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 |
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