#6 | On Collective Intelligence | Interview with Gianni Giacomelli

Ciao a tutti! Welcome to Engines of Creation podcast. I'm your host, Cristian Mastoronato.

In this podcast, I bring together my knowledge in complex systems with my experience in managing technology innovation and new product development to explore how successful products, organizations, and ideas emerge. In this new episode, we're meeting a great new guest, Gianni Giacomelli.

Gian is a fellow member of the Exponential Look community and he is the Head of Design Innovation at the MIT Center for Collective Intelligence and we are going to dive with him in the fascinating field of Augmented Collective Intelligence. When in this podcast we talk about intelligence, we usually look at it as an emergent phenomenon of complex computational topologies. This is called emergentist explanation, and it is the closest to complex system theory.

But it is not the only possible way to explain intelligence. David Chalmers, for example, called it the hard problem of consciousness. And for example, if you're interested to alternative explanation, you might want to look into idealist approaches like the ones from Bernardo Kastrup or Donald Hoffman. I personally find Hoffman's approach particularly interesting for its similarities with Wolfram's computational approach in his physics projects.

As we said, though, we are going to stick to the emergentist view, which is particularly valuable for us since one of the theses of this podcast is that innovation is an emergent phenomenon that arises from complex human network topologies within a specific organization or inter-organization, of course. which creates a direct link with collective intelligence, even if collective intelligence can explain much more than innovation, given its ties with sociology and social sciences.

What we're going to discover today is what happens when you augment collective intelligence, and Gianni will give us plenty of practical examples. Starting from a definition of augmented collective intelligence, we will immediately dive into its implication innovation, underlining how innovation at operational level is as important, if not more, than at technology level.

We will also discuss the impact of remote working in organizations and how managing weak social network ties is the key to success. Continuing our conversation about organization, we'll understand how looking at their complex system dynamics can help us identify the key elements to design and transform them into superminds.

This will lead us to discuss about the combination of artificial and human collective intelligence and we'll see some interesting examples across merger and acquisitions, sales and R&D department. So what is the key to become an enabler of augmented collective intelligence? According to Gianni, and I couldn't agree more, it is to be an anti-disciplinarian, working across boundaries and becoming a T-shaped individual.

Now let's dive into this great conversation. So let's go. Welcome Gianni. Thanks a lot for... joining Engines of Creation. It's really a great pleasure to have you here, and thanks for finding the time. First of all, Just, I think, would be great if you could introduce yourself a bit to the podcast audience for the ones who don't know you. Sure. Chris, thanks for having me here today.

Delighted to be here and just a free-flowing discussion. I've been working in innovation for all of my career, which is by now more than 30 years, but clearly the last 10 years don't look like anything. like we've done in the past. I spend a good chunk of my time as an advisor in the space of AI augmented collective intelligence, and I do that.

for the company where I used to be Chief Innovation Officer for many years, a company called Genpact, a spin-off of GE, IT services, about 120,000 people. I also work at MIT, at the Center for Collective Intelligence, where I head. the design lab and i'm a head of innovation design lab where we design prototypes that use artificial intelligence to augment collective intelligence

And most of my time has been spent across these disciplines in trying to, quote unquote, make the world a better place. I mean, I know it's a little trite, but the idea is. across, some people call it impact, others use ESG frameworks, but the environmental component, the social component, which also includes

making our people be better equipped for the future. And obviously a lot of governance, type work, finance, compliance, et cetera. That's been my career for certainly the last 10, 12 years or so. I used to work for technology companies. SAP had been with BCG for some time in the dot-com boom, startups in that period.

And I actually started my career in consumer products, in marketing many, many years ago. That sounds an impressive and very broad career. And yeah, definitely when you talked about augmented collective intelligence. As you know, this podcast... focuses a lot on taking and observing things from the complexity viewpoint, complexity theory and complexity management. And of course, the point of view of adding emergent behaviors of complex systems, which is one of the key elements.

that we do want to look at. And with this in mind, of course, looking at how a collective system can create augmented intelligence or, in general, emergent properties of any sort seems very, very focused with that. the core mandate of this podcast. And therefore, yeah, I'd like to read a little bit from you and have a bit of a definition of what art.

augmented collective systems or augmented collective intelligence and help our listeners understand a bit more about that. Sure. I think I'll share some definitions, but in the end, I think. It's probably useful to work on use cases and try to get people to think about it through the lens of what you can do with it. I think definitionally, collective intelligence is the ability for groups.

