Jolanta Kokosińska said in an interview in April 2018: “The 1st cohort of PwC Startup Collider was a trial, so to speak. We wanted to see how a program like that could work and if we had all the necessary resources and were ready for such an undertaking. Because we were successful, we decided to scale up within the structures of PwC and instead of only focusing on FinTech startups… in the 2nd cohort we had startups from as many as 7 different areas of innovation.”
What is the most interesting aspect of working with startups developing AI? Aleksander smiles and says that for him, working with startups that create intelligent systems is always an amazing experience, because those companies basically teach machines how to think. They ‘dissect’ the whole thinking process, redesign it, trap it in a box using software, using ‘magic’, and make hardware think – that is in his view the most interesting thing.
AI in the future? “It’s very hard to guess – to forecast – what the future will bring. However, AI probably will take a step back in a sense that it will become less visible, but not in the least less advanced. It only goes upwards from here.” The future is simple he tells us: “AI does the job for you, but you don’t see it. You won’t see a lot of robots running around the place like it often happens in sci-fi movies. Robots with red eyes, for example,” he adds jokingly and adds that machines and programs running them will be hidden and will be performing their tasks, for us, in the background. “This AI-hype with high AI-visibility, machines occupying human space, becoming more and more commonplace, all this will very soon, I think, start to slowly disappear and merge with our surroundings.”
Without a doubt there already are industries that are AI-ready. Those are the industries that use a lot of computers, ultramodern software, hi-tech hardware and advanced machinery. Aleksander sees 4 industries that will be impacted quickest: manufacturing, telecommunications, data communications, even retail, because of enormous amounts of data (our purchase history or preferences, for example).
“These are the industries where it is the easiest to implement AI, where AI researchers can experiment the most and where visionaries are also most welcome, because we need to make tedious processes smoother, quicker, more automatic and requiring less human involvement. The gap between human and non-human intelligence will be bridged in these industries quickest and soonest.” Will robots take people’s jobs away from them? “The employment structure will change on many different levels and at different speeds. So, on the one hand, the number of people doing repetitive, mundane jobs is going to gradually decrease, but on the other, new jobs for highly qualified technical staff supervising the robotic element will be created. It’s a win-win scenario provided we invest in future skills on a global scale.”
The road to a new, better world is bumpy. There are many challenges that innovators and visionaries face, and they make mistakes. “The biggest challenge for companies starting their AI journey is that they usually overcomplicate their solutions. They immediately start with extremely sophisticated algorithms, because they think it’s cool – they want to show off. However, it’s not what’s the best for them or their potential partners, not to mention the product,” Aleksander tells us.
Tomasz Wesołowski, who brought Edward.ai to life, agrees speaking from his own experience. “Our main mistake was that in the beginning we tried to create a big standalone product. Only later did we figure out that replacing an existing product with Edward.ai was not a good idea, so we decided to change our approach. We created a small AI-based add-on, so basically we created sort of a next layer in existing software. Now our customers don’t need to install new software or re-learn how to use yet another CRM tool, because Edward empowers those existing tools with new features and work becomes faster.”
“My recommendation is to start small – it doesn’t necessarily have to be deep learning in the early stages of your work. Sometimes smaller algorithms, less complex ones are what does the trick and that’s where the magic happens. Not everything has to be super huge and overly complicated, especially AI – unlike in science-fiction novels and films,” Aleksander explains.
Others, however, prefer complex algorithms. “What we do at AlphaBlues is we automate the customer service support using AI. It is precisely machine learning and deep learning that help us figure out what people actually want. Armed with this knowledge, we create bots that answer people’s questions and make them happy. Most problems that we face, the stress, the doubts – it’s all in our heads and that’s where we must fight our battles,” says Indrek Vainu, Estonian innovator behind AlphaBlues’s success.
PwC’s global expertise allows entrepreneurs from all industries to become better versions of themselves and the company learns together with them. “We have a lot of experience with AI, bright scientists are working with us on daily basis and so are industry specialists. We bridge the gap. Anyone interested in understanding how AI startups can become successful or in improving their own intelligent solutions and products is invited to ask us for advice,” adds Aleksander Fafuła.
There are risks, barriers, and problems that need to be addressed as well. “The danger for AI lies therein what people think when they hear Artificial Intelligence. I have already mentioned science-fiction. We have worked, for instance, on AI-based solutions for human resources,” says Aleksander describing people’s reactions as those of shock or disbelief. “What? A machine is going to fire me from my job? That’s actually what happens – people are irrationally afraid of machines and this is what hinders AI from developing even faster,” he recalls.
“People avoid it – just in case, even though it might be just an app installed on a smartphone or a program on the hard disk of a computer. And it’s smart only in a very narrow sense. It’s neither Terminator, nor a UFO. But then again it’s just human nature: we are very often afraid of what we don’t know, what we don’t understand, what is new or different,” concludes Aleksander.
Tomasz sees another problem: in his view, people have too high expectations and they find AI-solutions insufficient. “A lot of people expect that Artificial Intelligence will solve all their problems in 1 day. This is just the beginning – we will have to wait for many years until AI-based solutions start to work like a human brain. For example, our product is able to assist salespeople in basic daily activities, those most boring and time consuming. But it can’t do all the work. The human element must still be there.”
And does Indrek see any risk arising from AI and its rapid growth on global markets? He replies “Yes. But not the ones you think. So, no – no Skynet.”