Kyle HobsonApplied AIIndependent applied AI consulting
I help teams find a practical use for AI, make the instructions and review points clear, and test the result against the task. Bring a workflow, an early idea or a team that needs hands-on guidance.
Work directly with Kyle Hobson. A defined scope, a tangible deliverable and room to ask questions.
Professional work · ASU Learning Innovations
My 2024 proposal at ASU placed mentor judgment before and after task generation. Diagram reconstructed from the proposal; deployment was not established.
Selected work
Examples from my own work and contributions within existing teams. Roles and delivery stages are identified in each story.
Arizona State University · Learning Innovations
At ASU, I proposed a staged workflow for personalized learner tasks, with mentor review of both context and generated output.
AI Analyst
Proposed prompt architecture · deployment unverified
Prompt Compass
I built and released Prompt Compass, then developed local comparisons that check whether a rewrite preserves the original task.
Independent creator · workflow and tool development
Public prerelease · later local evaluation examples
Coursera · Guided Project
A published Coursera project guides beginners through making prompts more specific, comparing outputs and refining a draft.
Guided Project Instructor
Published Coursera guided project · 2024
Ways to work together
An AI engagement should have a task, a way to judge the output, and someone responsible for the next decision. We agree on those before choosing the scope.
Find where AI could help a task and where people need to make the decision.
You receive
A workflow map, requirements, review points and a proposed next step.
See scope and deliverablesTurn an unclear request into instructions and a result your team can inspect.
You receive
A scoped prompt or prototype, evaluation examples and handoff notes.
See scope and deliverablesHelp your team practice prompting, reviewing and refining work that matters to them.
You receive
An agreed session, guided exercises and reference material to use afterward.
See scope and deliverablesI make the inputs, decisions and review points visible. Your team can see what the workflow is meant to do, try the agreed examples, and understand where human judgment is needed.
We define the user, the work, its constraints and what a useful result looks like. Existing tools and examples provide the starting point.
I develop the agreed workflow, prompt or prototype and examine it against representative examples. We document failures as well as promising results.
You receive the agreed materials, review criteria and a walkthrough. Ongoing deployment, maintenance or training is scoped explicitly.
My experience brings together AI analysis, UX research, technical communication and teaching. At ASU Learning Innovations, I worked as an AI Analyst; I also created a published Coursera guided project on working with ChatGPT.
I build tools as well as explain them. Prompt Compass, my independent prompt-engineering project, has a public beta release. That work keeps my consulting grounded in actual workflows, constraints and things a team can inspect.
No. Start with the task and the tools your organization already permits. We can identify the requirements and constraints before deciding what to try.
Yes. A workflow review, a prompt comparison or a focused workshop can be a useful starting point. We agree on the deliverable and how to judge it before the work begins.
The proposal names the workflow, inputs, output, examples, review criteria and handoff. A prototype is an opportunity to evaluate a defined use case. Production deployment, security review, integrations and continuing support require their own agreed scope.
We agree on permitted tools, data access and review responsibilities. I start with representative or non-sensitive examples where possible, document limitations, and identify what needs a human check. Please leave confidential records and credentials out of the inquiry form.
Those results depend on the workflow, adoption and how performance is measured. We can define a useful comparison, but I do not promise a percentage improvement before the work is examined.
I propose the scope, fee, schedule and review points after we discuss your task. The complexity of the workflow, access requirements, evaluation examples and training format affect the proposal.
Tell me about the task and what a useful result would look like. We can find a concrete starting point.