to manifest intelligence that is superior to the individual parts again you know the old adage of some of the part being bigger than the individual parts so collective intelligence has been for the longest time my colleagues at MIT Tom Malone have defined it as the ability to generate level of intelligence, which is the ability to adapt to new circumstances in a deliberate way, in a way that seems intelligent, which means being able to do things that

lead to a better state for whatever the group is. Now, the augmentation component is an interesting piece. For quite some time, we had You clearly thought that AI would move that frontier in the first, I think the first real surge of AI application in the last seven years. has actually proven important but not as pervasive as the new one so the the ability for machine learning to actually be embedded into

processes and therefore organizational design and therefore large enterprise groups and ecosystems, et cetera, has kind of been a little more limited than what we expected. Certainly played out very strongly. marketing and social media and advertising and the creative professions, et cetera, but a little less so in enterprises. Clearly, you do have... obviously everything that touches technology, so take server farms, for example.

They use that extensively, but if you look at the back office of a bank now compared to seven years ago, it's changed, but not as much as one would say. And so here the question really is, how do we augment the collective intelligence of an entire back office of a bank, for example? I'm just using this as a very simple example of an augmentation of organizational processes that have been there for some time. And there's been a lot of organizational design consultants, et cetera.

you know the question really is how do you use the framework and the techniques of augmented collective intelligence to make that operation better, for example. And if you solve for that, you can solve for pretty much everything else, from education to healthcare to... supply chains and whatnot. So that is what I mean by augmented collective intelligence. I think that's actually probably the easiest way to frame it. Obviously, from my standpoint, there's also a hope.

that we can kind of redirect a little bit some of the attention that the concept of artificial general intelligence or AGI is having. Everybody's gone after that thing, I mean, these days.

you only hear about that, right? Where is this thing going to end? How far are we from artificial general intelligence? And I think one of the things that doesn't really get as much attention is the... concept of what is called ACI, then, augmented collective intelligence, whereby clearly you have a significant component of artificial intelligence.

neural networks, deep learning, everything that we've been playing with, but it's inserted into a context of, it's not just humans individually, right? I mean, I don't want to go back to the human in the loop thing, but really in the context of collectives and collectives can be networks of people and machines and those networks are built

in a deliberate way. And you have some of them, social media, our networks, they don't perform and work particularly well, but that's the concept, right? So the idea is, how do you embed increasingly intelligent machines into networks of humans and machines, not just augmenting one-to-one the person in a way that makes that collective, you know, MIT calls it supermind, stronger.

That's a little bit of a different articulation that also belies a little bit of an attempt to have a slightly different, very different in many respects, architecture. What are the processes? What is the technology? What is the role of the human, etc.? In an augmented collective intelligence, it's actually quite different compared to going down the path of artificial general intelligence.

This first introduction and all the example you gave us, they are very interesting because they actually probably uncover a lot of the ground of the things that I want to discuss with you today. And so it makes it easier for the rest of the conversation to unpick. and go deeper into some of the things you mentioned. So before going down into the augmented collective intelligence, which is a very, very interesting topic, I'd like to dive down first into the concept of

of collective intelligence within organization and how you can augment that. And I know you sort of explored quite a lot, some of these topics. And if you, I'd like to start from a couple of things that I've read from your. in your work. One thing I found interesting, you mentioned about talking about the back office of the bank, okay, is that we tend to think of innovation a lot related to new tech, but actually

innovation, the real innovation lies into improving the operational element of an organization. And that's really where you can tap into the collective intelligence. And I'd like to understand a little bit more your point of view on this front. And I think related to that, I know that we did some work into what you call the smart water cooler, the virtual water cooler, as an example or a nice way for you to...

uh boost innovation especially in remote organization so that would be it would be interesting to start from that front and on the operational and sort of back office side and our remote team can improve and definitely move into other topics i think that's a good frame yeah i i think it's important to start with operations i always use it not just because i know that world well but i used it because it's very

brass tack, very practical thing that makes people feel, especially leaders feel that this is not just social media type stuff. There's a lot of people keep on thinking about applications yeah everything related to marketing and maybe even sales clearly social media media is more straightforward and it's kind of true but that doesn't mean that these concepts cannot be used in the collective intelligences that we have built i mean just step back for a second the design of you know

we use we always talk about people process technology right the the design of those things that that go that goes into the operation so again you're underwriting alone right so you're doing a certain amount of risk management, you have parameters, you have people taking data from the client, suppose this is a mortgage.

