Simone Cicero 

Hello everybody and welcome back to the Boundaryless Conversations Podcast. On this podcast, we explore the future of business models, organizations, markets, and society in our rapidly changing world. Today I’m flying solo, but I am with a friend and I would say somebody with whom I’ve been exchanging a lot of ideas recently.

 

Andrea Gioia. Andrea is a partner and CTO of Quantyca, which is a company specializing in data, knowledge and change management, and is also the author of the book “Managing Data as a Product: Design and build data-product-centered socio-technical architectures” Andrea, again, has been a partner of many, many exchanges in the last few weeks. And so thank you so much, Andrea, for joining us on the podcast.

 

Andrea Gioia 

Charles, thank you very much for having me. My pleasure.

 

Simone Cicero 

Thank you so much. So you focus a lot of your work on how organizations can build useful information architectures, abstractions, semantics that are functional to their operations and in turn, let’s say architectures of information that can change how the organization operates.

 

And we had many changes, as I said, in the last few weeks and months because of our shared interest on how the question of semantics and ontology is becoming more relevant as, of course, we see AI entering the space of organizational development. 

 

And just a few days ago, three days ago, if I’m not wrong, we have witnessed the release of a seminal piece of work, would say, Jack Dorsey, former CEO of Twitter and now CEO of Block has released this very important blog post article on Sequoia’s blog, which is dubbed, I’m not wrong, “From heirarchy to Intelligence”, something like that. And it’s been very heavily discussed. And we agreed that it was a good idea to start from there.

 

Because on my side, felt like by reading the blog, the post, I felt like it was very relevant and it was very aligned with all of the conversations we were having. And the major piece, let’s say, of this article is Dorsey envisioning a future of blog as essentially a modular organization based on the principles of the platform organization. He even mentions Rendanheyi, which was quite interesting for me. He mentioned directly the idea of platform organizations that we have been pioneering for so long. he speaks about actually in visions on organization where the data and regarding the changes in the ecosystem, the partners, the usage of the products, the customers actually influence and somehow drive how the organization reshapes and rebundles to respond to that dynamics. 

 

He also mentions a couple of important topics that I think we want to talk about. This idea that besides the organizational modularity that is generally, let’s say, discussed as an enabler of this, he speaks about something that equals the company world model and the customer world model, the client world model, something like that. 

 

So it’s like he’s speaking about a representation of how the company works internally, its products, its units, its themes, its services, and on the other side, a representation of the ecosystem, the customer, the needs, and so on. So I think this is very much related to your work, and I know that you have been thinking about this for some time. So what was your feeling about Jack Dorsey’s blog and how does it resonate with your work?

 

Andrea Gioia

It resonates a lot. Just to begin with the starting part of the blog post, maybe for us is an obvious starting point, but it’s not so popular nowadays. is the context from which all the blog post is developed is the fact that the modern way we organize the work is unfit for the world we live in designed in the industrial age to work in a complicated world. Now we are in a complex world. So the change is not the exception. Change is the normal. And so the Taylorist organizational model is unfit for organization in order to drive and adapt in this context. The second very important point at the beginning of the post is the fact that thanks to AI, execution is becoming more and more easy. But when execution becomes easier, the bottleneck in the organization becomes the coordination. So how the different parts of the organization coordinate in order to reach the common purpose. Starting from these two points that resonate a lot with you. 

 

So we need to rethink the organizations because they are unfit to the world we live in. And the fact that AI is a tool that we can think of not just as a technology to adopt, to automatize existing tasks, but as a technology that can simplify the coordination problem within the organization. So this resonate a lot with me. And you have mentioned also the concept of modularization is a derivative of this thinking of reorganizing the way of working because when you cannot anymore in a complex world predict in advance what will be the state of the world two years from now and then optimize for that state of the world, what you have to invest on is the capability of adapting. So you cannot predict what will happen in the future. 

 

And so you must invest in being able to adapt very fast in the changing world, in a world that is constantly changing. Being able to adapt means modularizing the unbundled organization, modularizing it, and then being able to rebundle as the context changes. Unbundling and rebundling continues. The organization should be really, really agile. 

