Authors: Dr. Jochen Schilcher (SNGLR Capital), Luca Sambucci (SNGLR XLabs)
Autonomous driving is a key element of smart mobility and one of the drivers of the mobility (r)evolution. All ingredients of a typical technological innovation are there, the technology, or rather the various technologies, which are closely interdependent, the novelty of an innovation and the market or basically the market relevance, as some of the applications are already available on the market and others are being tested in various pilots around the world.
For quite some time now there are daily news and publications about autonomous driving, whether it’s about the state of technology, new pilots and achievements, failures and accidents, new fields of application, new business models, or predictions when the breakthrough will happen. Basically, from all disciplines as the topic is highly complex and cross-sectoral. Still, often all sorts of things are packed together, applying different definitions, over-simplifying or the opposite, creating unnecessary complexity, sometimes with a highly positive tone in the sense of “mobility for everyone in low-emission, clean, safe and just livable cities” and sometimes with a rather negative or questioning tone in the sense of “is that really what we want … and who is going to pay for all this?”. Generally, news is delivered with well-known companies – Tesla, Google/ Waymo, Apple, Uber to mention just a few as well as from traditional automakers like Volkswagen, Mercedes, Toyota or Ford – to position themselves in the race for technology leadership in autonomous driving. And presumably – for some of the listed examples – also to justify an enormous valuation, as a lot of VC is invested, and stock market stories were developed. Autonomous driving is considered as a kind of holy grail to solve many challenges of todays motorized individual mobility (basically cars) as well as within the public transport. Be it in urban, suburban or rural areas as a smart mobility system or in a broader smart city context. But whether that is the case remains to be seen, at least for now, and probably for several years to come. To put the multitude of information into a strategically relevant context, connecting the dots from various perspectives and disciplines as well as understanding the different stakeholder groups to allow a meaningful discussion is indispensable. This requires combining technological expertise with strategic thinking capabilities.
Technology
It’s important to understand the difference between a technology and an application. The former is generic by definition and the latter has specific functionalities and thus fulfills a concrete purpose or addresses specific customer needs. In the case of autonomous driving, AI is the technology, not the application. AI as a technology is used in various applications in autonomous vehicles, in traffic infrastructure, or as part of a variety of possible use-cases in a smart mobility/ smart city ecosystem. AI is an umbrella that encompasses many methods related to emulate intelligent human behavior. Buzzwords in this context are Machine Learning, i.e., when machines autonomously discover complex correlations from large amounts of data, e.g., via pattern recognition, or Deep Learning, when neural networks used for Machine Learning possess a large number of layers, allowing for deeper and theoretically better correlations. Both belong to AI and both are only as good as the quality of the data processed by the algorithms – which we will not go into further detail here. At this point we want to pinpoint that AI is a generic technology, not a specific application. The distinction is key, because the adaptation of technological innovations relates to the functionalities. Transparency and information about the underlying technology however increases the adaptation as the assumed, hypothetical risks tend to move into the background and the promised benefits into the foreground. Why is this the case? Well, the hypothetical risks are rather associated with the technology and less with the functionalities. The functionalities are much stronger related to the benefits, which makes intuitively sense, as they are easier to understand. As a company, you should be aware of this because the corresponding measures for marketing are changing. On the one hand, a company has to reduce people’s uncertainties in respect to the technology and, on the other hand, highlight the benefits of the functionality. Currently, almost only the benefits of the functionalities are shown without a corresponding education about the technology to eliminate people’s uncertainties, fears and doubts.
Innovation and Market
It is also important to clarify the term innovation. An innovation takes place when something new is introduced to the market. Technologies are normally not introduced into the market, as they are generic. Innovations with specific functionalities are introduced as physical or digital products, as systems or subsystems, as a process, etc., always addressing or creating a customer need. These novelties partly drive business model innovations, especially in the context of AI and smart mobility. An innovation without a market is an idea, nothing more, but also nothing less. If – next to the market – the technology is added to the idea, you end up with a technological innovation. Why is this important? When it comes to investing into the future, it’s different whether a company invests and manages a technology portfolio or whether it is about partnering in a new business model as part of a go-to-market strategy. Also, from a capability perspective it is different building up respective capabilities internally or partnering with externals probably followed by questions around the best-suited locations, new ways of collaboration and so on. Combining technology with strategy capabilities is the basis for successful technological innovation. Sounds simple, in practices it’s not as it means to combine science-driven technology with curiosity- and creativity-driven strategy.