You know, you get to collect the data and then, you know, there's an application that needs to spit out its own confidence scoring. And then you have a human and the rights at risk. All of that is the... tip of the iceberg of a larger collective intelligence which is in this case the bank the bank is a machine to manage risk that's the reason why it is in business that's the reason why it makes money that's the reason why it is regulated

What we've been doing for the last many years, many decades, and clearly more intelligently in the last four decades or so, we build organizational designs that help that decision-making process scale. And we do that with business processes. We do that with training people, hiring the right people, controlling the right people, a bunch of practices related to Lean Six Sigma, if you think about it, writ large.

and the underpinning technology. But all of that is a collective intelligence. I mean, we call it in a different way, but ultimately when you look at a workflow, a workflow is the equivalent of a collective intelligence of a dinosaur.

it still has an intelligence it still is something that picks up on many different ambient data points and tries to spit out a result that is appropriate for that organization now what we're saying is that you can do a lot more these days because you have the ability to build better workflows not just things that work

with fewer humans but really things that do better things i'll give you an example for many many years we've been having this discussion about the business model of insurers business model of insurers is in many respects a faulty one, meaning that they use the data to pay for things that went wrong. And so you cover the downside. side of the risk. But what if you could turn the business model of the insurer and the insurers would sell ways of avoiding things to go wrong in the first place?

Wouldn't there be plenty of people to buy that thing? And the reason why we can't do it is that technically it's not that easy. We wouldn't have the data, et cetera. And again, the organizational design hasn't been there, but it's possible.

The art of impossible has radically changed in the last few years. And so that's one of the reasons why I think we should really pay attention to this concept of operations, because in the end, they can create business models that we... we really never had the ability to to make flourish now When you look at one specific level of collective intelligence, you're looking at the people side of the story. That's the reason that you were talking about the virtual water coolers.

Let's double click on that. This is a slightly different level of the collective brain, if you will. So forget about the automated workflows and the technology that helps people do underwriting, etc. um i'll tell you a story that i think is an important one and you know this one hopefully will get people who are more of the managerial type less on the process design and technology type and more manager kind of people

fully understand what we mean by augmented collective intelligence. We know anecdotally that Having water cooler interactions is important. That's the reason why offices are very often built in a way where you have serendipitous encounters. You create space. I mean, if you look at the layout, it's made for... encounters that wouldn't happen otherwise you know canteen and watercolors and you know we used to have the photocopier machine and all that kind of stuff right

Now, we also know that during COVID, for example, we started working online and remote, and those serendipitous encounters actually kind of died out for most companies that had people remote. Companies actually still functioned. You still could do your underwriting of your risk. But what happens... uh is something that is aligned with the theory and there's been research in this space which is network theory what happens is in the end you end up with unfortunately

what is called weak ties. The weak ties in the network are literally, you and I are weak ties, right? We don't talk all that often, but we're kind of part of a community and we kind of talk at times. Weak ties are very important in the network diagram because those are the places where you get a lot of innovation signals, a lot of cultural signals. That's where you get very often as an individual.

leads for the new job, leads for a new business. So the strong ties have a different function. They run the data today, but the weak ties do your quote unquote horizon two and horizon three of innovation. Stuff that you don't do today, but you could be doing one day. And when those things die out, because you're not in the office anymore, you don't go to conferences anymore, et cetera, all you're left with is the strong ties. And that's super dangerous.

That's the reason why we, you know, my team worked on, we've been working on network analytics for a long time, also with MIT, my colleagues at MIT, and then, you know, since I've joined a little bit, the freighter. We try to create a massive database that continuously looks at the strength of the tie between people.

I mean, the company, as I said, 120,000 people, we start with about 45,000 people above a certain level of seniority. And then we calculated the pairs. So every two people have... a strength of tie between them and that that is recalculated periodically and basically what we said is we want to boost

every tie that is below a certain threshold, but also above a zero threshold. So you wouldn't want to do just random things between people who have nothing to do with each other, but you'd want to boost. ties that are not strong. And you do that by doing first this calculation of, you know, you use the data and you look at who are the weak ties and then what you do, you get a bot.

to look at two people's calendar and automatically find time between those two people, just 15 minutes, and this is not a real meeting. And then planting that into the calendar in an automated way. Obviously you need to. agree and accept upfront. But if you think about it, this is a virtual water cooler that comes to you as opposed to you going to the virtual water cooler. Thousands of people have used it. Obviously, I'm a big fan.

it does actually do something to the reinforcement of those weak ties. And I think the jury is still out in terms of broad adoption. Now people want to go back to the office. They're not going to go back fully to the office. The reality is that you're going to create as companies, you know, companies are going to create different network structures.

