SaaS, PaaS, and now AIaaS – Enterprising and forward-thinking companies will try to provide customers of all kinds with plug-and-play solutions powered by artificial intelligence for a myriad of business problems.
Industries of all kinds are embracing out-of-the-box AI solutions. According to industry experts, the global revenue of artificial intelligence software, most of which is artificial intelligence software as an online service (AIaaS), will grow at a staggering 34.9% annual rate, and the market will reach more than $ 100 billion by 2025. It sounds like a great idea, but there is one caveat: the “one size fits all” syndrome.
Companies looking to use AI as a differentiating technology for business advantage, and not just doing it because that’s what everyone else does, requires planning and strategy, and that almost always means a custom solution.
In the words of Sepp Hochreiter (inventor of LSTM, one of the most famous and successful AI algorithms in the world), “the ideal combination to get the best time to market and lowest risk for your AI projects is to slowly build a team and also use proven external experts. No one can hire top talent quickly, and worse, you can’t even judge quality during hiring, but you’ll only find out years later. “
That’s a far cry from what most of the online artificial intelligence services available today offer. The artificial intelligence technology that AIaaS offers comes in two variants, the predominant being a very basic artificial intelligence system that aims to provide a “one-size-fits-all” solution for all businesses. The modules offered by AI service providers are intended to be applied, as is, to anything from organizing a warehouse to optimizing a customer database and preventing anomalies in the production of a multitude of products.
There are several companies that claim to provide AIaaS for automated industrial production. Most of the successful data presented by these providers is based on individual case studies, with problems involving limited data sets and limited generic objectives. But generic AI solutions will produce generic results.
For example, the process for training algorithms to detect wear would be different for factories that produce different products; after all, a shoe is not a smartphone it is not a bicycle. Therefore, for “real” AI work, where smart modules really managed and changed production In response to environmental and other factors, companies developed customized solutions for their customers.
Many customers who were “burned out” by a bad experience with AIaaS will be more hesitant to try again, as they feel it is a waste of time. And use cases that required heavier AI processing did not give the results expected or promised. Some have even accused cloud companies of deliberately misleading customers, giving them the impression that standard AI is a viable solution, when they know full well that it is not. And if a technology it does not work enough times, it is likely that those who could potentially benefit from real AI solutions give up even before they start.
The goal is to standardize a solution that works well almost immediately and does not require extensive knowledge. Until now, the success of AIaaS has been to enable researchers to run complex experiments without requiring the services of an entire IT team to figure out how to manage the necessary infrastructure.
In the future, AIaaS is expected to allow non-artificial intelligence experts to use the system to get the desired results. That said, automated AI services online, even at their current levels, can greatly benefit industrial production, if done right.
Correctly done AI could provide huge benefits for the industry. Rather than giving up on AI, companies should delve into the AI services they are thinking of using. Does the solution provide for customization? What kind of support does the service provide? How is the algorithm trained to handle data specific to your use case? These are the questions companies should ask themselves when purchasing AI services. Vendors that can provide substantial answers, and back up their claims with real data on success rates, are the ones companies should work with.
Like all new developments that improve business activity, AI applications require a high level of expertise. Engineers working for large cloud companies have that experience, which means they could provide much more value to customers by helping them develop custom solutions. It is necessary to examine whether this can be done “as a service”, but the system that has been implemented at the moment is not the answer.