Live instruction that follows your work
Learn through finance and accounting tasks, with space to ask questions, try an approach, and review the result.
AI Center of Excellence · Devon Coombs LLC × Gaapsavvy
With Gaapsavvy, I bring Google Cloud accounting experience and Santa Clara University teaching to your team’s AI workflows. Learn, build, and review together through recurring sessions and practical support.
Devon Coombs, CPA, MBA · Assistant Teaching Professor of Finance and Faculty Director of Real Estate Finance, Santa Clara University
“Through training, workflow development, and direct support, he has helped our teams use AI more effectively in research, analysis, reporting, documentation, and decision-making.”
The recurring program
We focus on workflows your team owns. The scope defines the participants, priority use cases, support boundaries, and how we will assess progress.
Learn through finance and accounting tasks, with space to ask questions, try an approach, and review the result.
Work through the practical issues between a prompt and a repeatable process: inputs, task steps, output format, and review.
Choose useful applications, identify internal owners, and decide which work needs a controlled calculation or human judgment.
A repeatable approach
Identify a real task, the available inputs, and what a good result should look like. Agree a baseline and a way to assess the outcome.
Separate reliable calculations from work suited to an AI model. Provide appropriate context and define checks before the output is used.
Bring back what worked and what failed. Improve the workflow and its documentation, then help colleagues apply it consistently.
The teaching behind the program
PagerDuty · Controllership onsite · July 2026
A three-day onsite brought together about 40 controllership leaders.
Both improvement responses asked for more time with Claude or more workshops. Nine written responses across two questions; not ratings from all attendees.
Read the participant feedbackFrom a session to ongoing support
A participant can learn an approach in a session and still need help applying it to a new report, contract, or analysis. The recurring program gives your team somewhere to bring that work.
We use your approved platforms, including Microsoft, Google, OpenAI, and Anthropic tools. Together, we work through the inputs, instructions, calculations, and review each task needs.
The PagerDuty example documents a workshop I delivered; it is evidence of the teaching approach, not an evaluation of the recurring AI CoE program.
Explore a standalone workshopYour advisors

Assistant Teaching Professor of Finance and Faculty Director of Real Estate Finance at Santa Clara University. Founder of Devon Coombs LLC, with experience at Deloitte, Riveron, and Google Cloud.
SCU faculty profile
Founder of Gaapsavvy and partner in the AI Center of Excellence. Brings technical accounting experience to the program’s learning and workflow support.
Explore GaapsavvyThis private program is offered through Devon Coombs LLC and Gaapsavvy. Devon’s academic appointment at Santa Clara University is separate from the engagement.
Before you begin
It is a recurring program of live training, workflow support, and advisory guidance for your team, offered by Devon Coombs LLC in partnership with Gaapsavvy. We help participants apply AI to real work and develop the judgment to review the results.
A workshop is a focused learning event. The AI Center of Excellence provides recurring sessions and support as people apply what they learned, bring back questions, and develop additional workflows.
The program works across Microsoft, Google, OpenAI, and Anthropic tools. We plan around the platforms your organization approves and the access your participants have.
The program includes two live one-hour sessions each month and up to ten hours of monthly support. We agree the participants, priorities, support boundaries, and commercial terms in the scope.
A paid 30-day pilot is available. We define the initial use cases, participants, success criteria, and scope before starting. Ongoing fees depend on the engagement.
We agree a baseline and useful measures for the selected workflows, such as time required, quality of output, rework, and adoption. Results depend on the task, data, tools, and the team’s participation.
Find the right starting point
Tell us about your team, approved tools, and the work you want to make more effective. We can discuss a paid pilot or an ongoing program.