Just because now a bunch of people are going to be remote and they're going to go into the office maybe once a week, once a month, once a quarter, right? Some of that is going to be fine. Some of that is not going to be fine. And therefore, I do think that it's important to reinforce. and boost that part of the spectrum of the connectivity through digital means. That's the virtual watercolor example that has very, very simple practical implications for people.

I like a lot this example. To be fair, I'm pretty sure that I first came across your work looking into this virtual water cooler work. And that's where I started digging more into your work and I found out other things. I like the point you're making about looking into the network and the way a network in a company is set up and trying to reinforce, as an example, weak ties.

if we believe, of course, they're getting weaker and weaker with the virtual work. I like to get this approach one step farther even and thinking about how can we... But first of all, can we design organizations in a way that can more continuously and systematically? enhance this collective intelligence. I see this virtual water cooler more as a reactive solution of, we have a problem, remote working.

How do we fix it? Well, not that remote working is always a problem. Actually, it also created a lot of opportunities, but it was a specific problem, as you mentioned. But then how can we become more proactive into designing organization that can leverage?

which collective intelligence more systematically. Yeah. And I think we need to keep it practical because I think there's many, many layers of complexity. And obviously, you know, MIT and others have been working at this for quite some time. I suggest that people have a look at if they're interested. of what has been written uh i've written a lot about it um you know on there's a website with the url supermind.design there's quite a lot to be discovered there but

I think we need to take a little bit of a step back. This is organizational design, right? At some level, a bunch of people know what that is. I think people in senior leadership and in HR. do that routinely. It's just that typically it is done with boxes and lines, reporting lines, and actually also with workflows. That's the way we've been doing organizational design for.

for many decades. The problem though is that a lot of work actually happens irrespective of those things, right? Boxes and lines and workflows. Yeah, for sure, but also work happens. in another way in in ways that kind of organic you kind of feel as a manager that you know your people are going somewhere and talking to other people and you can orient that a little bit but how do you How do you make that systematic? This has to do a lot more with system dynamics than the normal org design.

Obviously, it's a discipline that's been around for quite a few years now. What system dynamics basically say is that you cannot control all the particles in a complex system. So you cannot do that. This is impossible.

physics beyond the point doesn't collaborate and actually the physics of the physics of organizational design also don't collaborate but there's a nice book that's called social physics that sandy pentland wrote many years ago the people should look into as well so this physics apply

So physics cannot control the single atoms or molecules, et cetera, very well, but you can still control what is called high leverage points. And so high leverage points are things that if you touch, the system turns. And it's just a matter of discovering what are the high leverage points in the system. And I kind of always recommend to people four simple things to start with.

you know you think of this as a in in the words of tom malone a supermind right so if i have a super mind how do i build one Well, it's really hard to start neuron by neuron and connecting and the synapses and the axons is that it's impossible to do. But what can you do to inject the right behavior into this? morass of organic stuff so you do four things first you try to understand how to connect the right part the right parts of the brain and connection means making sure that

For example, people are discoverable to each other, or even resources are discoverable to each other, which might actually mean machines. So discoverability and the ability for people to find each other. curation curation means that you get information we we are awash in information there's too much of it and our brain is actually doing what it can to keep out all the stuff that is irrelevant

So you need to help this super brain curate the right stuff. And curate the right stuff means eliminating a lot of things that are just draws, but also avoiding that you end up with monocultures. I mean, if you look at social media, one of the reasons why we have bubbles is that they do a curation that is not functional to diversity of thinking. But, you know, there's nothing inevitable about it. I mean, in theory, you could set up different rules.

The third aspect is actually given the ability for the nodes in this network to connect with each other and collaborate on the basis of the information that we've gotten. All of these can be done through machines and then really using machines to help compute. We see it now with generative AI, it's quite interesting. You actually see how literally generative AI in many respects is a curation machine.

