VIDEO TRANSCRIPT
The way that I recommend you work if you run internally, or, and which is the same as we work when we're engaging with is to help them develop a roadmap for AI is we start and we say, well, where are you at the moment? Let's just, you know, have have we got policies and the basics in place? Do we understand what's happening with our data. Now let's do some ideation. Let's understand the art of the possible and the desirable. What do we want to do? What can we think of that with the, you know, where where do we see that AI could move the needle, whether that's top line or bottom line or ability?
Take those ideas and look at the feasibility of them. Can the data and the, and the data structures and the organizational technologies support provision of that data in a timely and accurate manner?
What is the organization we need to build to be able to handle this as a production system? What capability do we need? What do we need to learn about AI? What practices and processes do we need to take on board, and how will we do this in a fair, equitable, secure and robust manner that is going to succeed as we would drive this into production and make our stakeholders happy.
The output from from my side for this is ultimately a a set of a set of business cases which she tend to do in the business canvas model summary format, and, a stack ranked roadmap of, ideas of, for progress, which can be built into into the IT and, technology road map for an organization starting with the lowest hanging fruit. Once the best ROI that can be delivered for the least amount of effort with the least amount of change, in order to start yourself down a road that doesn't involve months and months and effort in order to do the something that has value.
Another key's, tip I will give you, if and when you move down this road of trying to bind AI models into your internal or external platforms is exactly to look for the quicker wins first.
It's really, really important to to to deliver something that is successful and that that delivers results within a reasonable time frame because excitement initially will be overhyped about what AI can do for you. Your CEOs the board, the rest of the C suite, most of the employees, they're gonna believe that you're gonna change your world through AI And if you take six to nine months, just doing data engineering to come out with something that is actually then just a one tow in the water, you will lose a lot of that energy. You might lose a lot of that belief, and the AI transformation, will will will slow down and pause, and I've seeing that happen in organizations where the energy gets lost because they take too long. Expectations are too high and the delivery then is not just doesn't meet those long awaited expectations.
My view is get a road map in place set expectations going forward and deliver a series of small investigative projects but ones that have some value to the business in order to give you a platform for understanding what your budget for your next cycle will be, and how you can undertake larger projects that may require more data engineering and data preparation to be able to get the value from them.