Guest Post: Fridai – the first AI-powered gamer assistant

Today we have a guest post from Mark Englehardt, CEO & co-founder at Fridai. Fridai is an AI-powered voice assistant for gamers. Here he tells us about the history of the company and where they are today.

How we got started

We were three gamers getting together to build something truly awesome for other gamers - could be the title of how the idea came to life. We were passionate about games and knew a thing or two about technology – recognizing our own pain and find a solution for that was the first step.

User interviews

We conducted 119 interviews, to get a statistically good sense of what the biggest pains for gamers are. Most of these revolved around having to do too many things manually, switching screens or alt-tabbing out of games, as well as the distractions that can be flooding in.

We knew this could be solved by a gamer assistant, powered by voice technology so we embarked on a journey to start building a company around the product.

User interviews

Prototype building

After the interviews, we wanted to start testing with users as fast as possible, to get new ideas and feedback. We sent out a prototype to early adopters and people we did interviews with. Mostly we learned was that a more stable solution would be required to get to the scalable version of the assistant and that we needed to support games for guides and even tutorials. The retention rate for the first prototype was very low, but we started seeing a clearer path towards acquisition and features necessary to get better retention.

Getting into Axel Springer & Porsche Accelerator

After the prototype was a good learning experience, we decided to take everything further. We started applying to incubation programs, aimed specifically at idea-stage startups. After two months of applying to various institutions, Axel Springer & Porsche’s accelerator, APX, found us and after several interviews they sent us a term sheet.

It included an investment for 5% of our company and, more importantly, it gave us the opportunity to build a better version and start building out our company.

So, we left our jobs to start building this cool product we knew gamers and game publishers needed.

Getting into Axel Springer & Porsche Accelerator

Building our MVE - Minimum Viable Experiment (Essentially a learning vehicle, never meant to scale) - publicly referred to as an MVP

We fired up an early sign-up program to capture the attention of possible users and build a community around the product itself - while generating new ideas, having constant third-party opinion and involve users in development. The sign-up page included a little competition, too. We announced that only the first 300 users would get early access and the more friends you invited the further up you would go on the list.

Gamers started signing up by the dozens. We added 100 gamers to our early access list per day, organically, to the point where we had to close it down because we couldn’t support any more and had to announce the winners. But we managed to build a great community on Discord that turned out to be the biggest advantage of the early sign-up program.

During development, even though we could do some part of development we hired an external team to speed up the process. Our vetting process however was poor and we ended up with a team that lacked the skills and knowledge to build a product of this scale. After many failed attempts to release a working MVP, we decided to say goodbye to this team. They had slowed down our development not only in terms of how the product couldn’t live up to user expectations, but as a company we could not raise the next round of funding without having at least something presentable and in users’ hands. This was a major lesson for a growing company: even if you have the engineering mindset and skills internally, when outsourcing, vet the team properly with example tasks, ask for demos and use a product management system that gives you full transparency. We learned this the hard way.

We couldn’t risk any more delays, so we reached out to individual developers whom we had worked before and assembled a crack team of pros, who delivered on everything we asked them to. Later on, these people became our first hires.

Launch day

Our launch did not go as smoothly as planned, as we had to release with fewer features. We were second product of the month on Product Hunt and in the following months we increased our userbase to 4,000 monthly active users who performed 2 million interactions or 2.3 game queries per minute – making our launch a real success.

It was fantastic to see the numbers coming in, with only a tiny investment in marketing.

Launch day

User feedback - data collection

Users gave us one clear feedback after our MVP and it still rings in our ears. “You guys need to be where we are.” This meant, we either build on top of Discord or figure out how to improve on our existing Windows application to make it a standalone voice assistant that’s much more than that, a gamer assistant.

Based on user feedback we decided on a Windows-exclusive version which we released in March, 2020.

Working through COVID19

A pandemic is never good for business. However, from our perspective as it increased the amount of time gamers spent gaming, it was the right time to release our public beta for Windows. Usage has started picking up and we received more detailed and specific feedback from our users, as everybody seemed to have the time to put into helping us improve on the product. The biggest change we have experienced during the pandemic was that game publishers started reaching out to us, asking how we could work with them to create more revenue for their games – using Fridai as a built-in sales rep – and target gamers with sponsored content based on their behavior. This had been our consideration for a time already, so once we signed our first pilot deal we knew it was the right way to go – and we shall see what the future holds.