 

And now we arrive at the point that is discussed the most about the blog post, how we can coordinate the different part. As soon as I unbundle the organization, and unbundling the organization is something that is very complex. Maybe we can discuss later how unbundling the organization that we know is very complex task. Just suppose that we are able within the organization to reorganize the way of working in a way that is very easy to bundle and unbundle the different capabilities within the organization, it becomes very important to have a shared knowledge of the context within the organization. 

 

Because the organization at the end of the day is a unity, it’s a composition of different parts  that moving in order to reach a common purpose. But if I modularize and distribute the different capabilities and then bundle and then unbundle continuously, it’s important that everybody within the organization understand the context.

 

First, the identity of the organization. So why the organization exists in the first place? What are the value of the organization? How the organization plan to work and when one to move in the near future? And the other part is knowing the ecosystem in which the organization operates. Now, so the client work model and the inner work model of the organization. 

 

And this is basically the information architecture that I’m working on every day, so not just the data, but the data enriched with the context to understand the meaning of the data that the organization produce and consume. So basically, the conceptual model that describe what the data effectively contain, how to understand the data, and how to use effectively the data into the domain in which the organization works. 

 

What I do not agree completely, and maybe you as well, is the idea that this world model can be automatically generated by AI. I think that we have started in the data community, in the knowledge managed community to discuss this point at the beginning of the year when the concept of a contest graph has been popularized by a venture capital that they’re the contest graph as the next trillion dollar opportunity. 

 

And the idea of contest graph is to build this for the model directly using agent. So basically agents are using the weaving processes in the company to automatize the processes. But because they are on the decision path, so they see how things are decided and understand the reason why the organization take some decision because they are into the process, they are embedded into the process, they can learn and they can basically create these world model.

 

So you have not created it from scratch. You put the agent within the processes and this agent learn how the organization really operate in action, not through reading document and process on paper, but participating into the action, they can learn the world model and create this world model for the organization. 

 

I not totally agree with this idea. I think that in order to be able to put the agent within the processes, they should be reliable, otherwise people are not using an agent if it is a reliable, because they do not know the context and produce the results that are not so good with a lot of hallucinations. So the agent is not using. The person prefer to bypass the agent and do the work itself, because he does not trust the agent itself. 

 

But if the agent is not really used, within the process cannot learn how the process is executed in reality and build this world model. So I think that in order to have the agent that are reliable and use it, it can start to learn and adapt the world model. First, you need to build a minimal world model from scratch, find an agreement or key concept that define the identity of the company between the top business manager, the leadership team, or the person that basically have the role to define the identity of the company. 

 

So start to define these basic world model, then use the world model as the context to create this agent that participate in hybrid workflow together with person. And these agents are not only executor, but are also listener and learner. 

 

So they can, while executing things, they can learn and they can adapt the world model, of course, with the creation of human, human remain in the loop or as I prefer to say, human remain in the meaning. So there is a creation, but they can add a lot in the maintenance and the evolution, reducing the cost of adapting and evolving this world model. 

 

But I think that at the beginning, you have to create the world model from scratch. And this is not a simple thing because it’s not about how you represent the world model. It’s about finding an agreement of what is the identity of the company. 

 

Take the customer, for example. If we need to define what is the customer for the organization, probably there are different points of view. The definition itself of the customer is really political, move the power within the company. And so find an agreement between different points of view of a shared definition of the customer. What is valuable is not the definition itself, but the definition is shared. Different people agree on and from that, yeah, exactly. 

 

And as Brandeline say, find agreement is the most complex task within the company.

 

Simone Cicero

Yes, they’re most expensive as well. So that’s so many things, okay.

 

Simone Cicero

First, the first thing I know that as you were speaking, it looks like we spent the last 15, 20 years speaking about organizations need to be composable and distributed into modular capabilities and autonomous capabilities because they have to be adaptable to changing environments. We cannot predict.

 

We cannot understand properly because of the complexity. So we have to build this adaptability, right, inside the organization. So this idea of antifragile organizations, resilient organizations, and so on. Now, it looks like we have a technology that can somehow understand what’s going on, better than us.

 

At least can have a, I’d say, give, can, can a stab at the complexity of the world. And we kind of retreating and say, no, we don’t want you to recompose or rebundle of the organization. Even if you understand what’s going on, even if the organization is composable, we are scared of you doing, you know, wrong doing this in one way or another, right? Which is a bit of a knee jerk reaction for me of people saying, no, the AI cannot be managing anyone because it’s an AI, right? And they have been managed by bad managers for decades. 

