Bobby Reed · Builder, Integrator & Educator
I put AI to work for real teams -- and teach them to keep it working. Production chatbots and LLM integrations for clients. A decade of workflow automation that turned chaotic operations into documented, self-sustaining systems. And a career spent as the approachable translator between the technology and the people it serves. My own practice runs on AI infrastructure I built; I am the first user of everything I recommend.
AI chatbots & integrations in client service
Workflow automation in real operations
Every audience taught, no jargon required
User of every system I recommend
AI adoption fails the same way most technology fails: built from the tool backward instead of from the process forward. I start in people's offices, not at my desk -- learning what a team actually does, where the friction lives, and what "better" would mean to them. Then I build the smallest thing that helps, document it so it survives me, and teach until the team owns it. The goal is never the demo. The goal is the Tuesday afternoon six months from now when the system is still quietly working.
Discovery starts with the team's process, not the technology. The best automation candidates are found in conversation, not in a features list.
My depth is applied AI: LLM integrations, chatbots, retrieval, and process automation that plug into the tools teams already use.
Every system ships with documentation and training. At OU, my student workers owned and improved the automated workflows by their second year.
A decade of Zapier, Trello, scripting, and pipeline design -- treating repetitive work as a signal that a system wants to exist.
Faculty, library directors, branch staff, graduate students -- a career of making technology make sense to people who have other jobs to do.
Federal grant reporting taught me outcome discipline: define the metric, track it honestly, and report it whether it flatters the project or not.
AI in production, automation in operation, and the systems that keep working after the launch excitement fades.
Through DEV Limited: production chatbots and LLM API integrations for institutional and commercial clients -- scoped from each team's actual process problems, shipped into the tools they already use, and maintained as their needs evolve. The unglamorous truth of applied AI: the model is the easy part; the integration and adoption are the work.
Rebuilt multi-location 3D printing operations on Zapier, Trello, and Google Drive automation: intake, scheduling, tracking, and handoff ran themselves. Documentation and training were strong enough that student workers became the workflows' owners and improvers by their second year -- most still use those skills professionally today.
When a cyberattack corrupted a publication's database, custom Python pipelines did what no manual process could: recovered, cleaned, and restructured 1.4 GB and 10,000+ pages into a deployable archive -- automation built under pressure, documented for the record.
Client delivery, course management, project tracking, and a job-search operations pipeline -- all running on AI-assisted infrastructure I designed and refine continuously. I recommend nothing I have not lived with. The first user of every system I build is me.
The other half of AI solutions work: helping the humans. I have taught technology to every kind of audience for a decade -- AI is just the newest subject.
OCALD 2023 · Translating AI for institutional decision-makers.
Medical Library Association · Designed and taught this continuing-education course for working professionals.
Oklahoma City University · Teaching graduate software engineering students to build with AI tools -- critically, productively, and honestly.
College & Research Libraries News, 2024 · The craft of inviting non-specialists in -- the skill AI education depends on.
Applied-AI depth: integration, automation, deployment, and adoption.
If your team needs someone who can find the friction, build the solution, and teach everyone to run it -- and who measures success by whether it still works in six months -- I would welcome the conversation.