 16 (Electrical) and redistributed them into 21–28. Having no vocab for the reserved numbers is correct; our coverage already spans the divisions a real building bid touches.
Real remaining work: per-division quality tuning (04/05/06/26 have vocab but extraction quality varies by project type), and — only for industrial/process jobs — the Process Equipment Divisions 40–49. "Full coverage" = consistent quality across the active divisions, not filling reserved holes.

Hallucination harness / V3 eval + tuner partial

PLGself-serve

"Magic moment" trial — small/mid GCs upload a bid package → instant SOW + division scope (Core tier) live at MVP

Sub pull-through: every GC bid invite onboards its subs self-serve live

Sub self-serve trial via the $300M partner's sub network — sub uploads an ITB → scope + AI proposal (Essentials/Growth) live at MVP

Portal claim/signup converts invited subs into accounts live

SLG + MEDDPICCsales engine

🎯 Phase-1 goal: land our first client OUTSIDE of Chris. Chris @ Built Exteriors is our friends-and-family dogfood + referral source — the milestone is the first cold, self-won logo. the proof point

Build the pipe from Day 1. Stand up contacts/leads, SDR outreach cadences, and MEDDPICC qualification on every opp from Jul 1 — don't wait for product polish. on at MVP

SDR prospecting of SMB + warm-network GCs; AE discovery calls; lighthouse lunches/demos core play

SDR/AE on warm SMB subs via the partner network + Built referrals; MEDDPICC on the bigger ones core play

📞 Q1 daily cadence: ~22 SMB logos to hit $1M → see worked funnel math + per-rep daily targets in the Sales Engine, and customer-makeup paths in $1M Scenarios.

The Enterprise Pivot

Q4 2026 · Oct–Dec · Weeks 13–26

$15M ARR

Phase goal

TargetsGC $10M–$500M revSub $10M–$250M rev

Productwhat ships

GC Command Center — today this is distributed across Jobs kanban + Analytics + Workbench; the build is to unify it to build

Unified multi-job control surface (one screen: pipeline kanban, calendar, at-risk jobs)
GC bid-buyout orchestration — invite subs, track coverage by division, leveling queue (today's orchestrator is sub-pursuit oriented)
Analytics KPIs wired live: win rate, pipeline value, bids out, cycle time
GC-side SOW/scope path (currently scope-doc generation is sub-only)

Deep Diff engine — addenda/revision intelligence to build

Have today: addendum classification, dual trade+spec affinity on addendum pages, cross-doc entity comparison, document versioning + reprocessing.
To build: baseline-vs-addendum section-level diff ("Addendum 3 changed Div 07 waterproofing"), page redline, and a structured change artifact pushed to affected scope.

Analytics — KPIs, usage, funnel, at-risk jobs partial

PLGself-serve

Smaller GCs self-serve on Core/Pro; in-product upgrade prompts live

Note (Jeremy): self-serve is for smaller companies only. Enterprise is NOT self-serve — see white-glove note in the SLG lane.

Usage-based PQLs auto-routed to AEs (PLG→SLG handoff) partial

Sub self-serve scales across partner network; tier upgrades in-product live

Subs referring GCs upward → early GC PQLs partial

SLG + MEDDPICCsales engine

Start chasing enterprise — but chasing ≠ landing. reality check

Note (Jeremy + Brock): giants like D.R. Horton or Big-D will likely require us to refine, harden, and scale the product before we can win them. Q4 = enter the arena and start multi-threaded relationships; don't assume a logo just because we're in the room.

Build the white-glove motion — dedicated onboarding/implementation, SE support, exec sponsorship — before we pitch enterprise in earnest to build

Full MEDDPICC discipline on $10–500M GCs: Champion-building, EB access, Decision Criteria/Process, Paper core play

AE + SE pairs; first CSM hires for the bigger accounts ramping

Pivot MEDDPICC weight toward GCs while scaling SLG on $10–250M subs begins

📞 Q2 cadence shifts: fewer, larger deals · multi-threaded selling (3+ contacts/account) · longer cycles. Per-rep targets in the Sales Engine.

The Viral Flywheel

Q1 2027 · Jan–Mar · Weeks 27–39

$50M ARR

Phase goal

Targets — step upGC $500M–$1.5B revSub $250M–$1B rev

Productnet-new only

Sub-portal & relationship data already live (see Day-1 band) — Q3 is only what must be newly built context

Complete the sub-portal (today the shell is live; these tabs are placeholders) to build

Documents tab → shared plan room (currently "will appear here" placeholder)
AI Proposal generation in-portal (currently placeholder)
Bid-submission verification — auto-check a submitted bid for completeness vs. the spec/scope
Multi-GC unified view — one sub workspace showing competing GC invitations side-by-side (today only de-duped at job creation)

AI bid leveling — the comparison matrix exists but is manual numbers only to build

Normalize bids for exclusions/qualifications into adjusted apples-to-apples totals
Quantify scope-gap dollar impact per bid; flag coverage holes by division

Cross-org bid-flow graph — the net-new network capability to build

Today: per-org doc graph (Neo4j) + CRM encounters. There is no marketplace-wide "who-bids-whom" graph.
Build the graph-native model of sub → GC → job bid relationships to power routing, recommendations, and the true network effect.

