12 months ago, armed with little more than an idea and the right combination of tech and business expertise across its four co-founders, AI startup Parlamind became one of the first startups to be accepted into the GTEC Lab – GTEC’s year-long startup support program. Last week, having completed the program, I visited Parlamind in their new offices and talked to founder and CEO, Chris Wolf, about where Parlamind have come in that time and their experience as GTEC Labsters.

IMG_6861

Founder and CEO, Chris Wolf, in Parlamind’s new office.

Where was Parlamind one year ago, and where are you now?

“Then, we had our idea – to build an AI-driven customer care system for ecommerce – and we had a core team of four committed co-founders who knew, between us, how to build this idea and how to build a tech business. When we moved into the GTEC Lab, we had just started the technical groundwork taking the technical concept and developing it into a product. (Check out our blog post when Parlamind moved into the GTEC Lab here.)

Now, 12 months on, we have a working beta product, about a dozen beta customers, a rapidly growing team of 13, plus we’ve just closed a round of seed-funding.”

It sounds like you’ve come a really long way. How was being in the GTEC Lab a part of that journey?

“To be completely honest, right at the very beginning of the GTEC Lab program, one of the best aspects for me as CEO was the free office space. Having a place where we could work together at zero cost removed an existential pressure many early-stage startups face. It freed us up to focus on developing the core idea without feeling the need to immediately raise money from investors.

I’ve founded startups in the past but my three co-founders are new to the startup world, having never founded nor even worked in a startup before. I know that for them being around other founders from other GTEC Lab startups and having that exchange of ideas both formally at events and informally over the coffee machine in the kitchen was really helpful.

Beyond that I think one of the biggest advantages for Parlamind was the range of people we met through the GTEC Lab – investors, partners, and potential customers – brought in by the GTEC team and its founding partners.”

IMG_6867

Parlamind team hard at work in their new office

How did the GTEC Lab program – 0%, 12 months, no-strings attached – compare to being in a standard accelerator or incubator?

“GTEC offers a fundamentally different model for startups. Accelerators and incubators invest in startups, take shares and expect something from them at the end of their fixed time – follow on funding at least. GTEC doesn’t. Its approach is “come when you’re ready, leave when you’re ready, use the program for what you need it for”. This allows it to take more of a risk with the startups it accepts onto the program, to look beyond proven business models at more diverse tech startups. Without a rigid timeframe or program for startups to fit into, GTEC creates a space for startups which are truly innovating – whether in business model or in technology.”

And, post-GTEC, what’s coming up for Parlamind?

“Riches, world domination …. [laughs]. Seriously, though, we have one more product iteration in beta planned. Once that’s through it’s all about launching the first product version and, of course, sales.

Beyond that, we want to think about other areas the product could be used in. For now we’re very happy with our ecommerce focus; it feels right. But we’ve had a lot of interest from SaaS, banking, energy, utilities, and telco. We’re going to use the product to manage our own customer care this summer and get a sense of how versatile the AI is without starting from the very beginning again.”

IMG_6865Parlamind was the first AI startup in the GTEC Lab and this is your first AI startup. Are there specific challenges AI startups face which other tech startups don’t?

“Good question. With non-AI products, such as a straightforward e-commerce platform, you write a concept, deduct sprints, and then you have a product working 90-95% as you want it to work – 99% after bug-fixing. With AI, it’s a lot less straightforward. After building something you have to continuously think how to make it better in a completely different way. It is less about following a plan and more about being super agile and thinking how you can “teach” this machine to do what you want it to do, rather like you would teach a human. And then you might have to accept that your machine is not the fastest learner – and that you have to teach it something more often and with more effort than you hoped for. I totally underestimated this aspect of AI at the beginning. Now I know it and definitely keep it in mind for the next development phases.

It’s very early days for commercial applications of AI. Most of the technology we’re using is fresh from academic research – it’s technology at a concept stage rather than a ready-to-use tool. Tools will come in AI, like they have in website development and ecommerce. You can see it starting already with chatbots, but it could well take five years until you have something like a WordPress for AI. What’s great is that we have a head start at Parlamind. We already have a great team of high-level experts who know how to build these tools themselves.”

Thanks Chris. It was a privilege to have Parlamind in the GTEC Lab and we’re excited to see what comes next.