Co-Creation

From www.aiotplaybook.org
Jump to navigation Jump to search
More...More...More...More...More...More...More...More...Co-Creation

Co-creation between different companies can be an attractive alternative to the more traditional buyer/supplier relationships. This chapter examines different co-creation models, specifically from an AIoT perspective, before bringing in expert opinions from different perspectives, including enterprise, start-ups and venture capital.

Why AIoT & Co-Creation?

What are reasons for co-creation and partnering in an AIoT initiative? On the more strategic level, branding, access to an existing customer base, the exiting global footprint of a partner, access to physical assets or outlets (e.g. repair stations), or domain know-how and existing applications can be good reasons.

Since data are the foundation of any AI-based business model, this can play another key role. Often, data-related co-creation is about the federation of data from two domains, with the expectation that the sum here is larger than the parts. For example, co-creation partners can federate IoT-generated data from multiple machines in a manufacturing setting to support a more holistic OEE perspective for the end-customer.

The development of AI/ML algorithms can also be an interesting area for co-creation. Of course, this will often also be closely tied to the data side of things. For example, the massive costs of developing AI for autonomous driving make some OEMs consider a co-creation/strategic partnership approach in this area.

The need to combine different, highly specialized technologies in the context of AIoT can be another reason for co-creation and partnerships. This can be the combination of different IT enabling technologies, or the combination of IT and OT technologies. Especially with IT/OT, it is often the case that the required expertise is not found in a single company.

Platforms are another interesting area for co-creation. Especially in the area of platforms that need to combine know-how from different industry domains, or provide some kind of data federation, this can make sense. Co-creation here is not limited to technical co-creation. For example, if partners decide to create a platform as a joint venture because they want to pool data, this can also be seen as a form of co-creation.

Why AIoT & co-creation?

AIoT Co-Creation Options

Because many people have a different understanding of what co-creation actually means, the following provides a short discussion of common patterns.

In some cases, companies that are actually in a traditional buyer/seller relationship will extend this relationship, e.g., by creating a press release about a "strategic development partnership". Or, they will apply value-based pricing, i.e., the seller is participating in the business success of the buyer. In the case of smaller suppliers, it is also common that the buyer insists on a source-code escrow to have access to the sources in a worst-case scenario or even insists on a stock right of first refusal. These are examples that we would NOT consider to be co-creation, because they are too close to a traditional sourcing relationship.

Co-creation can be more technical or more focused on joint Go-to-Market (GTM). Technical co-creation usually means that -- at least -- two companies are combining their Intellectual Property (IP) in order to create something new. For example, this could be a company with deep industrial domain know-how partnering with another company that has deep AI experience. Another form of co-creation is focused more on combining two existing offerings into a new offering, but without a deep technical integration. This could mean truly combining two existing offerings or actually selling one company's offering via the channels of the other. Of course both approaches can be combined.

This does not have to be limited to 1:1 partnerships but can also lead to multiparty ecosystems.

Platforms are another area for co-creation. A partner platform will allow partners in a closed ecosystem to work together. For example, a platform for an industrial robot could allow partners of the robot manufacturer to submit applications, which are then operated in a semi-sandboxed environment. Because all the partners are well known and trusted, this approach will be possible without a fully secure execution sandbox. This makes sense, especially if the cost for developing a sully secure sandbox is prohibitive or technically impossible. A fully open platform will have made the investment to develop a secure sandbox, which will enable onboarding of unknown and per se untrusted partners. This will be even more important if apps are deployed not only in the cloud but also on physical assets. In an AIoT scenario, the sandbox will need to provide execution capabilities for code as well as AI/ML algorithms.

AIoT co-creation options

Expert Opinions

The following interview provides insights on AIoT and co-creation from different perspectives, including large enterprises, venture capital, and research. The experts are:

  • Jean-Louis Stasi, Senior Vice President, Strategic Partnerships with Startups, Schneider Electric
  • Dennis Boecker, Global IoT Innovation Lead at Bosch
  • Ken Forster, Executive Director at Momenta (VC)
  • Prof. Heiner Lasi, Director Ferdinand-Steinbeis-Institute

Dirk Slama: Jean-Louis, how is Schneider Electric managing co-creation?