PLGself-serve

The Mandate + SDR-into-GC-funnels are Day-1 plays, not Q3 (see top band) running

Amplify the loop the new builds unlock: completed portal + bid-flow graph make every GC's bid list a higher-converting sub-acquisition funnel scaling

SLG + MEDDPICCsales engine

MEDDPICC on $500M–$1.5B GCs; land-and-expand across divisions within won accounts core play

CS-led sub expansion inside GC accounts; scaled AE pods ramping

Convert the $250M–$1B sub base into anchored accounts; GC enterprise pods scale core play

📞 Q3 cadence: expansion motions + bigger logos. Activity model in the Sales Engine.

The Industry Standard

Q2 2027 · Apr–Jun · Weeks 40–52

$100M ARR

Phase goal

Targets — the giantsGC $2B–$3B rev (ENR Top 400)

Productwhat ships

RSMeans line-item costing → full Pre-Con OS not started

RSMeans = the standard unit-cost database for estimating. Zero code today — net-new integration to turn scope into priced line items.

RFI automation — code exists but is disabled re-enable + wire UI

RFI generation agent is built but skipped in the pipeline ("re-enable when RFI UI is wired"). Q4 = wire the UI + turn it on. Same for the disabled Risk-Analysis and Bid-Decision agents.

SOC 2 Type II — a marketing claim today; the real audit/cert must be earned to unlock the giants not started

Heads-up: Q4 product is mostly scale, security, and future features that aren't fully fleshed out yet. As we learn from Q1–Q3 selling, this list will sharpen — treat it as direction, not spec.

PLGself-serve

The viral loop is a Day-1 motion — by Q4 it's mature, not new running

Specific PLG to reach $100M: scaling levers

Every GC bid invite = sub onboarding at full volume across the network
In-product upgrade nudges (Core→Pro, Essentials→Growth) tied to usage thresholds
Usage-based PQL routing — high-intent self-serve accounts handed to AEs automatically
Referral loops — invited subs invite their GCs and vice-versa; activation-based incentives
Land-and-expand seats — more users/divisions per account self-serve

SLG + MEDDPICCsales engine

AI-SDR on the ENR Top 400 — automated prospecting at scale, feeding human AEs partial

Scaled AE org running MEDDPICC at volume; multi-threaded enterprise pursuits; land-and-expand across divisions core play

📞 Daily activities by role (every quarter) — exact SDR / AE / CSM cadences to defend the path to $100M live in the Sales Engine.

✦ Already shipped — ahead of / beyond the original roadmap

MEDDPICC sales engine — sales-orchestrator agents (relationship_researcher, scope_gap_analyst, meeting_briefer), qualification scoring, MAP steps (not in the roadmap at all)
Company Brain, agent guardrails, brain-grounded citations
Field-ops, CRM / relationship data (Deal Board, Forecast, Account-360), tribal knowledge, legal playbooks

📖 Glossary — every acronym, defined

ACV — Annual Contract Value: yearly revenue from one customer.

AE — Account Executive: the closer who owns the deal end-to-end.

ARR — Annual Recurring Revenue: total annualized subscription revenue.

CSI MasterFormat — industry-standard numbering that organizes a project into Divisions (e.g. 04 Masonry) and Sections (e.g. 04 20 00 Unit Masonry).

CSM — Customer Success Manager: owns onboarding, adoption, and renewal/expansion.

ENR — Engineering News-Record: trade publication; its "ENR Top 400" ranks the largest U.S. contractors by revenue.

GC — General Contractor: runs the project, buys out and manages the subs.

ITB — Invitation To Bid: a GC's request asking subs to submit a bid.

MAP — Mutual Action Plan: the shared step-by-step plan a rep and buyer agree to drive a deal to close.

MEDDPICC — Metrics · Economic buyer · Decision criteria · Decision process · Paper process · Identify pain · Champion · Competition. Our deal-qualification discipline.

MVP — Minimum Viable Product: first version real customers pay for.

PLG — Product-Led Growth: the product drives signup/activation/expansion (self-serve).

PQL — Product-Qualified Lead: a self-serve user whose behavior signals buying intent.

RFI — Request For Information: a formal question to the design team to clarify the documents.

RSMeans — Gordian's standard published database of unit costs (material + labor) for estimating.

SDR — Sales Development Rep: prospects and books qualified meetings for AEs.

SE — Sales Engineer: technical seller who runs demos and proves product fit.

SLG — Sales-Led Growth: humans drive the deal.

SOC 2 Type II — a security audit testing controls over time; required by enterprise buyers.

SOW — Scope of Work: the structured list of what a trade is responsible for on a project.

Sub — Subcontractor: the specialty trade (electrical, masonry…) that bids to the GC.

Perch · CRO workspace · centaur-timeline Preliminary · status = Jeremy + codebase judgment · confirm before external use