// projects · case studies
04 / client names private
// four builds · client names private
Four products that help people plan trips, understand product data, improve messy records, and produce videos—with practical controls around the AI.
Client names privateClient and product names stay private. We show the problem, what we built, how it works, and what we tested.
// how we build dependable AI products
We use AI for flexible work, while software and people control data access, costs, evidence, and important decisions.
Map the steps, decisions, and handoffs before asking AI to help.
Keep important answers tied to the data or material they came from, and say when evidence is missing.
Use software—not generated text—to enforce access, budgets, formats, checks, and critical calculations.
Ask for approval before costly, sensitive, or final decisions.
Give each tool and workflow only the data and actions it needs.
Save progress, avoid duplicate work, make failures visible, and support safe retries.
TravelTravel planning assistant
Client name privateWhat we built
We built a travel planner that turns a person’s dates, interests, pace, and priorities into a practical day-by-day itinerary using real places.
Product analyticsProduct analytics assistant
Client name privateWhat we built
We built an analytics assistant that lets teams ask questions about product usage in everyday language and get answers they can trace back to the underlying data.
Business dataData preparation workspace
Client name privateWhat we built
We built a workspace that imports inconsistent files, fills in requested information from approved sources, checks each result, and exports data ready for review.
Video productionAI video production workspace
Client name privateWhat we built
We built a video workspace where AI helps with research, planning, and production while people approve creative choices, costs, rights, and the final result.
// from difficult workflow to working product
We can turn it into a product your team can use, understand, and trust—from first framing through tested implementation.