Degree of Innovation and Market Impact
Furthermore, the type of innovation plays an important role, namely whether it is incremental or radical. The latter can be associated with disruptive, i.e., when something existing is “discarded” and something (truly) new is created. If this is the case, industries are redefined, market mechanisms change, new business models emerge, new players appear from other industry sectors, and so on. In contrast, continuous improvements are referred to as incremental. In the case of autonomous driving, the term “disruptive change” is often used, meaning radical innovation. The question, however, is whether this is the case. From our point of view, this can only be answered in a meaningful manner by separating the technology and the market perspective.
Basis for both are the autonomy levels (L0 – L5) defined by the SAE (Society of Automotive Engineers, see https://www.sae.org/). At this point it’s not about SAE levels, so only a brief digression: L0, L1 and L2 are so-called ADAS functions (Advanced Driver Assistance Systems). These functions support the driver, e.g., adaptive cruise control, lane departure warning, acceleration or braking. These automation levels are daily business of the R&D departments of all vehicle manufacturers incl. suppliers and accordingly in the field of incremental innovation with minor to no market impact (currently we are at L2+). L3, L4 and L5 are so called ADS functions (Advanced Driver Systems; the “A” for assisted is dropped). Here it’s getting tricky for L3 and L4 as starting with L3 the car has an increased autonomy and the human’s role is changing from driver to an (active) supervisor. This introduces a novel set of challenges like human attention, human-machine interaction, legal accountability or ethical crumple zones. Strictly from an engineering or technology point of view, L3 and L4 are not completely different and that’s why we might still talk of incremental innovation from a technology perspective. However, what makes L3 even more difficult to design than L4 is basically the switch from “driver” to “supervisor” and the required differentiation of “high-risk, safety-critical” and “low-risk, non-safety-critical” driving conditions. The differentiation of L3 and L4 is thus primarily made via definition of the parameters and the triggered reactions. At L5 the system does not need a human at all, which is definitively the “game-changer” as even the vehicle will probably look totally different compared to today’s vehicle designs and configurations.
What should we think about? Whether L3 or L4 … what matters is the switch from “driver” to “supervisor” as this creates a difficult situation for the AI technology. A simple, binary categorization according to “risk” versus “no risk” no longer seems sufficient to cover the multitude of autonomous vehicle use cases. What is required is a “more nuanced, differentiated approach”. At this point, technology takes a back seat, and the market with its regulatory framework takes the driver seat to develop respective guidelines. This is when radical innovations emerge from a market perspective resulting in totally new business models within a new industry or better cross-sector “smart mobility”. The radical nature of the innovations is obvious as almost everything is affected, from the business model, the underlying value chains and its operations, the companies involved, the roles and interaction of private and public transport (e.g., robotaxis, shuttles), the cost calculations and the profit pools, etc. Above all, the users of mobility, the humans, will probably (or have to) change their mobility behavior when buzzwords such as shared mobility, mobility-on-demand, or ride-pooling are developed in the context of so-called mobility-as-a-service (MaaS) concepts – which by the way does not necessarily mean the individual mobility will disappear.
Key Take-Aways
As stated, technology and strategy must go hand in hand when it comes to autonomous driving and its role in smart mobility strategies. This is even more true within a broader smart city context. Some key take-aways:
-
First, it is key to understand AI as an enabling and basically an exponential technology in order to be able to develop future scenarios. With regard to the adjective “exponential,” this may require going further than what is currently conceivable. And this means to understand the opportunities but also the limitations of the technology and its applications on a timeline.
-
Secondly, it is essential to distinguish technology and innovation when thinking about scenarios. Investment priorities and investment decision, capability requirements, partnerships as well as go-to-market strategies might look totally different if doing so. Especially when it comes to the launch strategy and the accompanying communication strategy, it is highly important to consider the distinction and the effect on the different target groups. This refers to the human-factor in so called customer-centric strategies and the difference between technology-push and market-pull and should not be underestimated. In many cases, the adaptation of technological innovations does not follow a rational pattern but is characterized by habits and behavioral patterns, risk appetite, curiosity, basically often a kind of irrational behavior – especially when we talk about a basic human need like mobility.
-
And third it is important to understand and accept the kind of arbitrary nature of political decisions resulting in respective regulatory frameworks. In this context to be mentioned are especially the significant differences of political systems and decision processes in different geographies.