 

And now they are scared about AI managing them because of what I don’t know, right? So that’s the first thing I was thinking about. So basically the idea here, we have composable organizations, platform organizations, at least some of our organizations are going in that direction. So why don’t we agree that why don’t we use AI to manage the organization?

 

Why we rejecting that idea. And I think that what we are seeing now, especially when you, for example, say we need to start from something we have agreed on. I think this is largely because we want to, at least we want to have the capability to observe what’s going on. 

 

More than actually steer, because we are also kind of OK that agents can update the ontologies and the formation architectures as we go, because we understand that our human capability to do context modeling is very poor. People do not like to sit in front of boards, spend time to agree. It’s expensive. It’s terrible. Everybody that has runs in a workshop knows that. I always have this experience of customers showing up and say, I have this 70,000 characters architectural document that I built so that the organization can agree on it, but that do not really engage on sitting in front of a board for two weeks and trying to figure out a shared understanding of what’s going on. 

 

So I think that largely we are asking, we want to have this because we want to observe, want to be, I have this kind of feeling that we are in control of what’s going on. Which is OK. I think it’s OK. It’s a good idea to be able to observe what’s going on. Still, I’m not 100 % in agreement that the AI is going to do a worse job than humans in managing organizations. I’m not convinced. 

 

Actually, I kind of convinced the other way around, that AI is going to do a better job of managing. But still, I agree that we have to observe. Because at the end of the day, we are the ones that can somehow exert meaning on what’s happening. Because we are embodied entities. We know what’s going on is going to have an impact that we want to, let’s say, control.

 

So that’s more or less where I’m sitting at the moment after your first round. Does it resonate with you? Do you have any feedback on that? 

 

Andrea Gioia

Yes, think that, yes, I agree. We need not to be scared about artificial intelligence. We need to know, we need to understand and think about what is the real role that it can play in a new way of working. Because I think that coordination and decision-making are two interconnected things, but are two different things. 

 

Because the way you coordinate influences the possibility that you have to design a way of working and your decision-making process. Now, correlated, but are not the same thing. I think that at the end of the day, at least in the near term, decision-making is something that remains upon humans. Instead, coordination should be moved as fast as possible on artificial intelligence. 

 

Because in this way, first, when I have to define the first agreement about our knowledge model, our meaning, our operator, it is our identity, we have not to o be scared and play a lot of work to define all exception or define a model that should survive for the next three years or for the next strategic iteration. It’s not important. It’s just to have a good enough agreement on key important points. That’s all, because then artificial intelligence can help you evolve it over time. 

 

And you can also have some different and not completely aligned definition if they are not completely orthogonal, but you can have slightly different definition, for example, of customer in a different domain of the, or subdomain of the organization because artificial intelligence is really good to do the contest mapping. 

 

So you have not to define a standard that is shared 100 % by everybody in the organization. You should have a common model, simple enough. And then you can have different specializations, what the domain-driven design called the ubiquitous language. And then it’s very easy. And you have not to do the contest mapping as described in domain driven design with interfaces or protocol or whatever. 

 

You have a different pattern that formalizes how you move from a ubiquitous language of one context to another context. Thanks to artificial intelligence, this work of translating, do the translation maybe between different definitions, slightly different definition can be done by artificial intelligence. 

 

This is what, for me, is the coordination capability – translating without having a super specific standard that everybody agrees on that defines everything. We do not need that anymore. We can have good enough model and the artificial intelligence can translate. Artificial intelligence can get tacit knowledge. The ones that is in teams, can formalize, can enrich the model, can materialize, can combine the explicit knowledge and then can move the knowledge exactly in the place where that knowledge is needed to take the decision. Is the artificial intelligence that takes the decision? No, it’s the person at the edge of the organization that has the need to take a decision, but the context is built automatically in order to make the better decision from artificial intelligence. 

 

And if it is not able to take the decision alone or needed to interact with other part of the organization, maybe part of that speak different languages, the artificial intelligence can do the translation and simplify that stuff. Not infinite meeting, but artificial intelligence can translate the meaning and be sure that for the work that we have to do together, we define the common language. My language is translated to your language and we define a common language to do the work without doing 100 meetings to agree upon what we are talking about. 