Jean-Louis Stasi: Schneider Electric has twenty lines of business in different areas of energy management and automation – each one roughly with revenues of a billion Euros. Each has its own market environment and its own R&D roadmap. Each is facing different regulatory requirements, which has a huge impact on innovation as well. Take the highly regulated electricity grid market versus fast moving areas such as data centers. Each business has its individual set of conditions in which they are operating, and that is nothing new. However, this defines the appetite for co-creation for each business differently and the different ecosystems they are focusing on. The starting point is always the same: what is the problem that the business is trying to solve? They see the opportunity and the value of the market, but they do not have the required skills and resources internally. They understand that this is going to happen in the next two to three years – so they know they have to move fast. Cyber-security is a good example. That is the next big thing for every industry customer in the world. If I do it myself, it will take me six, seven years. The alternative is to find a start-up as a partner and then scale up together. Of course, going into such a partnership mode is a structural and strategic decision.

DS: So the main rationale for co-creation is time-to-market?

JLS: No, not necessarily. There are other factors such as better understanding of market cycles or specific use cases. Another aspect is the management of business model evolution vs. technological game changes. So I do not want to emphasize time-to-market only, but the point is it has to solve a real-life problem. It has to solve a big problem and has to happen in the next three to five years. Those are the three key conditions.

DS: Dennis, what is your take on this from the Bosch perspective?

Dennis Boecker: I agree in general with what has been said already. What I would probably emphasize more is that we're coming from our strategic search fields. Many of these strategic search fields are cross-divisional topics, because major trends such as mobility, construction and building technologies cannot be limited to a single line of business. The strategic search fields agree with the executive management, and are aligned with the different lines of business. The interesting question then is how you address the different strategic search fields. We have identified four different ways of doing this: Corporate Product Innovation, Strategic Partnerships, M&A, and Start-up Co-Creation. Corporate Product Innovation is an internal approach in which we focus on building up our own Intellectual Property (IP) and capabilities. Strategic Partnerships are for big projects, taking more a consortial approach. M&A for us is the enhancement of our own IP and our core capabilities. Finally, Start-up Co-Creation focuses on very concrete portfolio elements, or very specific problems we need to solve. Here, time-to-market is usually more important than IP. Say, for example, we are involved in a large bid ourselves, where we have identified a gap in our own portfolio. A start-up is able to address this very quickly. We can even work on several portfolio elements with different start-ups. This is a good way of creating an ecosystem. You share a problem, and then you start working on concrete projects, either in a more strategic partnership, with start-ups, or even a combination thereof. Co-creation for me is not a one-to-one model, where two companies work together to create a specific piece of IP. We truly want to build open ecosystems around our strategic search fields, including multiple partners. And then applying any of the four collaboration models which I just explained, depending on the situation.

JLS: Interesting. Schneider Electric has a very decentralized model of control. Each line of business is pretty much autonomous in their selection of partnership models and specific partners. So there is no centralization of this element. At the end it produces a central result, where each business has its own iteration on that. An important element is the intuition between a start-up and the larger structure. A start-up is not necessarily three guys in a garage. It can already be scaled by a team of 200 engineers, which have already scaled to a certain level of product-market fit. As we grow more confident in the potential for scaling their business, they become an M&A target for us. However, this is a continuous process, and it is not decided in the early stages. It takes some time for a partnership to evolve into something more, before it becomes strategically impactful for us at large and we are prepared to make a move in the direction of M&A. Another important aspect in this is the geographical footprint. Because what we see is of course markets are evolving at a different speed, depending on where the innovation is happening. Is it US? Is it Europe? Is it China? And each market also has its own way to do things. So we also have this dimension that we need to integrate into our model because as you develop that globally, there is no one way to do that things. Each kind of region has its own specifics, depending on the maturity and culture.

DB: I agree with you on the necessary evolution of partnerships. There is definitely a chance that the portfolio approach and the respective relationships with the ecosystem develop from one quadrant into the other. But in the very beginning, I think you need to be very clear if this is something to share with the ecosystem or if it is rather something that you want to own the IP ultimately yourself. That is something that you need to be very clear about.

