Applied AI Architect — Commercial & Enterprise
When to Use
- Design end-to-end AI architecture for a B2B product feature or internal copilot
- Choose between RAG, fine-tuning, agents, or workflow orchestration for a use case
- Define multi-tenant isolation, data boundaries, and customer data handling for AI
- Map enterprise requirements: SSO, audit logs, residency, DLP, human oversight
- Compare cloud AI platforms (managed endpoints, private networking, quotas)
- Author AI-specific ADRs for engineering, security, and procurement
- Plan POC → pilot → production with eval, safety, and ops gates
When NOT to Use
- Write application code, prompts, or eval harnesses →
ai-engineer, prompt-engineer
- General non-AI integration ADRs →
senior-system-architecture
- Context packing and token budgets →
ai-context-engineer
- Memory store implementation →
ai-memory-developer
- AI risk tiering and policy documents →
ai-risk-governance
- SOC 2 evidence packs →
compliance-engineer
- Token cost improvement program →
ai-token-improvement-plan-engineer
- AI ops cadence and release governance →
ai-lead-ops
- Business case and operating model →
business-consultant
Related skills
| Need |
Skill |
| Build and ship RAG/agents |
ai-engineer |
| Cross-service platform ADRs |
senior-system-architecture |
| Governance and use-case classification |
ai-risk-governance |
| Security architecture and controls |
cybersecurity, information-security-engineer |
| Data platform for embeddings/warehouse |
data-architect |
| Product tenant isolation |
product-infrastructure-security-engineer |
| Red-team before launch |
ai-redteam |
| Commercial contract terms |
commercial-counsel |
Core Workflows
1. Classify context: commercial vs enterprise
| Lens |
Commercial (product) |
Enterprise (internal IT) |
| Users |
Customer tenants |
Employees, partners |
| Data |
Customer content + config |
Corp IP, HR, regulated data |
| Identity |
Product auth + tenant ID |
IdP (Entra/Okta), groups |
| Isolation |
Strong multi-tenancy |
Network/VPC, private endpoints |
| Buying |
SKU, usage metering |
EA, private offer, on-prem option |
See references/commercial_vs_enterprise.md.
2. Shape the solution
- Job — user outcome in one sentence; success metric
- Pattern — Q&A RAG, task agent, workflow automation, codegen assist
- Data — sources, freshness, PII, retention, who may see what
- Model — capability vs cost; routing; fallback
- Control plane — prompts, tools, policies, human approval points
- Plane separation — ingestion ≠ inference ≠ logging (scale independently)
See references/reference_architectures.md.
3. Non-functional requirements
Document targets for:
- Latency (p95 first token, end-to-end task)
- Availability and DR for inference path
- Cost per tenant/session and guardrails
- Auditability (prompt, retrieval IDs, model version—no secrets in logs)
- Safety (tier, blocked topics, escalation)
- Compliance (residency, retention, subprocessors)
See references/enterprise_nfr_integration.md.
4. Platform and vendor selection
Score options on: data handling, private link, model catalog, SLAs, metering, exit strategy.
See references/platform_selection.md.
5. AI ADR and review
For one-way doors (vendor, data leaves region, autonomous agents):
- Options, NFR impact, security, cost model
- Migration and rollback
See references/ai_architecture_decision.md.
Pair with senior-system-architecture for org-wide integration standards.
6. POC → production path
| Stage |
Architecture focus |
| POC |
Single tenant, manual eval, no prod data |
| Pilot |
Tenant isolation, observability, limited users |
| GA |
SLOs, rate limits, on-call, eval in CI, red-team for tier-2+ |
See references/poc_to_production.md.
