What a cohort actually ships.

Problem

Senior associates were using ChatGPT on personal accounts to draft client memos and run document review. Leadership wanted them to keep the productivity but lose the data-exposure risk.

Delivered

Four-week cohort. Week 1: prompt engineering on the firm's enterprise-tier tool. Week 2: data classification + redaction patterns + privileged-data handling. Week 3: Word + Outlook + Excel automation patterns. Week 4: each associate shipped a prompt library covering five of their highest-frequency tasks.

Outcome

Capstone deliverables: 25 prompt libraries, each averaging 38 templates. Average self-reported time saved per associate per week: 4-6 hours. Zero incidents in the post-cohort 30-day window.

Problem

Operations leadership wanted the team to stop being scared of AI tools, start using them where appropriate, and not use them where not appropriate. They wanted this in one half-day session, not a four-week cohort.

Delivered

Half-day live bootcamp. Two 90-minute modules: (1) what AI actually is and where it works, with hands-on practice on the team's tool of choice; (2) the four-bucket data-classification framework, applied to the team's actual workflows.

Outcome

Post-session survey: 92% reported the framework was 'immediately applicable' to their work. Six attendees nominated themselves to become internal AI champions inside their sub-teams.

Problem

C-suite wanted to set policy on AI use across the company but did not have a shared mental model of what AI does, how it fails, and where the policy boundaries should be.

Delivered

Two-hour workshop. One hour of plain-English explainer using the [whatisai.io](https://whatisai.io) curriculum. One hour of policy workshop: the team drafted the company's first AI acceptable-use policy on the whiteboard.

Outcome

Draft AI policy in hand at end of session. Final policy circulated by legal within two weeks. Saved an estimated 3-4 months versus the original 'have legal write it from scratch' path.

Problem

Civilian engineering staff at a federal district office need to integrate AI tools into their day-to-day work — drafting, document review, memo summarization — without violating internal-data handling rules or the agency's IT acceptable-use policy.

Delivered

Proposed scope, 4-week SDVOSB set-aside delivery. Week 1: prompt engineering on the agency's approved AI tool. Week 2: internal-data privacy guidance aligned to FAR 52.224-3 + agency-specific records-handling rules. Week 3: no-code automation patterns for the engineering-staff workflow. Week 4: capstone with each attendee shipping a reusable prompt library.

Outcome

Scope and approach delivered in response to USACE Kansas City RFQ (W912DY-26-Q-0001). Apache-3 prime, no subcontracting layer. Award pending.