Makoto Suwamoto

AI Workflows & Experiments

No-Code Payroll Reconciliation for a Major Sports Event: Used Claude with Google Apps Script and Google Sheets to reconcile ~25,000 timesheet records for 1,200 hourly staff across two incompatible systems at a global sports broadcaster. Built scripts to merge data, remove outdated entries, and recalculate pay, then added formula-based checks to flag improbable hours and missing logs for managers to review. Helped ensure on-time, legally compliant payouts with no unresolved payment issues and contributed to bringing total payouts in under budget.

Context Management (“Master Prompt”) for Internal LLM Workflows: Maintain and iterate a ~30-page “Master Prompt” that gives our Enterprise Google Workspace Gemini Pro structured context on org structure, roles, approval paths, and recurring workflows across corporate and strategy domains. This allows colleagues to generate first-draft emails, briefs, checklists, and decks that usually need only light edits instead of being written from scratch. Periodically prune and refactor the prompt to stay within context limits and keep procedures current.

AI-Assisted Matching for a 60-Person Alumni Event: Designed the survey and matching logic for a 60-person university alumni networking event. Collected attendees’ goals, topics they enjoy sharing vs. learning, conversation style, and career stage via Google Forms, then used ChatGPT’s o3 reasoning model to tag and score responses. Assigned participants to two rounds of 4-person tables that balanced shared interests with complementary expertise and avoided repeat pairings, creating an optional “AI-matched” track alongside open mingling.

Regular Testing of Current LLMs with Real-World Experiments: Regularly test ideas from practitioners like Nate Jones, Ethan Mollick, Allie Miller, and Conor Grennan by running small, real-world experiments to try first hand the different capabilities of different models. Currently focusing on AI Ops (looking into Rachel Woods’ AI Exchange AI Ops bootcamp) and exploring how LLM-based adaptive assessments (like AI Cred) can incentivize deliberate practice at scale in domains previously not possible, and thus increasing fluency across various domains of human expertise.

Makoto Suwamoto is a Corporate Operations Specialist based in Tokyo with a rare combination: the strategic judgment to know when AI shouldn't touch a problem and the technical sophistication to build serious solutions when it should.

They've architected a 30-page "Master Prompt" system that functions as organizational memory, built custom Apps Script automation for complex payroll reconciliation without writing traditional code, and developed verification techniques—including adversarial "canary traps"—that catch AI failures before they compound. Their approach to context management and model limitations reflects genuine understanding of how these systems actually work.

What makes Makoto's profile worth exploring: they're quantifying 90%+ time reductions while most practitioners are still figuring out basic prompts.