Playbook

AI workflows & ethics

Leverage modern tooling without compromising code quality, client trust, or IP.

Field guide

Scoping AI-assisted builds

AI compresses implementation time — it doesn't remove discovery, QA, or client education. Quote human review hours explicitly.

Proposal line

AI-assisted development: Implementation accelerated with AI tooling. All output human-reviewed. Testing, a11y, and security verification billed at standard rates.

Avoiding AI debt

Generated code you can't explain is liability. Require tests, linting, and ownership docs before handoff.

Review checkpoint

// Before merge: can you explain every exported function?
// If not — refactor or rewrite before client handoff.

Drop-in assets

Copy, download, ship

Every asset exports as plain text, a Word document, or a PDF — no markdown files required.

Drop-in asset

Client-Facing AI Use Policy

Drop-in policy block for proposals and contracts.

Downloads as TXT, Word, or PDF — ready to edit and send.

AI Use Policy (Client-Facing)

What we use AI for
- Scaffolding boilerplate, draft copy, research summaries
- Test case generation and documentation drafts

What we do manually
- Architecture decisions, security review, accessibility audit
- Final code review before merge — AI output is never shipped unchecked

Data & privacy
- Client data is not submitted to public AI tools without written approval
- No training on client IP

Disclosure
We disclose AI assistance when it materially affects deliverables. You receive human-reviewed, owned codebase — not unverified generated output.