3 Key Considerations for EHS Leaders When Adopting AI Tools
Transcript
Hi, everyone. I'm Phil Benson, VP of product here at Blackline Safety. Thanks for joining me for this session. We're gonna talk a little bit today about, how to apply AI or at least the beginning phases of how we recommend you apply AI to your data and to health and safety projects kind of as as a specific topic.
So, there are three ways that we really wanna go about tackling this. And so from from our learning here at Blackline, we've been spending many years working with data and probably last last three where we've been pretty serious about doing data science projects and ML and AI projects. And the advice that I give people whenever they're starting this, because this is really us in our journey, is to, you know, don't don't worry about AI yet.
Don't start your project with thinking about that. You wanna kinda think about three things. You're gonna your first one is gonna be consider what your problems are. You wanna be really clear about what the problems are that you are trying to solve in your business, and and and that's actually the most important way to move forward. Because if you're just trying to apply the tool to some topic, it's gonna be really hard to understand what your outcomes are gonna be, it's gonna be understanding how to how to test that against real world scenarios. So you wanna start with your problems because those ones have clear objectives.
The next is I sort of call like know your data, and and that is a familiarization with the data you already have within your control and that you already understand. And you really wanna make an inventory of that. A lot of people get very hung up and very worried at the beginning stages of projects because they know they don't have quite enough data and they know that data isn't quite as clean as they would need it to be. And those are both very real concerns and will sort of decide the outcome of what type of data project you have, but don't let that stop you from understanding what you have and figuring out how to apply it to those projects. So you start with what problem am trying to solve? Then you really think about what data do I have on hand?
And then the third one is just understanding AI as a topic, understanding the tools available to you. And it could be, you know, as, you know, in-depth as fine tuning your own model, like using a foundation model and fine tuning that for an application that you have. Or it could just be really using a a gen AI model to sort of, help you understand more about your analytics, right, so you can very quickly be able to dive into your own analytics that are available using natural language and sort of get those visualizations. And it could be building your own model.
It could be building an ML model, you know, through more traditional data science techniques to just do some predictive modeling to understand the data you have. And as you learn about this as a generalized topic, as you understand the difference between AI and ML and Gen AI and what a large language model is and what it can do for you, as you're learning about that as a topic, you'll be thinking back on those last two pieces, which is what data do I have and what problem am trying to solve?
And that's really what we find is our best projects are the ones that really coalesce into, you know, how do we understand this as a topic, what do we have available, and what problems can we solve for our customers. And think it's very true for the customers we work with when they have their own AI projects.
So thanks so much for joining me in this first session. We're gonna have a few other sessions where we get into it a little bit, a little bit deeper. So thanks a lot. Have a good day.