DS: Let us switch the perspective from the enterprise side to the start-up side. Ken, Momenta is an investment fund that specifically focuses on AIoT companies. What are you telling the companies in your portfolio: look for a buyer-seller relationship with large enterprises, or focus on strategic partnerships?

Ken Forster: We see it as an ecosystem of many-to-many, with many actors contributing to the overall value of solutions. Let me first outline the actors and then I can reflect on the role Venture Capital plays. I will generalize four key actors in the AIoT ecosystem: incubators, innovators, incumbents, and the implementers. As investors, we broadly operate as an incubator working to grow young companies that are often innovators. The large strategic Operational Technology (OT) players are generally incumbents. Finally, we have the end users, implementers, those leveraging our collective technologies and tools to create business outcomes. Between these actors, Venture Capital sit at the intersection of innovators, incumbents and implementers accelerating the velocity of innovation. Early stage companies often do not have the resources or experience to dance with the elephants of industry, so we bridge the gap, investing behind the innovators to accelerate their capabilities to play in the larger ecosystem. While our initial value is in providing seed to later-stage capital, the larger value is often the acceleration we bring via our AIoT ecosystem networks. As an example, we organized a consortium of four of our portfolio companies last year to demonstrate an intermodal container location and condition monitoring demonstration for the U.S. Department of Defense. The pilot demonstrated the use of AIoT technologies to track the location of containers through the full journey of shipment while recording the condition of the cargo within those containers, including security. This was truly a team effort with each portfolio company providing a critical piece of the solution. The DOD in this case was the implementer actor, our portfolio companies the innovators with Momenta acting as the incubator. Of course, once product-market fit is demonstrated, the incumbents will often play the strongest role: providing the 'voice of the customer' backed by their own size and momentum to scale up these innovations to the enterprise scale.In summary, we operate at the Venn of innovation - bridging the innovation of startups with the enterprise scale of large industrials. Our tools are capital, securing key leadership and teams, and activating partnerships early and often with companies such as Bosch, Schneider Electric and other market leaders.

DS: Thanks. Now let us look beyond AIoT-enabling technologies for a moment. Heiner, another key ingredient to AIoT is data. In your research, you are focusing on ecosystems for data sharing to support different AIoT use cases. Care to explain?

Heiner Lasi: As you said, an AIoT use case needs data to operate on. If these data are created and processed within a single company, everything is fine. However, in many cases, you will need to cross these boundaries. Either because you need to combine data from different IoT-objects/ companies, or because companies need to access data from other IoT-objects/companies in order to create business value. Let us take a logistics company that is running a fleet of forklifts. The initial IoT business cases here were very much about operational efficiency. This was all internal. But now we are seeing “… as a service” business cases. This means the logistics companies and forklift manufacturers must share operational data. In addition, insurance and finance companies are becoming involved in these business cases. On the technical level, emerging concepts such as cooperative Data Spaces can help here. However, the issue here is not technical integration but rather the creation of a partner ecosystem that is developing the required levels of trust to share operational data between the partners. This is not about technical trust; this is about trust on the business and even the human level. Transparency is important. This is what we are trying to address with our concept of data coops. These data coops are a manifestation of a data-centric ecosystem, with clearly defined rules of engagement, and a data space as the underlying technical platform. The rules of engagement must address how – for example – the logistics companies, the forklift manufacturers, and the financial companies are providing and accessing data, and how the benefits are shared. While AIoT is an important technical enabler in all of this, it really comes down to creating an ecosystem of partners who trust each other. And it takes time to develop the required level of trust.

JLS: This is an extremely critical point. What we are seeing at Schneider Electric is that there is a very strong relationship between the time you connect an asset with a sensor, and the time you get tangible business value out of it. Time-to-value is of the essence. I think we – all the industrial companies – have a long list of IoT use cases which have created zero value. Anything beyond three months, you lose the momentum, you lose the sponsorship. If you are thinking too broad, try to integrate too many stakeholders, things get too complicated. Every factory you talk to is adding new requirement. The conditions in the factories are different. The cloud infrastructure is different. The level of knowledge is different. It never ends. And suddenly you are caught in the “pilot dead end”.