It curates what exists out there in corpora that are massive and kind of makes them digestible to humans, really. And so those are the four things that we could... try and think about and get machines to fundamentally do. And so that is how you do that. I think even a step before that, For organizational designers these days, what they need to do is to look at, again, four things. One is to get those nodes illuminated. Again, we talk about

How do you find each other? How do you find the right resources, the right people? Second is you need to energize the network. So energizing the network, if you're a CEO, a lot of what you do. is to create energy in the network by having what is called a reward function for the company that is partially embedded into culture, that is partially embedded into the systems of rewards, your KPIs, et cetera. So it really set in an orientation.

in that direction, but also then, you know, putting your money where the mouth is, right? So energy. And some of this is intrinsic, some of it is extrinsic. So, you know, designing the incentives for this whole system to go in the right direction. Third, you need to think about feeding this brain with the relevant knowledge. And the last component is really trying to get people to collaborate, people and resources to collaborate.

then obviously you can do that with technology and you know these technologies that do that which which again is this aspect of connectivity this aspect of curation the aspect of collaboration the aspect of computation So if you think about those, and this is actually quite practical, Chris. I mean, this is not a, oh, you know, this is a good conversation for Thanksgiving. It's you can actually, you have a lot of those tools already.

i mean if you're a ceo in a company you know that you have collaboration tools like i don't know you use microsoft teams you use slack you use zoom you use that kind of stuff you have information feeders because you probably have sharepoint and some companies have news clipping services and you know competitive intelligence all that kind of stuff The energization side, you do that already. I mean, you do KPIs, you do town halls, you do, you know, people.

fly around the world to go and shake hands with people all the time. And so a lot of that is happening. And the illumination is kind of done, I think is done. Again, the example of the water cooler is a good one because people typically do it with offices or they do it with, I don't know. yearly retreats, right? You go and meet your peers in your broader company, etc. But that illumination, again, requires a lot of network analysis and it very often is not done.

intentionally and systematically. So those I think are quote unquote low hanging fruits. Just looking at the tools that you have and try to think a little more systematically about how they.

they reinforce one another. That was a very good overview. And I like the fact, again, we moved from... kind of being reactive to be more proactive in the way you can design an organization and and i find particularly interesting the fact that yeah we can become more systematic in a way we introduce toolings in the organizational design which is not something that is done very often as far as i can see i can still see a lot of organizational design more as

whiteboard exercise sort of, instead of taking a more statistical and network analysis approach as the one you're suggesting, which is, as you mentioned, we have the tools, we have the capabilities, we have oftentimes in actually people dedicated to network analysis for specific product design.

But the same actually competence that you might have within a company can be actually used also internally to improve the organization itself and we don't often think in this term. So that's really relevant. I think now we definitely moved and covered enough ground looking at.

sort of how it works in organization, to move to the next steps. We kind of looked at what we can do now, what can be the next sort of more proactive way to build an organization. And then there is more a future looking view that you already mentioned in your introduction. which is really the how can we call that the sort of the artificial collective intelligence what maybe we could call it even super mind so how can we

How does this future look like for you? How do you envisage this augmented collective intelligence work with this combination of artificial intelligence and human intelligence?

Well, I mean, obviously there's an academic component, but I want to stay super practical. I mean, if you think about it, all companies are in the business of outsmarting each other, right? um and they do want to have the collective brain to be more effective not just smarter more effective which actually also includes an element of yeah you can be smart but you need to be able to act fast and execute seamlessly etc

very often the companies that win are all about execution, right? I mean, you have a good strategy, but then you out execute your competition. A supermind should be able to help people do that. I'll give you a couple of examples, again, so that we can keep it fairly practical. By the way, if people are interested, I... i wrote about 30 odd scenarios for the future they are called their own supermind.design if one looks into the resources something called futures 2030.

If you think about it, imagine you're buying companies, right? M&A. M&A is one of the hardest things. CEOs buy companies all the time. This is a way of buying yourself into capabilities, buying yourself into growth. The majority of M&A transactions fail, fail to result in meaningful accretion too. shareholder return, people who've been doing M&A know why. Sometimes it's the price, but very often it's just the complexity of getting people embedded into your company.

Cultures, organizational structures, the weak ties, and in many respects it's like... injecting a new brain into an existing brain and trying to hope that, you know, that thing embeds itself, etc. And that's one of the reasons why the M&A transactions, the acquisitions that were most successful typically been. fairly small things with very good technology product ideas embedded into a machine that can out execute those little things. But if you think about the future of that.

This is a very clear problem that CEOs don't really have a good solution for when they talk to the boards. And they keep on doing that because you have to do it. Sometimes you have no choice, but it's really a hit and miss. Well, a super intelligence constituted by an existing company plus an acquisition or a variety of acquisitions could be facilitated by the things that we talked about before.

