About Ashley Cohen
Twenty years building software. Now teaching the people who use it.
I'm a Software Engineering Director at Trimble, where I've spent the last several years training entire departments on AI adoption. I started working with law firms after two attorneys — from opposing sides of my own family law case — independently asked who had built my case materials.
Trimble
Where the methodology was built and proven.
At Trimble — a publicly traded technology company — I serve as a Software Engineering Director. Over the past several years I have led AI adoption across four departments: product, engineering, support, and leadership. Each rollout was built from scratch. Each team had its own processes, its own document types, and its own resistance to change.
The first step was always the same: audit the workflow before touching any tools. Map where time goes. Identify the tasks that are repetitive, document-heavy, or precision-dependent. Only then introduce AI — not as a general capability, but as a specific solution to a specific problem the team already knows it has.
The result: 40+ professionals trained, across teams ranging from individual contributors to senior directors, reporting complex tasks completed in half the time within the first month. The engagement continues with biweekly check-ins because adoption doesn't happen in a single session — it happens when someone catches a drop-off early and corrects it.
The legal practice
How the methodology met family law.
I had a complex family law case of my own. Using the same workflow-first approach I had developed at Trimble, I applied AI to my own legal documents — organizing years of records, financial evidence, and correspondence into 66 structured exhibits, a complete case timeline, and full hearing preparation materials.
Two practicing attorneys — from opposing firms in an unrelated matter — saw the work product and independently reached out asking who had built it. Neither was pitched. Both became clients. The Sprint is the productized version of what I built for those first two engagements.
Methodology
Why it works where other AI training doesn't.
Workflow-first, always. Most AI training starts with tools. I start with the work. Before any software is introduced, we map your actual processes — document types, volume, where time goes, where errors happen. The AI gets inserted into that map, not presented alongside it.
Role-mapped adoption. A litigator's needs are not a transactional attorney's needs. A managing partner's needs are not a paralegal's. Every engagement is built for the specific roles involved — the prompts, the workflows, and the training examples are all drawn from their actual work, not a template.
Ongoing monitoring. The most common failure mode in AI adoption is not resistance — it's attrition. Teams are genuinely excited in the first week, then slowly drift back to old habits as complexity and edge cases mount. Every engagement includes biweekly check-ins to measure usage and intervene before the drift becomes permanent.
Credentials
The short version.
- Software Engineering Director, Trimble (publicly traded technology company)
- Twenty years in software engineering and technical leadership
- 40+ professionals trained across product, engineering, support, and leadership
- Four department-wide AI rollouts, each custom-mapped to existing workflows
- Legal methodology built during an active family law matter — 66 exhibits, structured binders, hearing preparation
- Inbound demand from practicing attorneys who saw the work product before any pitch existed
- Hands-on practitioner with Claude, GPT-4, Cursor, and NotebookLM
Working with me starts with a discovery call.
Tell me a little about your practice. I'll tell you whether the Sprint, a workshop, or coaching is the right fit — and what the engagement would look like.