Add AI to Your Workflow
I look at how the work happens today, find the steps where AI actually helps (reading documents, searching what you already know, drafting, sorting) and build it into the tools you already use, with clear limits on what it does without a person checking.
How it works
For teams losing hours every week to work a model could do.
What's included
- Workflow mapping
- RAG over your documents
- Assistants & agents
- Fits existing systems
- Human review built in
Who it's for Teams that already run a process (documents to check, questions to answer, data to move between systems) and spend hours a week on the parts a model could handle.
How it works
- Watch the work. I map each step with the people who do it, including the workarounds nobody wrote down.
- Pick the steps worth automating. Usually two or three, not the whole process. Each gets a rule for what the AI may do alone and what a person must confirm.
- Build it into what you already use, instead of another tool your team has to learn.
- Measure it against how long the step took before.
Where I've done this
- Career Genie, my final-year project, answers from a verified career and scholarship dataset through RAG, so it can't invent a scholarship.
- The chat on this site answers only from published work, and the AI tools I use to draft posts can create drafts but never publish.
Not a fit if You want a chatbot because a competitor has one. Start from the step that costs you time.
Related work
Where I've Done This
Career Genie: AI Career and Scholarship Guidance
My final-year project: a prototype that matches students to career paths and scholarships from a verified dataset, using RAG so each recommendation traces back to a real source.
Let's build the right thing
Tell me what you're building, or what's slowing your team down. I'll tell you honestly whether I can help and what I'd do first.