So going and doing an organizational network analysis and using tools that support the rewiring of not just the strong ties, because that typically gets done. but the weak ties between an acquired organization and the acquiring organization. So that could be done. We have the tools to do it as long as you put people on the same platform. Some companies like, for example, Microsoft, if you look at the Viva side of Microsoft, does give you the ability to understand.

you know, almost like run a CT scan of your organization where you can identify the parts that have not been integrated properly and really, you know, build things like the one I just talked about before, right? And so... you can improve your your chances of a successful integration in an mna by doing something like what we talked about before

And M&A is also not just the people side. I mean, ultimately, it's the knowledge that these people have. And so you think about knowledge graphs now that are a form of curation of knowledge. only talk about generative AI, but we tend to forget that knowledge graphs are very important because they are the ones that help you connect the dots between the content side of the knowledge.

Most companies have some form of knowledge management, probably built five to 10 years ago when we had very poor artificial intelligence related to semantics. Those tools don't actually surface the right knowledge at the right time. They don't surface the right people with the right knowledge at the right time. So all that kind of things could be worked on.

with the capabilities that we talk about. And again, in the case of the M&A transaction, if you have two companies that are large enough, it's actually not that easy for a company. for a person in a company in the acquirer, say, to go and find capabilities in the acquired entity above and beyond the hierarchical connecting lines.

but as you know hierarchical connecting lines are just one way of connecting those you know and reasons why i mean it's like strong ties but those are those are not the only ties that exist so Again, you know, I envisage a future where even things like M&A could benefit from the augmentation of collective intelligence and the rewiring of companies in a way that... The entirety of the network is now accessible and powered up so that you really compound on the intelligence of each other.

as opposed to trying to do it manually, which is very often what has been done. There's other examples that one could go down, one could talk about sales, we could talk about R&D, et cetera. But I thought that it'd be important to explain. with concrete examples, what the future looks like. And then if people want to look at more examples and scenarios, they can certainly go into the document that I referenced. Definitely going to put your link to your website.

into the the podcast description so people can follow the link and find all the details because yes you actually provided already quite a few references i'd like to get back to some of your examples actually maybe I like actually the reference you gave me around sales and R&D, but give me a minute. Before that, I was thinking about a specific example I met a few months ago.

I'm pretty sure you're aware of Anonymous AI, which are doing experiments on swarm intelligence and really kind of using people, a swarm of people for decision making. Do you see that as a sort of realistic scenario? or still there is a way to work on that front. Yeah, I mean, Anonymous has been around for some time, actually, one of the more established companies. They do actually a specific thing.

So once you have a network, meaning you have identified the people, you can get them to take bets on bets, meaning it's based on behavior, it's based on how confident the actual. It's a head, so it's a mind-body interaction that has some signals that reflect into basically the user interface, and you can tell if people are really... confident about what they're saying or not. And that is one form of collective intelligence augmented by AI. The AI there understands how you're using your mouse.

your finger on the mouse, et cetera. And that's a very good example of one of the things that you could do with augmentation or collective intelligence. But it's a very specific thing. I mean, you think about it, you have your network. You've taken people into one place. It's synchronous. meaning that the decision is taken with everybody there at the same time and obviously kind of uh it dispenses with sorry the everything that needs to come upstream and downstream from that decision, meaning

The collection of the fact base, bringing people up to speed with the debate, all that kind of stuff, it doesn't do that. And it's also what it doesn't do, it doesn't put... things into practice and then learn from the results of the natural experiments and the experience, et cetera, that the organization and many people might have then taken the decision to have. So it's one, I mean, there's a workflow.

in collective intelligence and that is a very useful component of an end-to-end workflow. It certainly wouldn't do things by itself. I mean, the same way, by the way, we've been having things like prediction markets for many years now. Eli Lilly and other companies have been using them for many, many years. basically you bet on ideas that you think you're given money kind of fictitious virtual money you bet on ideas that then inform r d and marketing and that kind of things um

Those are useful, but they are not the only thing that you can do. The unfortunate thing I think with augmentation of collective intelligence is that just like scientific process design 30, 40 years ago. It isn't one thing. There's a lot of process design, there's a lot of human resources stuff in there, there's a lot of technology design. So it's a little bit of people coming up to speed with all these tools of which unanimous AI is one of them.

and then they can be embedded into the end-to-end fabric of the collective intelligence. That actually helps to somehow distinguish between the approach. wider collective intelligence versus swarm intelligence. And why, as you mentioned, we are, there is an overlap, but it's not the same thing. And one is more narrow, probably the other one is wider.