When to load references
- Commercial vs enterprise →
references/commercial_vs_enterprise.md
- RAG, agent, copilot patterns →
references/reference_architectures.md
- SSO, DLP, logging, residency →
references/enterprise_nfr_integration.md
- Cloud AI platform choice →
references/platform_selection.md
- AI ADR template →
references/ai_architecture_decision.md
- Production hardening →
references/poc_to_production.md
1---2name: applied-ai-architect-commercial-enterprise3description: Guides applied AI solution architecture for commercial (B2B SaaS, multi-tenant product) and enterprise (internal copilots, regulated corp data) contexts—reference patterns for RAG, agents, and copilots, platform and model selection, data boundaries, identity, observability, cost at scale, and POC-to-production hardening with security and governance gates. Use when designing customer-facing AI features, enterprise knowledge assistants, vendor LLM integration architecture, AI ADRs for commercial or IT stakeholders, or mapping AI components to SSO/DLP/residency requirements—not for implementing RAG code (ai-engineer), generic multi-service ADRs without AI (senior-system-architecture), AI policy registers (ai-risk-governance), or contract redlines (commercial-counsel). For engineering management of vertical AI product squads (hiring, roadmap, launch governance), use engineering-manager-vertical-ai-products. Labeling PM: product-management-human-data-platform—not this skill.4---56# Applied AI Architect — Commercial & Enterprise78## When to Use910- Design end-to-end AI architecture for a B2B product feature or internal copilot11- Choose between RAG, fine-tuning, agents, or workflow orchestration for a use case12- Define multi-tenant isolation, data boundaries, and customer data handling for AI13- Map enterprise requirements: SSO, audit logs, residency, DLP, human oversight14- Compare cloud AI platforms (managed endpoints, private networking, quotas)15- Author AI-specific ADRs for engineering, security, and procurement16- Plan POC → pilot → production with eval, safety, and ops gates1718## When NOT to Use1920- Write application code, prompts, or eval harnesses → `ai-engineer`, `prompt-engineer`21- General non-AI integration ADRs → `senior-system-architecture`22- Context packing and token budgets → `ai-context-engineer`23- Memory store implementation → `ai-memory-developer`24- AI risk tiering and policy documents → `ai-risk-governance`25- SOC 2 evidence packs → `compliance-engineer`26- Token cost improvement program → `ai-token-improvement-plan-engineer`27- AI ops cadence and release governance → `ai-lead-ops`28- Business case and operating model → `business-consultant`2930## Related skills3132| Need | Skill |33|---|---|34| Build and ship RAG/agents | `ai-engineer` |35| Cross-service platform ADRs | `senior-system-architecture` |36| Governance and use-case classification | `ai-risk-governance` |37| Security architecture and controls | `cybersecurity`, `information-security-engineer` |38| Data platform for embeddings/warehouse | `data-architect` |39| Product tenant isolation | `product-infrastructure-security-engineer` |40| Red-team before launch | `ai-redteam` |41| Commercial contract terms | `commercial-counsel` |4243## Core Workflows4445### 1. Classify context: commercial vs enterprise4647| Lens | Commercial (product) | Enterprise (internal IT) |48|---|---|---|49| Users | Customer tenants | Employees, partners |50| Data | Customer content + config | Corp IP, HR, regulated data |51| Identity | Product auth + tenant ID | IdP (Entra/Okta), groups |52| Isolation | Strong multi-tenancy | Network/VPC, private endpoints |53| Buying | SKU, usage metering | EA, private offer, on-prem option |5455**See `references/commercial_vs_enterprise.md`.**5657### 2. Shape the solution58591. **Job** — user outcome in one sentence; success metric602. **Pattern** — Q&A RAG, task agent, workflow automation, codegen assist613. **Data** — sources, freshness, PII, retention, who may see what624. **Model** — capability vs cost; routing; fallback635. **Control plane** — prompts, tools, policies, human approval points646. **Plane separation** — ingestion ≠ inference ≠ logging (scale independently)6566**See `references/reference_architectures.md`.**6768### 3. Non-functional requirements6970Document targets for:7172- Latency (p95 first token, end-to-end task)73- Availability and DR for inference path74- Cost per tenant/session and guardrails75- Auditability (prompt, retrieval IDs, model version—no secrets in logs)76- Safety (tier, blocked topics, escalation)77- Compliance (residency, retention, subprocessors)7879**See `references/enterprise_nfr_integration.md`.**8081### 4. Platform and vendor selection8283Score options on: data handling, private link, model catalog, SLAs, metering, exit strategy.8485**See `references/platform_selection.md`.**8687### 5. AI ADR and review8889For one-way doors (vendor, data leaves region, autonomous agents):9091- Options, NFR impact, security, cost model92- Migration and rollback9394**See `references/ai_architecture_decision.md`.**9596Pair with `senior-system-architecture` for org-wide integration standards.9798### 6. POC → production path99100| Stage | Architecture focus |101|---|---|102| POC | Single tenant, manual eval, no prod data |103| Pilot | Tenant isolation, observability, limited users |104| GA | SLOs, rate limits, on-call, eval in CI, red-team for tier-2+ |105106**See `references/poc_to_production.md`.**107108## When to load references109110- **Commercial vs enterprise** → `references/commercial_vs_enterprise.md`111- **RAG, agent, copilot patterns** → `references/reference_architectures.md`112- **SSO, DLP, logging, residency** → `references/enterprise_nfr_integration.md`113- **Cloud AI platform choice** → `references/platform_selection.md`114- **AI ADR template** → `references/ai_architecture_decision.md`115- **Production hardening** → `references/poc_to_production.md`
daemon-blockint-tech