"It's naivety that allows people to do great things."
Markus Rautio, Chief Technology Officer at ultimate.ai, talks about data security, NLP and daily challenges of deep tech companies.
Can you briefly describe your role and tech mission at ultimate.ai?
Originally, my role was about building the website of our product. This means the dashboard and the chat connections from our AI to the contact center platforms. Today I am still working on web development. Additionally, a big part of my job is to talk to our current customers and prospects about our technical capabilities and the needs of those clients. As an example, we have a weekly meeting with one of our biggest customers where we discuss how we can improve the product for them or introduce the product to new avenues or to a new channel - such as email.
My mission is to improve the daily work routines of our clients. Customer service is a very stressful job with high turnover, and I want that work to be as pleasant as possible. I do see further possibilities to support the agents and the customer service operation as a whole, and our automation and suggestion engines are the first steps towards that ultimate goal.
"We always use AI as a tool to help our customers instead of creating AI for the hype's sake."
Your company just received funding from the EU project Horizon 2020, which aims to support deep tech companies in Europe. How does ultimate.ai’s deep learning solution stand out?
The part that gives us a lot of competitive edge is language agnosticism; as many other companies are only focused on English. But I believe it's only one factor. I think, on a deeper level, we stand out as being very practical in our approach, as we always try to use AI as a tool to help our customers instead of creating AI for the hype's sake. As a smaller company, we need to use our resources efficiently and have a super tight focus on creating what brings the most value to our customers. This often leads to creating deceptively simplistic, but very effective solutions that have huge amounts of thought behind them. Bigger companies tend to have a much wider focus, which leads to them being average at everything - good at nothing.
Great AI needs great data: How do you ensure data quality while considering data security at the same time?
When we started, we realized very quickly it's easy to make a bot that responds to one or two questions such as "hi", but super hard when the customer has 50-100 different intents with overlapping questions. It's a lot of work maintaining all that data and content, and it would be easy to give that responsibility to our customers. However, one learning we had was that customers have a hard time maintaining the quality of the data. That's why we have built our customer success team with the knowledge for keeping data quality high, and a product that makes it easy for our customers to do the work. The same applies to data security. We take GDPR very seriously and have built-in tools to mask personal data out of messages coming to the bot.
"Don't try to solve every single part about human language, but enough to build a solution that creates value."
What is the biggest tech challenge for an AI startup from your point of view?
I can only speak about problems in the natural language processing field, but I think it's the sheer scope of the problem we are trying to solve. Human language is an obscenely complicated system filled with all sorts of semantics, logic, and context that philosophers and linguists have studied for hundreds of years and still makes the human mind go dizzy. We are trying to make an AI that understands such language with only a handful of people. This comes back to what I said before about focus: the challenge is to figure out how can we choose a focus for our AI, so that we don't try to solve every single part of human language, but enough to build a solution that creates value with such limited resources. Let's assume we suddenly hire hundreds of researchers. We could throw them into NLP problems and use those resources without necessarily getting any products out of it. That's why it is so important to know what aspects of your work you want to prioritize.
What tech advice would you tell yourself if you could travel back in time 3 years ago?
Maybe this sounds a bit weird, but I would tell myself: "just do what you are doing now and it'll work out." I think it's naivety that allows people to do great things because you're not paralyzed by the size of the problem you are solving. You keep solving things as they come, and those small things together turn into a big mountain you have climbed at the end. Another reason is, I am super stubborn about things until I see it for myself so I wouldn't even believe my own advice. I think that's part of being a startup founder, you don't let anyone tell you what to do.
Which source of information do you use to stay up to date on what’s going on in the tech world?
I keep my toolset of skills fresh by reading documentation and articles relating to building good web services, but generally, I don't keep up with the tech world. There is so much noise out there that it mostly ends up being distractive to my work.
Stay in the loop
Subscribe to our email newsletter. No spam, just occasional insight from our experts.