Going back to, I mean, you mentioned you had a few other examples that you can present as example in sales and R&D. And I like it because, at least in my experience in company, they tend to be.

also culturally at the two extremes of a company that's only the pipeline of creation. Yeah, so look, sales and R&D are interesting because For many companies, that's how you live or die, right? I mean, if you execute well in one or both, you're competitive. If you don't, you're dead meat, right? Especially in enterprises that really have... strong competitive pressures. Let's look at them in isolation and let's look at them together. If you think about sales.

So sales at the end of the day is the membrane between the enterprise and the outside world where you're asking for money. in exchange for your products and services. And so sales needs to be doing a really good job at translating back and forth. It's like a membrane in a cell, right? Translating the environment back and forth. And very often sales doesn't do that well because maybe there's too much variance. Some salespeople perform well, others don't.

um some focus on on the wrong products it's not just strategy a lot of it is just execution day-to-day execution and now you look at the augmented simple example of augmentation of that is One of the first examples of generative AI being applied in the enterprise is where salespeople start writing emails with generative AI.

right um so that's we that we've seen for many months now it was one of the first ones to come out and if you think about it what it does really is an ai that has been built on a corpus of collective intelligence itself i mean but basically what it does it regurgitates in a very smart way things and billions of people have been saying in billions of interact trillions of interactions and you know you can write all of a sudden

the right headline for your email, which is 90% of what people will ever see, right? And so that is a very good example of augmentation of collective intelligence. But the other thing that you can do and that you've been able to do somewhat in... for some time, salespeople need to have a little bit of background on the company and the person that they're talking to. And very often they don't have all that much time to spend on it.

might have been using google etc but that process is manual and and lengthy etc and and instead now what you can do is uh you know clearly get faster to crafting more relevant

messaging and more relevant information back and forth. So augmentation of collective intelligence there is important. R&D is another good example because R&D very often has the problem of interfacing with the world through a bit of a kind of a keyhole and that's the reason why many companies including say png started working on open innovation so that the r d gets

outside of its potential silo. R&D people very often are very technically smart people, but very technically oriented. So they tend to kind of silo themselves into their own domain not unlike academia academia by the way and so they become very very deep in a very narrow thing but the best r d as for example you know, the device that you probably have in your pocket, you know, the iPhone is cross sectional. So, you know, R&D of Apple, you know.

A decade and a half ago was people who knew material engineering as well as AI, as well as whatever version of AI was there, as well as... design engineering and design thinking and very deep in design. And so that cross-sectional thing is very important. So if you think about R&D and the augmentation of the intelligence of an R&D is really the ability to sense the environment.

and connect dots from somewhat unrelated spaces. And sensing the environment could be, for example, the sensing of the environment can be things like what is called low intensity, high momentum signals. So things that are coming at you from the very frontier of the space that you're monitoring, but its space is compounding.

And so those are the things that R&D needs to pick up because they need to start preparing for an interaction and inclusion into what they're doing. And the augmentation of collective intelligence can do that. I mean, if you think about the fashion industry, a lot of the fashion industry does.

they scour social media to identify low momentum, sorry, high momentum, low intensity signals. You really want to see something before everybody's seen it. And so that's a good example of collective intelligence.

The combination between sales and R&D is also another example of augmentation of collective intelligence. Sales is typically the first place where the signals get processed very often for R&D to... to be able to harness and do something with it, but also especially in environments that use lean startup methods.

The Salesforce is super important to do quick iterations on the ideas that R&D and product in general has come up with. And so the ability of those two groups to really see eye to eye. build strong connectivity of the not just the strong tie but also the weak tie side of the story have maybe some common watering holes where

they can maybe look at weak signals together and the salespeople may be able to contextualize things like that. Typically, no. every sales person will be able to do that and not every R&D person will be able to do that but there will be people who kind of act as connective tissue between the two and if you give them

curation environments, collaboration environments, and the incentives to do that, they will probably do a better job at identifying opportunities and also developing, again, with lean startup methods. the next set of opportunities that then gets baked into funding pipelines and then you know product roadmaps so those are examples of individually sales individually r d but also the combination of sales and r d

can benefit from this infrastructure of augmented collective intelligence. I like a lot the way you covered that. There are a lot of opportunities to be picked in the single departments, but also... about announcing collaboration and making closer and closer loop. In a previous episode of the podcast, I was talking with Luca Dallana. I'm also talking about closing the feedback loops and making it closer as an element of improving.