 

It can be the agent to say, pay attention, we are talking about the customer, but you two guys, I have a different representation of the customer. So when we are talking about the customer for this specific problem, what we intend exactly. So can do the sense making for you, do the translation for you, and point out what are the points of which you should agree for that specific task, to execute together that specific task. It’s a potent weapon for doing coordination and to redesign the organization thanks to this capability of organization of coordination that we do not have before. It’s very important because of course, we now return to an important point of the post. You can reduce a lot the middle management because coordination is basically aggregating, translating tacit knowledge and explicit knowledge, combining explicit knowledge, moving explicit knowledge to the one that need it, create new tacit knowledge and create this loop.

 

This can be performed by AI very well. The decision will remain on human and you need to have a lesser middle management because you can have someone that decides the identity and the direction and then have different capabilities or different capabilities grouped into bundles that then have all the context that they need to produce the result that are aligned with the purpose of the organization. So you can distribute the intelligence distributed decision-making process. But at the same time, thanks to this coordination capability provided by EIE, you can be sure that the organization move in the same direction. The organization remain an organization. It’s not distributed in a world of capability that are not anymore an organization. So the organization, the different part autonomously move in order to reach the same purpose.

 

Simone Cicero 

At the end of the day, we end with purpose, which is an interesting place to end for me because I was thinking some notes and.

 

The question I had, you speak about getting into some description, some context description, some context modeling. The question is, do we head? Where do we start? What is the starting context description we want? Because at the end of the day, where do we put this boundary where we agree with each other trying to understand what the company is about, what the organization is about?

 

What are the customers? What are we doing? What is the value? What are the capabilities? Because I can think of not putting this boundary. Let the agents figure it out. That’s one approach. And especially, think, if we embrace a perspective of let’s create an efficient organization. Let’s look into the perspective of Jack Dorsey. Jack Dorsey is saying, once you understand what customers want, the organization can be constructed by the agents. 

 

So the agents are going to reorganize the capabilities of the organization in a way that they can serve the customer. They can create excellent products. We just have to oversee what’s going on. So the question, so if you embrace this perspective, and let’s say that you have two organizations serving the same customers and optimizing for efficiency, these two organizations are becoming the same, which is a bit what’s going on with software. 

 

Software, what I mean with this? Let’s forget about this note on the software for a moment. 

 

So the question, let’s say, if we don’t put the boundary of what we want to observe, in service of efficiency, because if we just have to serve efficiency and effectiveness, we may just leave it to the agents. So the question is, where do we put the bugs? So what is it that we want to observe then? If we do not want to observe anything that is not related with efficiency, so what do we observe?

 

What do we observe? What do we optimize the organization for? What is the purpose of the organization? So at the end of the day, we end up with understanding what the hell we want the organization to produce in the world, right? 

 

So asking question about what are we optimizing for? What is this organization? What does this organization exist for? Which is a bit the same with software. That’s what the parallel I was making because at the end of the day, today, if you look at how we build software with agents, right?

 

The problem that keeps repeating is you are not clear on what the hell you want from this piece of software. So the agents are always refactoring things because you are not able to explain to them what you want from them. I was doing this software development project for Boundaryless and one of my partners told me, why are we not using a software company? And I said, because we don’t have clarity in the specifications yet.

 

And so he said, so why are we using agents if we are not clear with specifications yet? So the point is, what do we want from the organization? And at this point, it makes me think about that we have essentially a constraint on defining what we optimize for. Are we optimizing for sustainability? Are we optimizing for, certain qualities on the work we do? Are we optimizing for

a certain type of outcomes that we can generate or how we just say, no, we don’t do this or something like that. And this is extremely important also. 

 

And this brings me to the second part of the conversation I wanted to ask you. We are Italians, right? And I think people listening to the podcast will end up in experiencing it a bit of a mix between an episode of The Sopranos and one of the videos of the Sushi Mango. 

 

But the point is Italy is a very special place in the world, right? It’s a very huge economy. It’s a very manufacturing related economy. So he has a lot of companies that really are not just about software like in the Silicon Valley, right? They are building stuff like machines or pieces of engines or whatever, right? 

 

So the question, it becomes very interesting because the software development is kind of, it’s proceeding what’s going to happen with agents in the real world, in industries where there is a relevant footprint. 