To Heiner's point about the data: the challenge is to ensure that these data are turned into a tangible benefit for the customer in a very short amount of time. To achieve this, you have to maintain OT leadership in the team. You must give them the value, which means you also must give them the leadership. If you give the leadership to the IT guys, it is finished. That is our experience.

HL: I agree. However this is not only about IT vs OT. This is also about creating a link between business experts from different industry domains, such as OT and financial services, e.g., to enable “as a service” business models. In this example, you need to combine the experience of running the OT side with the experience of understanding the inherent financial risks.

JLS: Let us take supply chain visibility, e.g., in the food and beverage industry. This is a good example where we need to unify data to obtain increased visibility. It took them years to realize that yes, it makes sense to share data, because we have the same suppliers and we want to make sure that this food product truly is coming from the right place. It took perhaps five years to reach this level of trust in this ecosystem. But then we are coming back to time-to-value: People do not have five years to get to the outcome. They have to create value in the next six months. So there is this big tension in the AIoT space regarding the time it takes to build trust. Maybe eventually there will be a technology such as blockchain to intermediate the trust, but I think thus far this has not yet scaled beyond one or two verticals. I want to reiterate this, because I think AIoT has been failing in many, many cases because of this lack of value in a given time frame. In addition, I think we are still on this journey where it takes more value creation to justify a vertical integration, end-to-end.

DS: Does this only apply to what we have been calling the long tail of AIoT, or are we also talking about the short tail of AIoT here?

DB: Even for the high-impact products on the AIoT short tail, you need to ensure that you are delivering value creation along the way. If you have this huge AIoT short tail opportunity which you know will take you three to five years to deliver, you better make sure that every couple of months on the way to it you show tangible value creation. This is what I was referring to earlier on regarding portfolio achievements and strategic partnerships. On this level, it will take a longer time, and you cannot expect to get the full outcome in three months, because it’s just too much and too complex. However, you need those complementary, smaller items along the way to show that AIoT makes sense. Admittedly, with physical products involved, finding a stable Minimum Viable Product is much more difficult, but still…

KF: In the past, AIoT products often required full stack solution development. As an example, Nest had to develop the full stack of software, hardware and connectivity to build a smart thermostat. Today, this development is more horizontal allowing solution developers to choose from best of breed components leveraging standards. In this momentum from highly vertical development toward more horizontal solutions, there is even more opportunity to work with an ecosystem of companies that are aligned around domain specific use cases.

HL: We see a similar trend for data coops. Usually, they are initially clustered around domain-specific use cases. Therefore, the data coop with the underlying data space is the platform, with the different use cases representing long tail opportunities. Because there is an upfront investment in setting up the initial platform, it is key to immediately show value creation with the first use cases.

JLS: If you are talking about high-impact products such as the first iPhone, I agree. They take significantly more time and higher investments. Even if we are talking about IT/OT integration, it is important to understand if you are talking about the device manufacturer, or the implementer. If you are talking about the manufacturer of a new AIoT-enabled device, I agree with your characterization. However, if you are talking about the implementation side, where you are addressing many factories in different locations, you need to do this in a very consistent and predictable way. You will need a very simple and easy way, so that every operational team can implement this in less than three months. Than you can reach massive scale because you are making things so simple. Then, it becomes like the SaaS (Software-as-a-Service) model. This is where AIoT today starts to become relevant for industrial companies, because end users are able to scale this up themselves.

DS: Thank you all!

Tradeoffs

To conclude the discussion on co-creation and strategic partnerships, let us take a short look at the pros and cons. For example, some of the risks and drawbacks include:

  • Technical, legal and commercial complexity, as well as generally increased stakeholder complexity leads to a significant increase in project risks
  • Colliding interests can lead to failure or at least unbalanced partnerships
  • Limited visibility and lack of transparency can impose significant risks

Some of the benefits include:

  • Significant business and technical synergies
  • Unlocking creative potentials that would not be possible within a single company
  • Access to markets otherwise not easily accessible
  • Combining speed and agility of startups with global reach and execution capabilities of incumbents

Before making a decision on co-creation vs. traditional sourcing, these factors need to be carefully weighed.