the way we sort of make decisions, taking into account that our teams and our organization are complex adaptive systems. And so you want to make sure that these loops are as close as possible and definitely, as you mentioned, sales and R&D. they can sort of make.

the success of an organization in a singular fashion, but also together. We definitely covered a lot of ground today. And I think, yeah, there is a lot of football thoughts for the podcast listeners, plus a lot of references I'm going to put. into the description to your website. You mentioned also a few books that readers can, if they're interested, can dive into. So I would close our greatest in chat with that.

question is becoming my usual final questions. And so can you describe what are for you the characteristics of a modern leader who embraces complexity? It's a fantastic question, by the way. I think at the end of the day, one of the reasons why the things we talked about is being done at scale is because our leaders were built. like us right in in times in which we didn't have the augmentation of collective intelligence we could talk collective intelligence competently probably since 2010.

And so leaders weren't, we didn't go to business school talking about this. I'm pretty sure that there wasn't a business school course about it. So the first thing about the leaders is being a little anti-disciplinary. You know, at MIT, this concept of anti-disciplinary has been around for some time. The idea is, you know, fundamentally, you want to have you don't want to be blinded because of.

what you think you know today, what you think that the disciplines are today. You want to end up with a mindset that allows you to... be maybe even just colorblind to some of the disciplines that you have today. What is R&D? What is sales? What is marketing? What is HR? What is the role of each one of those? Because if you're going to those swim lanes too fast.

then you'll not have the ability to rethink the brain, the collective brain, the right way. And, you know, think about it. I mean, if you were one part of the brain, you would say, oh, I am the neocortex. I don't know what an eye is. I rely on somebody else to do all that kind of stuff. And as a result, you may actually miss a little bit the forest for the tree.

if you will. So I think the importance of being cross-disciplinary, anti-disciplinary, being what, in other words, is called T-shape, right? The I-shape is a narrow person, a person who has narrow skills, but very deep in that space. T-shape. is a person who has that but also has the ability to go across many disciplines. We need more T-shaped people. Just the way we have T-shaped cells in our body, we need T-shaped leaders much more than in the past.

Any other suggestion for sort of any other characteristics that comes to your mind or mainly this one, which is, to be fair, pretty important and not obvious. I guess, you know, related to being anti-disciplinary in T-shape, you can think about what is in there. It means you need to be curious, needs to have the ability to learn fast and to be a really good beginner.

right in in the things that you do so you know since the obsolescence and the uh the skill obsolescence half time is is kind of dwindling right so the skills are becoming obsolete much faster than they used to in the past You need to be able to pick up new stuff much faster. Maybe you don't need to be an expert the way we used to be building our expertise in the past. You just spent 20 years doing the same thing. But you absolutely need to be a big...

really big at being a good beginner. And also ultimately, I think maybe for a generation of people, this is a little against the grain. Also realize that you are not the brain. You know, the brain is the collective brain and pretty much all you're asked to do as a leader.

is to weave that connective structure that is well beyond yourself and i know that a bunch of things you know from competencies to uh capabilities and and incentives for many people don't go in that direction but the the leader of the future would be really a weaver of nodes in a collective brain much more than a very strong person with a very strong team in somewhat of a silo.

And the ability to play that role is, I think, the defining factor for success in the future for many people in many jobs. That's great. I couldn't agree more.

Those are very, very interesting and relevant characteristics for a modern leader. And so thank you very much, Johnny, for being here today. And yeah, I hope to be able to continue this conversation in the future. Thanks, Chris. My pleasure. you

Thanks for listening to this episode of Engines of Creation. I hope you enjoyed learning from Gianni Giacomelli how we can all enable and take advantage of augmented collective intelligence. If you did and you want to learn more, be sure to follow Engines of Creation and leave us a rating and a review.

And don't hesitate to share this episode on social media and with friends and colleagues who might be interested. If you have any questions, comments or feedback, feel free to reach out to angelsofcreation at masterdonato.co. I'll be glad to continue the conversation. It's now time for a break and we'll skip August's episode, but Angels of Creation will be back in September with another episode where we'll discuss about time, Japanese folklore, and the art of mapping and much more. Arrivederci!