Who remembers the biotechnology hype in the early 2000s, or the nanotechnology hype ten years ago? While the media no longer focuses on the billion-dollar bio- and nanotechnology startups, the technologies as such are now more sophisticated, and influence our everyday lives much more extensively than ever during the hype.
These technologies are present in energy production, construction, clothing, pharmaceuticals, almost everywhere. Their development and integration into old and new products is accelerating at a fast pace, regardless of public attention.
Against this backdrop, I can’t help wondering whether our current “AI hype” is following the same pattern that keeps repeating itself when a new technology reaches sufficient maturity. As we know, neither bio- nor nanotechnology was new as such when the general public became aware of them; there had already been major breakthroughs in these technologies years before they caught media’s attention. This is completely analogous to what is now happening with artificial intelligence.
Artificial intelligence today is based on machine-learning algorithms that were developed decades ago, and the exponential growth of computing capacity and big data. However, there is no single identifiable innovation in the basic technology explaining the AI boom. This is a fact to keep well in mind for companies developing products and applications with ”new technologies”.
”Yet Another AI Startup”
As with bio- or nanotechnology, a product based on artificial intelligence shouldn’t be developed with a “technology-first” approach, either. This is easy to forget during a hype, as businesses are set up first and foremost to develop the technology, and only afterwards focusing on the use case. I’m sure this is one of the reasons why many of the startups fail: they lack a clear product focus or vision. In the current AI boom, the saying ”Yet Another AI Startup” has almost become a cliché.
The lack of product focus or vision results largely from a gap between technological expertise and product expertise in a specific application sector. For instance in the context of legal services production, it’s not easy to find lawyers who, in addition to understanding the challenges of their profession, would also be able to identify the technologies providing solutions to the challenges. And vice versa, there are few software developers with in-depth understanding of a particular business area. Multidisciplinary expertise should be pooled into teams, but this brings other challenges.
In an attempt to mend their skills gap, technology startups intentionally abandon the clear product focus, and instead take a purely market- or user-driven approach, looking for ”creative solutions” enabled by the new technology to defeat their clients’ challenges. Abandoning the product vision in the name of overly agile development could be tempting for startups trying to justify their shortage of skills in applying technology to business for themselves, or potential investors.
As the vast majority of companies seeking external funding are those struggling with the talent gap, it is natural that most of the media attention goes to companies trying to mend their shortage of product skills with overly agile development: they need the media coverage as a way of raising finance. But despite the hype, in reality the ”new technologies” will find their niche as part of old, ordinary products, bringing improvements by not transforming the product completely.
Just like with any other product, a service product that’s been molded over the years into a well-working form shouldn’t be abandoned, or re-developed on a customer-driven basis, despite technological advances enabled by artificial intelligence. At first glance this might be tempting, as most of software development is done based on agile methodologies, which often drives projects to abandon strong visions.
It’s worth keeping in mind that to apply agile development methods in no way excludes a strong product vision, in any technological field. For a more detailed discussion, see e.g. Juho Kerppola, Agile software development – Case study about Varaamo, Metropolia University of Applied Sciences, 2016 (in Finnish). It is a misinterpretation of these methodologies, which as such are efficient, and which we also apply here at Papula-Nevinpat, that drives people to abandon their vision.
Clear product focus determines technology choice
I’ve been Chief Digital Officer at Papula-Nevinpat for about two years now, but have been developing processes to streamline the production of legal services for roughly nine years. During this time the office technology has taken huge steps forward. In our product development the core has always been the service product, not the technology or a specific development methology.
In our product-driven development the main goal has been to integrate new technologies into the complex system that is the backbone of our process in ways that would best serve our operations. Although agility is one of the key elements in our software development projects, we have made great efforts to ensure that our strong product vision would not be obscured due to overemphasized agility.
As office technologies are evolving, our service product is evolving as well. Over the years, we have seen emails replacing letters and faxes, network drives and document management systems replacing paper document archives, large displays replacing paper folders, and an ERP system replacing Excel spreadsheets. We have integrated our ERP system into external databases and application extensions over interfaces. Through all these changes we have managed to maintain our company’s profitability by focusing on our clear product vision.
Now we are integrating machine-learning components into our systems. Artificial intelligence, or machine-learning algorithms, are in many ways disruptive technologies, which enable us to create new products, business models and structures to support our operations. As we are developing systems based on artificial intelligence, we are also keeping in mind that integrating disruptive technologies into an old product is simply a natural part of the product’s long-term evolution.
Develop the product as a whole, not technology
In AI-based, as in any other kind of product development, it’s important to consider the product or system as a whole in order to integrate new technology into the product in an optimal way for the whole system. The main purpose should be to develop the whole product, not technology. Whether technological development will actually improve the product varies along the product’s lifecycle. Even if potentially disruptive new technology was available, it doesn’t mean that the tried and tested components of an old product should be abandoned just in the name of agile development.
I would argue that despite the prevailing hype, the vast majority succeeds in benefiting from new technologies by improving old products based on a strong product vision, although this is not always the impression we get from the media. And the majority of successful product development projects, whether they deal with artificial intelligence, biotechnology, or nanotechnology, are found within this large pool of technological development, which despite its size remains more or less ignored by the media.