Key takeaways
- Generic prompt training produces generic use. People go back to email.
- Train by role: intake, drafting, research, follow up, administration.
- Adoption is present when the work still happens if the enthusiast is on vacation.
Why “we showed them ChatGPT” fails
People remember a clever demo. They do not change Friday afternoon because of it. In a professional services firm, the useful question is not “can you write a better prompt?” It is “which part of your week should no longer wait on you?”
Training that stays at the tool level creates a few power users and a quiet majority who tried it once. Training that stays at the workflow level gives people a job to do with the tool attached.
What practical training looks like
- One role at a time. Intake staff, associates, marketing, and operations do different work. Do not train them as one room.
- The firm’s examples. Use your templates, your tone, and your actual last proposal — not a generic sample.
- A written exception path. What the person must still review. What they must never send.
- A before-and-after number. Time to first draft, time to first client reply, or documents completed per week.
Responsible use is part of the training
Employees need to know what not to paste into a public tool, what stays inside the firm’s environment, and who approves client-facing language. That is not a legal seminar. It is a short, firm-specific rule sheet attached to each workflow.
TechStack includes this training in AI for professional services so the systems and the people arrive together. A tool without training becomes shelfware. Training without a system becomes a story people tell about a Tuesday in August.
Put AI to work in the firm
See how TechStack helps professional services firms use AI for client response, knowledge, marketing and sales, documents, administration, and team training.
Explore AI for professional services