Finding Nemo HIRO™

I have to start with a confession. We haven’t lost HIRO™, our cutting-edge AI platform. Nor have we lost our charismatic CEO on one of his many trips to see potential investors or educate the broader public on the benefits of AI. We also haven’t lost anyone from our sizeable R&D team in any of Frankfurt’s myriad of apple wine bars. The title is a shameless plug to grab your attention. In a slightly stilted way, the real question is, how do we market HIRO™? How do customers actually find HIRO™ and understand its unique capabilities?

When I use the word “unique” I don’t mean the usual marketing hyperbole, but the fact that there is no commercially available product out there that we could be directly compared with. In terms of philosophy, we are close to Google DeepMind. Just like DeepMind, we are developing an approach that can learn to solve any complex problem without needing to be specifically taught how. But you can’t buy DeepMind off the shelf. The good folks at Google do some wonderful work for a leading London eye hospital but otherwise, it is not commercially available. For many, DeepMind is the poster child for progress toward Artificial General Intelligence (AGI). Crucially, what Arago is doing is having a comparable mindset but applying the research commercially. For many years we have worked on use cases around IT Automation, such as IT support and network monitoring and are now aggressively advancing toward business process-centric use cases.

Finding Nemo HIRO™

Pixar ©

However, having highly differentiated capabilities is a double-edged sword when deciding on the most effective marketing strategy. We could either use a significant amount of money trying to create our own “complex problem-solving AI” market segment. That could be a facet in the ever so noisy AI discussions. However, there is a risk of stakeholders becoming tone-deaf. Or, we could fall back on established notions of AI, even though they might not tell the full story of our capabilities. For me, the obvious one would be the notion of Intelligent Automation. While it doesn’t cover all our capabilities, it could get us a place at the table where the majority of sourcing decisions on innovation and AI are being discussed. As Chris Boos, our CEO, keeps pointing out, his firm belief is that anything that is a process can and will be run by an AI. This view is an antidote to many client discussions that we have, where we are being asked: “where is the AI?” You can easily translate that into “where is the magic?” However, unlike the technologies that we carry on our smartphones or the compelling adverts on AI that tackle mankind’s big problems, AI in the enterprise is largely unspectacular. Rather the crucial issue is about managing the inherent variability in these processes and using AI to scale and extend existing automation deployments. It is about issues such as scheduling, prioritizing, orchestrating and ultimately automating process steps. If anything, the “magic” should be seen in being able to do this at scale and without limitations in terms of the nature of the underlying processes and data. Crucially, HIRO™ allows you to do all that without needing to rip and replace your existing investments, rather it is leveraging and thus ringfencing those.

General AI: Diverging AI discussions in academia and the applied commercial world

The notion of General AI is a segue to the diverging discourses on AI in the academic and the commercial world. In the academic world, there is an enthusiastic discussion going on around this notion. General AI or AGI is seen by many as the Holy Grail of AI itself. While there is no common definition of General AI, in Arago’s view it is a universal AI and denotes a system that can handle tasks of different areas and origins without needing to rebuild models or robots. It applies experience from one area to another and thus learns faster. The crucial element is progressing toward knowledge transfer which is being enabled through a semantic connection between different areas of knowledge.

Yet, in the context of discussions on socio-economic implications and the commercial world, this enthusiasm for technological progress often is seen with rather more negative connotations. For many, General AI is equated with discussions around the so-called Singularity – when machines conceivably become smarter than humans, thus running the danger of taking over from them. For others, there is a conflation of reference points ranging from Sci-Fi movies to the fear of job losses. In the end, it is probably futile trying to be forensic about those connotations, but it is a critical consideration when deciding your marketing strategy.

Automation and AI are in the eye of the beholder

To bridge both worlds, we are starting to position ourselves as Integrated Intelligent Automation through an AI platform strategy. The positioning around Intelligent Automation is meant to emphasize the commercial application of our advances in AI. It allows us to address the horizontal processes and help clients with one of the most pressing issues, namely scaling their automation deployments. But our platform approach goes far beyond just scaling automation. Clients can build their own applications on top of our platform. This opens opportunities for scenarios such as automating IoT devices and providing a centrally managed approach for Autonomous Driving that go far beyond Intelligent Automation.

Regardless of how we want to be perceived, in current client discussions, we often get compared to narrow or point solutions such as RPA, Runbooks or AIOps. Occasionally we get pitched against automation platforms of service providers or IBM Watson. What this tells me is that many clients are still experimenting with AI and looking to incrementally optimize their processes rather than redefining their digital operations. Yet, the latter is actually what Arago is proposing. To help organizations fend off disruptors and become “AI ready”, HIRO™ is providing them with one platform and one data pool that allows them to dynamically scale their automation projects. The average automation rate achieved for business processes, IT Operations and new business models is a staggering 87%. By retaining and re-using the operational knowledge it has captured, HIRO™ is offering clients significant benefits in a short period of time. What many stakeholders are unaware of is that most AI solutions require training from scratch for every task. Through knowledge retention, HIRO™ can build on and progress from every problem it has solved. Furthermore, by integrating a broad set of algorithms into our proprietary assets, clients can overcome many of the limitations of individual algorithms and approaches. The limitations are most notably the inability to recognize the appropriate context of situations or dealing with changes in the environment or processes. Put in simple terms, the capabilities of HIRO™ are a far cry from the narrow applications of Machine Learning or Chatbots that dominate the discussions in the industry. Our R&D team led by Chris has created new sets of algorithms that go way beyond my very confined technology knowledge. Crucially, they have created a hybrid algorithm, using components from reasoning, learning, and semantics. This allows for two things. First, explainability because we use human-readable rules as the foundation for decision-making. Second, having one engine for multiple problems. Without pretending to fully understand all the nuances, I can confidently proclaim that we really are in a category of our own. Suffice it to say it is Arago’s duty to explain and demonstrate those capabilities as well as the results achieved. Collaborative intelligence.

Bottom-line: Beyond the moniker, AI is actually all about a change in mindset

While it is challenging to market the capabilities of HIRO™ in these early days of AI effectively, it is a luxury problem to have. Because Arago is driving ground-breaking research advancing the fundamentals of AI while at the same time applying those unique capabilities commercially. Yet, to take advantage of our capabilities, clients have to fundamentally change their mindset. The goal has to be to progress toward the management and automation of knowledge. They have to move beyond just optimizing a plethora of tool sets. But they also have to resist the temptation to sprinkle some Machine Learning across their operations in the vain hope of accelerating their digital journey. Rather, HIRO™ allows clients to manage the complexity of digital operations with one engine end-to-end.

However, to achieve those goals clients must blend technology, process, and people. Whatever the moniker, all those innovations typically subsumed under Intelligent Automation and AI require a different mindset and a different way of working. The only way to get there is proactively pushing change management at all levels. We might not have found Nemo, but our cutting-edge technology is one of the most compelling change agents that I am aware of.

Tom Reuner, Head of Strategy

Arago Redaktion 3. June 2019