Where we use AI
Strategic Labs operates a multi-expert AI workforce on our partner platform Coheria. We use AI in:
- Research & opportunity audits — multi-expert AI panels analyze your stack, processes, and goals to surface high-leverage pilots.
- Solution design — AI experts propose architectures, RAG strategies, agent workflows, and evaluation rubrics; humans finalize.
- Implementation — AI authors first drafts of code, prose, configurations, and infrastructure-as-code; humans review every line that ships.
- Delivery operations — AI handles ticket triage, customer comms drafting, monitoring alert classification; humans approve client-facing messages.
- Internal operations — proposal authoring, sales follow-up drafts, and contract first-passes are AI-assisted; the operator owns every send.
Where we don't use AI
- Client identity verification — humans only.
- Final approval on client-facing deliverables — humans only.
- Contract negotiation — humans only (AI drafts; humans negotiate).
- Legal interpretation of regulations specific to your industry — we recommend qualified counsel; we do not give legal advice.
- Medical, fiduciary, or safety-critical decision-making for end users — out of scope at Strategic Labs.
What this means for you
- Every artifact you receive from us is reviewed by a senior human operator before handoff.
- Every claim, citation, or recommendation in our deliverables is auditable via Truth Oracle (we will share the audit log on request).
- We disclose AI use in every client communication when asked directly.
- We do not pass off AI-generated work as fully human-authored when asked. If a client requires an all-human work product, we'll tell you upfront whether we can do it (often: yes, just at a different price and timeline).
Your data
Your data, your prompts, and your context are processed inside your tenant. We do not train shared models on your data. We retain operational logs for the duration of the engagement (and per the retention period in our DPA) and then purge.
Models we use
We rotate across leading frontier and open-weight models (Claude, GPT, Gemini, GLM, DeepSeek, Llama, Qwen, Mistral, others) depending on the task, your residency requirements, and cost considerations. The Truth Oracle pipeline runs every important decision through multiple models so no single model's biases dominate the output.
Questions
Email hello@strategiclabs.ai with subject AI Disclosure.