 

So if we embrace this perspective – What can we say about the future of organizations? When agentic capabilities are going to be widespread at the information layer, and possibly also start to penetrate industrial production, automation, and so on. 

 

So what’s going to happen with organizations then?

 

Andrea Gioia

That’s an interesting question. Let me say that I see an intermediate future. The intermediate future is a scenario in which some companies will thrive, some company will become able to scale at a velocity that is not comparable with other companies, and they will win in their market. That’s for me a situation that is pretty clear. 

 

I mean, the ones that will be able to, first of all, unbundling. Because I said before, now organizations are monoliths. They’re monoliths designed to optimize specific value streams and all the capabilities, yes, are modular in a way, because we have a function, we have function, have capability, but are designed like puzzle pieces that can be composed only in one specific configuration.

 

And this the configuration that optimized the actual value stream. So they are not able to recompose easily different capabilities, different models in different configurations. And in this scenario, all the middle management, the third person, the third power that are within the company are basically the guardian of this fact that the bundling structure that we have within an organization are made by capability that are not easy to be unbundled because when you unbundled they lose their roles, no? Because they govern a bunch. 

 

Simone Cicero 

Yeah, they kind of defending, keeping it hard to unbundle.

 

Andrea Gioia 

Exactly. Maybe it’s not something that they are doing intentionally. Sometimes, yes, they are doing it intentionally. But of course, the organization pushed to do optimization, efficiency. And so they also tried to make this efficiency trading off the modularity, to put the piece more together in optimizing that specific configuration, not keeping them able to be unbundled in the next year.

 

Simone Cicero 

One specific configuration.

 

Andrea Gioia 

Yeah, so the company that will be able to modularize in the real way, creating modules, capabilities that can be reconfigured and are able to do this reconfiguration faster, are the ones that will be able to scale faster than the other. Because e now the company in a very complex market, they were designed to work in a complicated, not complex (fabio please add “market”), see a lot of weakness and fight this weakness trying to scale to increase dimension, merge and acquisition. The better scenario for a company is to become a monopoly. 

 

But it’s a sum of weakness. So because me and you are weak, we are not able to scale because the bottleneck of scaling is not in the dimension, it is in the organization of work because when execution became cheap.

 

It’s not a problem of the scaling, it’s not a dimensional problem, the scaling is an organizational problem – It’s a coordination problem. So the company that redesigned their way of working, doing better organization, so solving the bottleneck around the coordination, we scale very, very fast. No matter if the other companies are merging in a big worldwide company that are complex, slow to move, slow to adapt, are giant, that are going to die, basically. 

 

So this is the first scenario. Then there is your provocation. understand now if when we have redesigned this organization around this new model of work leveraging AI, maybe AI is also better than us not only to do coordination, but also to take the decision to define the boundary, to redefine even the identity of the company.

 

And I think that maybe this could be a scenario. I don’t know. But in that case, that company that is completely managed by AI is a company in which the value of the company, the market value of the company is equal to the cost of token they spend to run upon. 

 

Because you can replicate the company as much time as you want. As you do the company in your way, I do the company my way, but your way and my way, it’s not our way, it’s the way of the AI. So the AI will replicate the company with a marginal cost very easily. There is not a competitive advantage, because everybody have AI, everybody can create the company in that way with that level of efficiency, with that product, can replicate everything. 

 

So where would be the competitive advantage. In the short term, the competitive advantage is the capability to adapt, bundle and the bundling, scaling on the organizational model side, not scaling on the measure. In the longer term, for me, the competitive advantage will be, so if a company that survived because it’s able to adapt very fast, the competitive advantage will be where we can bet against the model. So when we can steer the model in a different way to produce a bigger value.

 

So understand this leverage point, the company that have not only the organizational model, the AI, the automation, the capability to reconfigure, but people that have a clear idea where we can bet against the model, where we can leverage our creativity to say, maybe this is theoretically the best way of doing this thing, but I bet against it. Probably doing these other things, know better humans than AI, and maybe we can do this thing a little bit different so we can steer.

 

So I see that maybe in this scenario, just to picture the image that I have mind, you know, curling, you know, this very strange Olympic game and all that. And basically, the e-mob in the future can launch the stone, but the company that are better and have a real competitive advantage are the one that influence the moving on the stone, steering it a little bit, moving the ice before the stone.

