AI tools are easy to buy and hard to use well. A subscription takes minutes to set up; changing how a firm works takes months. Professional services firms also have more at stake than most, because they hold confidential client information and are trusted to get things right. Five questions help separate the tools worth adopting from the ones that only add cost.
1. Which task, whose time, how much?
Start with a specific task, such as drafting routine correspondence, summarising long documents or preparing first versions of reports, and estimate how much time it takes today. If you cannot name the task and the time, you are buying a tool in search of a problem.
2. What information will go into it, and where does it go?
Find out what the supplier does with what you enter: where it is stored, how long it is kept and whether it is used to train the supplier’s models. Decide which kinds of client information may never be entered. Your professional body or regulator may have guidance, and your own adviser should review the policy.
3. Who checks the output, and how?
AI output can be fluent and wrong. Name the person responsible for checking anything that reaches a client, and decide what checking means for each type of task. The firm, not the tool, is accountable for the result.
4. Is there a clear process for it to fit into?
A tool applied to a muddled process produces a faster muddle. If the way the work is done today is unclear, fix that first, or at least write it down, before automating any part of it.
5. How will you know in 90 days whether it worked?
Set a small pilot with an owner, a time limit and a measure, such as hours saved on a named task or fewer rounds of review. At the end, keep the tool, change how it is used, or drop it. All three are good outcomes if the decision is based on evidence.
These questions are the starting point for Hexarch’s AI consulting. If your firm is weighing a tool, or if staff are already using one without a policy, a first conversation can help.