[Music] welcome back to the Enterprise agility podcast I'm John Lim and we've got Fred Champlain and James Ramsey back again to explain how strategic portfolio management and Enterprise architector are crucial in successfully implementing AI across organization at a high level SPM is used to prioritize and EA is utilized in looking at dependencies whereby we can pull that all together and ensure we are funding the right gen initiatives is that on the right track James no no absolutely so you think about um that that whole approach there's there's an element of right we're going to this AI transformation is is across the Enterprise so the key element is we need to understand what those use cases are going to be and we need to be able to prioritize them based on the value going to deliver to the business and also our ability to deliver them so that that's one element of it but one one of the interesting conversations I've had with organizations is the the greater that upfront piece in terms of well let's crowdsource what those AI use cases are going to be maybe they're doing it through their internal employees their St their uh internal stakeholders or sometimes actually reaching out and getting feedback on the products and services that they delivering from the customers and what the customer expect from an AI perspective so that sort of element uh I'm seeing organizations really getting a benefit from from leveraging SPM at the front end of the process but one of the other interesting aspects is understanding what the impact of AI technology is across the technology landscape and again that feeds into from an Enterprise architecture perspective and some organizations have been quite um guarded in in terms of the the process that they've taken to implement those use cases because they've been unsure about how they're going to track the uses of AI technology with inside their technology state so fundamental if you think about it Ai and the use of AI has got significant benefits he going to provide the organization in terms of value improve efficiency improve productivity you list goes on the N but on the flip side to that is potentially impact from a risk and compliance perspective so what organizations are saying to me is what we need to have is a clear picture of where we've got AI deployed inside the organization so from an Enterprise Architects perspective and using our our Enterprise architecture capabilities it's important we're able to capture that AI usage so if you take business applications as a as a as a a level in my estate I I may have procured some business applications that are commercial off the shelf and as part of their capabilities and the features they provide they're using AI as part of that application so there's there's a AI embedded within that set of solution you've then got business applications inside the estate which are providing AI type capabilities to other applications inside our estate maybe things we we've developed internally and then the third leg may be where we've got applications which are leveraging AI capabilities or external llms outside the walls of the organization so I you often use a a really simplistic use case to get this across but imagine you've got a uh a global payroll system inside your organization and you want to enhance that some AI capabilities and and somebody's had the great idea that we're going to actually link this out to chat GPT now one of the things about that is there you've got to understand well there there's some sort of digital integration going on between your internal P all system and external chat GPT so that's something that needs to be modeled in terms of your architecture model um but also then if you think about it you also then need to understand what that data flow is going to be so from a service now perspective we use we have this concept we call information object so as we model those digital integration we're also being able to model the the information objects that are passing back and forth between an application one application and another application now that becomes important in terms of visibility for in this AI use case because you think about it maybe this payroll system leveraging chap external chap DPT and maybe we modeling that we're passing some employee type information between the peoll system and chat GPT in order to get uh information back now that not that's probably not the great idea in the world sending out employee record information bank account details and that that sort of stuff so and James that that's really where were yeah really comes in and and also informing from a risk perspective right as we think about governance um in in Risk controls and compliance right um which is as you mentioned a lot of customers have concerns around this very legitimate especially when you go off off your premise um and that really introduces uh the need to start modeling those Integrations and being able to have an understanding of of where those dependencies are right absolutely yeah exactly where I was going with that it's that it's that risk perspective because what what you need to be able to do by by modeling it in your you know from from a diagramming perspective and modeling what those digital Integrations are understanding what the flow of those information objects the Enterprise Architects are now able then to share that information digitally with the the RIS teams so they can understand and identify the policies and controls that need to be applied to your peoll system in that context to understand what we need to comply with and from that we can calculate the risk profile associated with that application linking out to an ex external llm which ultimately is going to show high risk and going to prove to be a bad idea but the the the the key tenant is you have that ability to model the application where where and how it's leveraging Ai and also down to the the that you're looking to to pass between the two which I think from a resiliency perspective it's it's sort of critical to what we want to achieve absolutely absolutely well James this is a great conversation great discussion around some of the the real life things we're seeing with our customers as they um consider how they Implement Genai um all the all the the boundaries the the the issues uh from a application portfolio management perspective from a Enterprise architecture modeling perspective and digital Integrations so thank you very much James this has been a a great conversation around what's really what what's happening with our customers so appreciate your time and and thanks for joining no problem Fred [Music]