 

The big direction of the store maybe would be something that AI can define, but that little difference in the structure of the terrain, of the ice in front of the store is what make the competitive advantage. Because if a company is replicable at a very low marginal cost, have no moat, have not competitive advantage. You can create it, I can replicate it at the same price, that is the token price.

 

Simone Cicero

I mean, yeah, competitive advantage is, I think, big candidate to go bust with AI. That’s an automation. That’s pretty much, I think, something we can agree on. again, as I was listening to you, another interesting point, you said in the intermediate steps, the winners are going to be the ones that modularize, because using AI, not using AI, you can be more able to reconfigure and respond to the market signals. 

 

But this is, I think, a signal that we have seen already. It’s coming already. Then if I think about how to modularize, so modularize is an interesting concept. Something that we have seen is that modularization needs to go with a modularization of the business model. So it’s not just a modularized capability, but this capability actually needs to have a product, an offering, a P &L.

 

It needs to be able to deliver a promise of value to other nodes in a chain of promises. So you don’t just modularize capabilities. You modularize into microbusiness, microenterprise capabilities, microenterprise capabilities. So there needs to be a distribution in the organization of the capability to run a business, so to deliver a value proposition consistently with SLAs, predictably, and so on.

 

What happens again, as we have seen, 

 

And this is by the way, something that I think it’s a first signal that we get that is imminently critical for the concept of an organization. Because when your organization is made of multiple micro businesses, the idea of the organization doesn’t really count anymore because these micro businesses can do business across organizations. 

 

So that’s especially when there is a technology that can coordinate, integrate across organizations, agents.

 

Andrea Gioia 

And in particular, if I can stop you on this point, when the market is very complex and you want to create optionality, this module, a lot of this module, you want to outsource them. You want just to control and coordinate them. That is the source of your competitive advantage. But a lot of these modules that now are part of the company because they are composed in a monolith, to create optionality, you want to externalize. 

 

You do not want to have within the boundary of the company, you want to only have the control of this module. And so control how the bounding make value, but not having the control of a single component, not for all the capability, but for some of them.

 

Simone Cicero

Yeah, I think I’ve read something very interesting, and I will try to put it in the podcast note recently, which was a blog post on, which was making the point that, of course, vertical integration delivers certain qualities to the customer experience. 

 

But nowadays, whenever you vertically integrate, you have to really consider – two things. So basically, you can vertically integrate even things that you don’t really own. So you can use other things, especially if the other chains of the pieces of this chain are, let’s say, well-documented, easily contractualizable, predictable. So they are a good piece of business. They can be integrated in your vertical integration fairly easily. So you don’t really have to own them. 

 

Sometimes, if you depend on those like a it’s dead or live for that situation, you may want to build it on your own. But essentially, the question is, when you have a very strong capability technology, sorry, coordination technology in the market like AI, whatever is not a counterpoint for you, it’s a liability, essentially.

 

Andrea Gioia 

Exactly. Maybe you want to vertically integrate by buying external company because you want to remove that capability from the market. So you don’t want the data company can use that same capability that you are using because it’s a piece of your key values. I don’t know, for example, Amazon right now on the supply chain, basically I’m buying all the companies that compose the supply chain because from one side, they want to reduce the risk. 

 

And it’s very clear that the supply chain is the competitive advantage of Amazon from the retail part. And the other part, they want to share some knowledge, some important knowledge that are important for this capability to execute the supply chain as efficient as they are able to execute it to other companies that can also coordinate that capability if it is freely available on the market.

 

So they prefer to put it inside to hide and to not make available this capability to add that capability.

 

Simone Cicero

Yeah. If you think about software, that’s already very hard because software can be replicated fairly easily. But again, if I think about the real economy, that’s another story, right? Because there are things which have a real impact, a real, let’s say, footprint. And we have also an episode of the podcast with Sangeet Choudhary on the Sandwich Economics that speaks about this very, very eminently.

 

But that’s a good point also to close the reflection because especially, I mean, think software-based economies or knowledge economies are kind of skewed from one perspective. So my perception is that in that part of the economy, it’s going to be very difficult to maintain a competitive advantage. 

 

So if you are only into this type of economies, very hard, maybe network effects somehow can save you. I didn’t buy the idea that you can somehow commoditize very easily these big network effects like Airbnb or something like that, right? Because I’m not convinced yet on this. 

