Rene Pajta - The Enterprise Version of the One Person Unicorn

0:02 Good morning everybody. this session will be slightly different because we talked a lot about what we building how we improving all those technologies and I'm focused more about how these technologies are adapted in the organizations how they are actually used with v engineering we see a lot of things being created but struggling to get the adequate usage and and live through in the organizations. my I'm I'm basically a chief architect at at Microsoft advising the the CTOs and CIOS of Fortune 500 organizations. I focus on their AI transformations and both how to build with AI but also how to build agents which are worth to purchase and create the new business models for them.

0:47 And also at the same time I running as a technical lead our own sales or customer success account transformation where where we have around 700 people and we building our own agentic systems to help us to run accounts better and uplift the the quality of how we run our own business and I would like to start maybe with a a concept called enterprise unicorn right we heard about like one person unicorn needs to be and this a bit looking at what it takes maybe how that could look like in the larger organizations which has over 10,000 people and what it will be as a shape and form in there and I believe a lot of that going towards rather than talking about the headcount is going about how we going to operate in let's say a few years and I call it either a mission control or a command center basically having a human an age and a shared state where we have a pre-built signals based on which we can act in the in the organizations even before we ask a questions. The chats are more things of dealing with exceptions and trying to capture those as a new signals which we have to make the decisions in the in the organizations.

2:07 and it really like I would say a simple test of this will be that how many meetings we will need to have to sync internally as an organizations in ideal state we won't need those meetings because everybody's on the same page as we capture that intelligence within the organization and as usual it all starts with with the waves right we started with chats with assistants now we kind of have the micromanage management of agents where we run the task they run for a while but then they still need our steering or we don't have developed the trust into them and we want to move to the delegation and remote execution what we see with promise of the open claw automated agents running parts of our businesses and I still would say that majority or a lot of the enterprises large organizations are in that micromanage state because of couple of things that we just have a the system and access to the systems which needs to be connected to really done the work. For example, I'm a salesperson. I need to connect to like three to four different systems and I need to have access to those combine that information, right? To bring and build that one mission control system.

3:20 But that's not enough, right? I need an intelligence how we as an organizations are using that and that usually the vendors cannot build that needs to be built by the organization internally and therefore there are still I would say in the micromanagement state because we have those connectors but the that company intelligence still needs to be developed on in the organization. So I would I would say like moving to the delegation piece will really require that vider system access to have all that information to bring it in one place plus the dusted knowledge of the organization to develop the trust on top of those workflows and because of how we see the how the organiza or let's say the technology evolves we see new models every 3 months we see new frameworks harnesses to being positioned the CIS and CTOs I have really tough questions there like how I find the clear direction, how I bet on what technologies to evolve and upskill my organization to survive such waves of different technologies. Imagine like if you have a 10 people in your organization, it's pretty much a daily business to also learn and adjust.

4:35 However, if you have an organization of 10,000 people, it requires a lot of coordination, efforts, communication, enablement, strategy and if there is no clear direction to that usually what happens is either people wait or they start building on their own which then creates one part is just stailed organization the other one is basically creating a shadow IT with engineering that's even booming and then it's very difficult to create a governance right Ideally the companies wants to operate like a small city with clear laws but also freedom to experiment within those boundaries.

5:14 and that's also where the efforts then then going really in the in terms of standardization right we need to have standardization in terms of the governance in terms of the of the policies we do have to enable those laws to to to be around and then allow those on the edge actually to build and innovate because there is that tested knowledge captured in the organization we don't know centrally let's say in the organization what's happening what are the nuance is how the organization operates and we need to have both of those. So a lot of questions come to what I buy and what I build. Where is the mode and where is a commodity in that space and I I would say what we a lot see like the the models the agent harness the connectors that's all a commodity which you ideally focus on on buying but then the really the the mode for the organization is their company intelligence. how I capture that intelligence of the my organization and an example from from our side is I'm running a report we have a reporting system which has around 400 reports right I have a connector to it but that connector is just giving me access to go and query those reports however what I need to do is really to understand how my expert teams are using those reports when they use what and that's basically what we also do as One layer is just the connector itself. The other one is going and crawling those system creating kind of navigation like mental model capturing how my employees are going and and retrieving that information, right? Because that allows me to actually have agents which I can trust, have agents which are making less mistakes, they are not kind of this junior intern employees. They are experts and they're even better than I am in doing that job. and that also then with with the latency and the cost actually reducing because they make less mistakes on achieving that specific task. and and that's that's basically like where we saying also this company intelligence is something durable.

7:25 it's where you want to capture those nuances, right? You can have a let's say global sales book that can be shared globally, but then the nuance is how we handle let's say our accounts in Germany or how the others handle accounts let's say in Japan. It's very different nuances which you can't globalize and generalize and needs to be captured on that edge level. And once you know like what you want to build and what do you need to let's say buy it also comes with the kind of the operating model meaning that you have that central spine where you're building let's say the key connectors to the systems right you don't want to have your field to go and do the authentication and authorizations integrations to to the the key critical systems you want to have that done centrally. So you provide that access to the data for for the teams but then also on the other side you for the edge on on the on the edge for for your employees allow to let's say build around it. they focus more on how to capture the the knowledge of of that dayto-day task and operations because if we have just everything centrally we create a go well governed systems but which which are missing work because they are not doing exactly what we on the edge expect and on the other side if we just let everybody do the innovations on the field side we're creating and booming the shadow shadow IT which is very difficult to govern and with that I would say coming back again to let's say that one person unicorn it won't be let's say one person doing the work it will be everybody having kind of being a CEO in a boardroom having their own set of agents and workbenches which allow them to take the decisions on daily basis imagine I think our work will shift to the point that we have a queue of let's say your colleagues behind you asking questions on different topics and need to you make the decisions. So it will come also with a new type of mental loads on us. We are not used to do decision- making on constant basis. We are very difficult with content switching right. So these things will be creating a challenges for us mentally as well how to adjust at the work and also the trust will be very important because we can't if we can't trust our delegates who do the work for us we won't be able to make those decisions and those two things will be shifting and at least what I see across the enterprises so if I should leave with one it would be buy the models buy the commodity and build that intelligence in your organizations.

10:18 Thank you.