How do you think AI could be your secret weapon in handling risks?

Sorry for intervene. The key for using generative AI (ChatGPT for example) is to design the prompts (answers) in the right way. If you create a prompt in the right way generative AI can say you the sentiment (that is the word it is used. I do not like it) expressed in the text you add to the prompt or when generative AI analyze documents.. For example, this is widely used to analize customer feedback in service desk answers. Other example: we use it to calculate NPS (net promoter score) from surveys. For an AI entity (software or whatever) this is the most simple thing to do because is a matter of clasification. Remember: all in AI is probabilistic based on the knowledge base and to inference the best way to add existing text to the result, then what most companies forget is, when you use things like generative AI, almost a new business unit has to be created to work on results quality, including it lawyers, diversity specialist, etc. In the case of working with risks as Kerry wrote just to have people that validate the answers in terms of you are looking for is enough.

Assuming we have a large amount of data, for example, lessons learned database from past projects, it would help, but getting enough data to identify risks is a challenging task, in my opinion.

Thank you Sergio. We have the tools, but I was more referring the way of gathering the data to find out the current project sentiment and possible hidden risks. As you referred, we already analyze the sentiment of customer feedbacks. Now, how do do that for our internal project team. One way could be to collect feedback similarly as we do with customers, but getting answers can be challenging. To really dig the silent risks, we would need to access the raw data like meeting transcripts or Teams channel discussions.

Yes, I agree that the lessons learned from the past projects are important data to gain more insight to possible risks that you might face and how to potentially mitigate those. Instead, this approach don't identify the specifc risks that there in the current project. This is the aspect that I'm interested in and working with.

For us, "client/customer is the next on the process chain" so we collect feedback from all customers in the same way. In the other side, for example, you can use meetings recording (video and transcripts) to collect feedback BUT IF AND ONLY IF you put in place policies and rules about that and each time you will start the recording you notice to everybody about that. You can collect feedback from emails and from comments in project site (yammer, ms team, whatever you use). There are lot of way to collect feedback. The generative AI architecture published by Google in 2015 has democratized the information and is just a matter to imagine it. BUT because of that I wrote in my first comment that one thing is mostly forgotten is that when organizations start using generative AI almost a new business unit has to be created (I am using almost for not being category) where you have to put there not only technical people but other roles like lawyers, diversity and inclusion specialist, linguistic, etc, etc. The cost is higher not because the tools, the tools are for free, because the associated structured you need to create not matter you use it internally.