 

If we look into the real economy, I think what we can see is that the AI, it’s another step in the direction to push the power in the hands of the customer. Because again, if we come back to the idea that AI is going to build the perfect organization, or rather is going to compose the perfect set of capability to serve the customer, then the customer is at the control. 

 

So actually, at the end of the day, who is dictating the meaning is the customer. Because as a customer, can say, I only want to purchase a piece of shoes that are built with fully sustainable, recyclable materials. And then if on the market you find someone that can produce this fully sustainable, let’s imagine, for example, you can 3D print the shoes. And this is one capability. Another capability is shipping the shoes to me. And then there’s another capability that is creating the materials from fully recycled process, for example. 

 

Then an agent can combine the perfect shoes for me. And then what is going to count is what are the constraints that they put in the production capability that I’m bringing on the market? Is it sustainable? Is it regenerative? Is it, I don’t know, a huddle or something like that, right? I think that’s somehow describing what could be this direction we are going. I don’t know if maybe it makes sense for you or it’s just too much far-fetched.

 

Andrea Gioia 

No, no, no, I think that, but already today, when I talk about context world, I talk about the fact that in some market that are not the rule by network effect, okay, which may be sometimes difficult to understand if the product is the platform or the product is the user, the thing to be the user, but instead it’s the product. 

 

But in the other market that are not platform economy based on network effect. So I’m not governed Instagram or other social media or Airbnb. I need to be there and follow the rule and not govern it because it’s not easy to create another Airbnb, not from the point of view of the software, but of the network that is inside. So the value is the network, not the software. 

 

But this is an exception. The reality is that other companies have to face the fact that the market is not anymore a mass market, but is a market composed by a single person that have different needs and these needs change very, very fast. And so basically they are an economy of speed, of customization, of variety. They needed to be able to produce a very specialized product and change the product that they are offering to the market as soon as the demand drift and the demand drift very, fast. This is the problem that the company has to face.

 

And so I agree, in that market, the customer rules the market. Because if someone else is faster and can produce what I asked for with the right price, we win. And so you have basically to create organizations that can adapt faster, produce exactly what the customer needs, do variation, do experimentation in an economically efficient way, and maybe creating some sort of value proposition, that has to be a quality application to that, value proposition that also build loyalty and engagement with the brand. 

 

Because the other thing that we have not talked about, of course, there is the efficiency to satisfy the customer, but there is also the value of the brand, the trust that we work together, we have the same proposal. I share your values. This is also an important thing.

 

Simone Cicero 

Yeah, listeners can look into the podcast we had recently with Jasmine Bina. 

 

I think it was very much on these topics of know, brand resonance and meaning. I think that’s a good complement of the conversation we had today. So as we go towards the closure, as always, we ask our guests to share some of what we call break rams. So I know you have been doing your homework. So what are your breadcrumbs?

 

Andrea Gioia 

First, I’m happy that you have not asked me what I foresee for the future of consultancy, but maybe this is a topic for another.

 

Simone Cicero

For another episode. I think we had one with Louis David Benayer recently, which wasn’t very encouraging, I must say.

 

Andrea Gioia

Okay, so I’m not going to watch that. Just a joke, just a joke. So I I read a lot of books. I have two channels to gather information and knowledge. The first one is discussion. I like to discuss with people about problems that we are facing and how we are trying to solve this problem. I think this is very interesting.

 

And I also have a lot of conversation with chat GPT, Claude, whatsoever, because when I see an idea or when I have an idea, I like to stress test it and see if it’s really a new idea or if someone else have already think about it. And usually 90 % of the cases someone have already had this idea 100 years ago, maybe everyone had a Nobel Prize on that idea. So I like to find out this thing and go directly to the source, read the original material related to that, to that idea. 

 

Because in a lot of conversations, we are talking on topics that have been already very well formalized in the past and I prefer when it’s possible to discover it and go directly to the source.

 

Then I read a lot of book. I try to read a lot of book. Of course, I’m busy, but I try to save some time. And I have three types of interest. Consultancy, of course, data and knowledge management, and organizational design. So I have some suggestion in terms of book for each of these kind. First of all, for the ones that like data management and the future of technical architecture within the enterprise, I suggest the Dave McComb book on the data-centric revolution. think that another big transformation with organizations is that we will not have any more application at the center and then data integrated later. We will have data at the center and application will be by the code around and come and goes, but it’s not the real value. The real value is the data. Then we build the application over a data manager centrally. So this is not a good news for software as a services.

