Keller Williams Realty International (KWRI) · Case study
A Faster You with AI: the transaction playbook
I researched the full listing and buyer processes and authored A Faster You with AI. The playbook mapped practical AI use to each task, with pressure-tested prompts for familiar models and setup guidance for specialized tools.
Corporate headquarters · Industry & Learning. Research, prototypes, and enablement for the Keller Williams network of agents, Market Centers, and regional teams.
The starting point
Agents needed to see where AI fit in the work they actually did, from the start of a transaction through closing. A general list of AI tools did not explain which tool to use, how to begin, or when professional judgment was still required.
My responsibility
I led process research, mapped the transaction, researched and tested tool options, developed prompts and process guidance, and worked through editorial and legal review before presenting the material.
Tools & methods
Agent research · Workflow mapping · Claude · ChatGPT · Gemini · Editorial & legal review
Mapping the listing and buyer processes
I researched the transaction through agents and industry sources and used Gary Keller and Jason Abrams’ process charts to check the map. The full playbook covered 17 listing processes and 12 buyer processes.
I checked for missing work rather than treating the existing outline as complete. That review identified a missing listing-side repair-negotiation process, which I added to the map.
Testing an AI approach for each task
Surveys showed Gemini, Claude, and ChatGPT were among the platforms agents used most. I required a foundational approach for each process using those models and developed an accompanying pressure-tested prompt.
I also researched purpose-built products for advanced needs. I tested whether they could complete the task reliably and meet the workflow’s compliance needs, and considered actual usage, vendor maturity, and longevity when recommending them.
Writing instructions agents could follow
I translated the research into process cards with a goal, how-to guidance, a prompt, and a watch-for section. Specialized tools had setup instructions and one-pagers, so agents would not have to repeat the research just to begin.
I also created a separate 20-step introductory guide. It was a lighter entry point into the work, rather than a different count of the processes in the full playbook.
Refining the guidance through review and presentation
Editorial and legal review changed both the recommendations and the wording. I retained human responsibility for pricing, timing, and negotiation, revised steps to a hybrid approach where needed, and incorporated the required compliance language.
The completed work was presented, including Jason Abrams’ presentation of the introductory transaction content. I taught related workflows to national and international audiences and used the research in existing agent playbook updates.
Outcome
A transaction guide grounded in researched tasks
A research-led playbook covering 29 transaction processes, supported by tested prompts, specialized tool instructions, and a separate 20-step introduction. The materials were presented and shared as practical guidance for agent work.
Let’s connect
What could work better?
I’m interested in AI innovation and solutions roles where research becomes useful systems.