
How lawyers can develop their prompting skills by process thinking
As part of the adoption program, I’ve been giving training to law firms for several months now and the main objective I have applied to myself is writing the perfect prompt. Every training we dive into technology, combining the new developments with the legal sector. Saga helps legal professionals to eliminate time-consuming, repetitive tasks, allowing them to focus on what really matters. As we call it, the human interaction is what really makes the difference. The Saga-platform comes with some advanced prompting capabilities. The opportunities are endless once the lawyer understands how a prompt impacts the outcome of LLMs. Combining legal expertise with a high level of prompting will, without doubt, outperform the competition. How can we reach this superhuman level in prompting?
Embracing process thinking
Lawyers sometimes struggle with process thinking and in general consider every legal case as unique. The idea of standardizing legal work will feel counterintuitive when treating each case as something special (also, the billable hour method doesn’t work here). Therefore, when observing a typical law firm steering by the traditional hourly billing-model, they find process-oriented thinking unnecessary, as it aims to minimize effort and automate repetitive tasks.
But what if…
Here comes a what if-question. What if a lawyer develops skills in process thinking and as a result delivers top quality work in less time than their competition. During the adoption program, I notice lawyers light up when we find some useful prompts they can use for their own work. Writing a good prompt that produces the desired output can be challenging. It’s not just about phrasing the prompt correctly but also about identifying which tasks or processes AI can handle most effectively.
Difference between processes, AI use cases, and prompts
The other day I invited an insolvency and restructuring lawyer over to our office and explained to him the objective of today: writing the best prompts by thinking through processes of his daily work. Hmm… Not sure whether he liked the job or not. First, I’ve showed him some examples of what the Saga AI-platform was capable of and gave him some instructions on how to think in the correct way.

Image 1: processes, AI use cases and prompts
Step 1:
Try to write out all the processes you bump into when dealing, for example, with a restructuring and insolvency procedure. We came up with like nine processes.
Step 2:
Which of these processes are best suited for AI? Is it one that involves large volumes of text? Do you need AI to understand the content and facilitate discussions by evaluating arguments and assessing the validity of certain considerations? AI is very good in analyzing and reviewing extensive documents or contracts. Saga also offers a role that reasons as a counterpart in cases, challenging you on all your arguments. We call these processes that are best suited for AI to solve, AI Use Cases.
Step 3:
Formulate a prompt for each AI use case—easier said than done. Crafting the perfect prompt requires key elements such as context for the case, a clear objective to guide the AI’s reasoning, instructions on which documents to reference, and guidelines on structuring the desired output, and so on.
One step closer in finding the perfect prompt
We consider these steps iterative. Repeating these steps and finetuning the prompt each time, can deliver a great prompt that will outperform the competition. This is what I would call a perfect prompt. A prompt you should not only keep to yourself but share it with the rest of the firm. For some it’s easy to come up with perfect prompt while others struggle with creativity, process thinking or just finding time to do so. Embracing AI is something we cannot avoid easily anymore in this industry. Support your fellow colleagues’ journey embracing AI by leveraging them with the perfect prompt.