 

Simone Cicero 

For which you need semantics.

 

Andrea Gioia 

Yes, you need semantics. Yes, you have data and you have semantics and then application basically are UI, API to assess the data. And this is in fact, is the reason why the market evaluation of software as a service company is going down because we are moving in a data center from application-centric world to a data-centric world. And the book of Macomb is fantastic.

 

For the one that have still these I don’t know how it’s called, but like to work in consultancy. I suggest two of my favorite books, also about consultancy. One is Donald Schön,the reflective practitioner, basically thinks that the practice. It’s not more a practice that execute technical rationality. So there is not a theory that the practice execute as it is from the book, but every case is different. And so you have to apply what you know, but by thinking in action, experiment, change and adapt to the single case that you have.

 

Simone Cicero

Yeah, I think it’s been suggested a bunch of times already. I think two or three times, if I am not wrong.

 

Andrea Gioia 

I love it. love it. And I like so Edgar, Edgar Schein process consultancy. That is, this is very important for consulting. Consultant is not the one that gives solution. He’s the one that ask the right question and steer the customer to find its own solution. This is very important. We have not to feel that I go to a company to say, okay, this is the best organizational model for you.

 

I have to help them to find their best organization and model. That’s an important thing. And then about organization, I like very much the idea of Stafford Beer?, the father of cybernetics on a viable system model and fractal organization. It’s yet aligned. It’s an evolution of the hierarchical, but more in a cybernetic way. 

 

I like it a lot because it can be a step in the transition between a very distributed company. Maybe it’s not the final step, but it can be something incremental that can be accepted. So I suggest this book, The Fractal Organization. It’s very nice by Patrick Hoverstadt that explained the viable model of beers and how the organization can be fractalized and share the decision process. 

 

And one of the last book that I’ve read, maybe I think that

 

Andrea Gioia 

Sangeet has been also your host in the podcast. Yeah, he’s been three times on the podcast. I’m trying to get him the fourth time, but he’s traveling too much.

 

Andrea Gioia 

It’s very loyal. It’s very loyal host. It’s a recurrent Reshuffle. I think it’s interesting. Most of the topics that we have discussed about coordination and the reframe ideas not as a technology to adopt to automate, but as a way to do better coordination are here that are discussed in this book. We have very good examples from other industries at that period of time. 

 

So if you want a story to tell. To better explain this idea in less technical way at the cocktail party, you have a lot of examples.

 

Simone Cicero

Yes, yes, yes. think he did a good job in this in the last book. And I think the book is excellent in explaining the macro thesis behind that. Also on the VSM, I must say that there is an episode. you want, you can look into the library. There is an episode of the podcast with Mark Lambertz on the VSM. 

 

I agree VSM is a good – It’s actually a good frame to look at distributed organization to ensure that you don’t just look at a bunch of cells, but more as an organism. So if you still believe that having one organization makes sense, think a VSM is very much a framework you want to, the viable system model is a framework you want to engage with because it gives you an idea of what are the functions that these organizations as an organism has to offer. So like a strategy function, exactly.

 

Andrea Gioia

Yeah, a live organ from a machinery to optimize to a living organ to grow, to grow basically, not to fix. There’s nothing to fix. It’s something to evolve and grow. Yeah, I like it.

 

Simone Cicero 

Yeah, so good. Thank you so much for your time. Let me remind the people to read your book as well. 

 

Andrea Gioia 

Managing data as a product. Then there is a very long subtitle, but you can skip it. It’s managing data as a product.

 

Simone Cicero

Yeah, yes. Yeah, it’s a fresh book a couple of years ago, if I’m not wrong. Thank you so much. It’s been a pleasure to have you, a pleasure to have the conversations with you in the last few weeks. And I hope you also enjoyed the conversation.

 

Andrea Gioia

Likewise, yes, very much. Thank you so much.

 

Simone Cicero

Thank you so much. And for our listeners, of course, you can head to Boundaryless.io/references/podcast, and you will find these podcasts recording with all the transcripts and all the suggestions that Andrea has shared today. And of course, until we speak again, remember to think Boundaryless.