Operating Infrastructure Co.Companion Brief
v1.0 OPS-ARCH-2026-01 May · 2026
One-Page Brief · Companion to Architecture Doc

The operating layer underneath your AI.

DataBrainOutputGoverned by default

We install the operating layer underneath your AI tools — the validated, governed source of truth every department draws from — for knowledge-work SMBs ready to make AI useful across the company.

Three layers · core architecture
01

Data Layer

Connect · clean · classify · permission. SOPs, CRM, email, proposals, quotes, tickets.
Buyer promiseAI can find and use our information.
02

Brain Layer

Decision rules · retrieval · prompts · guardrails. Encodes SOPs, approvals, tone, memory.
Buyer promiseAI understands how we work.
03

Output Layer

Email · chat · CRM · docs · forms · tickets. Lands inside existing tools, not a new app.
Buyer promiseMy team uses it without learning another system.
Ideal customer
  • 35–150 employees, $10M–$50M revenue
  • Cross-functional content, decisions, and specs flow daily; brand and accuracy both matter
  • Digital agencies, strategy consultancies, B2B SaaS / product cos, marketing & creative firms, professional services
  • Modern stack + exec sponsor + paid-audit willingness
Service ladder · commercial model
A

AI Infrastructure Audit

Diagnostic + roadmap + buildout SOW

$3.5–10K2–3 wks · entry
B

AI Business Brain Buildout

Install first 1–2 workflows end-to-end

$15–60K6–12 wks · build
C

Managed AI Operations

Run · tune · expand · QBR every 90 days

$2–12K/moongoing · operate
Workflow kits
  • Kit A · Business Brain  · retrieval over SOPs & approved content
  • Kit B · Brand & Voice Source of Truth  · validated brand reference all marketing AI draws from
  • Kit C · Customer Operations  · draft customer replies grounded in approved product truth
  • Kit D · Cross-Functional Drafting & Approval  · route AI drafts through approvals, with audit trail
Governance baseline
  • Role-based access mirrors source ACLs
  • Source-of-truth rules per data class
  • Human approval gates on external actions
  • Audit log of every AI invocation
Year 1 · revenue model
Audits, 4/mo × $5K × 10mo$200K
Buildouts, 2/mo × $30K × 10mo$600K
Retainers, 10 × $4K × 8mo$320K
Total Y1 revenue$1.12M
55–75%
Gross margin
~$45K
Avg ACV
~29%
Recurring Y1
Buying triggers
  • Senior brand or product lead leaving, taking institutional voice with them
  • PE sponsor or board asks "what's the AI plan?"
  • Cross-functional review backlog is slowing launches
  • A prior internal AI experiment quietly failed
  • Customer-facing inconsistency went public — marketing page contradicted what product shipped
Defensible wedge
  • Cross-functional governance IP, not technology
  • SOP & decision-rule encoding
  • Governance scaffolding from day one
  • Productized, templated delivery
Implementation roadmap · six phases
Phase 1

Validate

Wks 1–8 · interviews, demo, sell 3 audits

Phase 2

Productize

Mo 3–4 · templates, governance baseline

Phase 3

Deliver

Mo 4–8 · first buildouts, retainer attach

Phase 4

Managed Ops

Mo 8–12 · SLA, dashboards, QBR motion

Phase 5

Verticalize

Mo 12–18 · 2–3 vertical workflow kits

Phase 6

Scale

Mo 18+ · channel + repeatable sales

8-week sprint · validate the wedge
Week 1

Interview 10 customers

Beachhead verticals · pain & WTP map

Week 2

Build vertical demo

Anonymized data · 15-min walkthrough

Week 3

Sell 3 paid audits

3 × $5K · signed SOWs · references

Weeks 4–8

Convert 1–2 to buildouts

First case study seed · template v0

Operating Infrastructure Co. · One-Page Brief OPS-ARCH-2026-01 · 01