Original article was published on Artificial Intelligence on Medium
Taxy.io: Making taxes human — chapter 1
The startup story — from idea to MVP
Two years ago we embarked on a journey without a clear path. We, four guys with computer science, engineering and management background, founded Taxy.io, the innovation leader of AI based automation solutions in tax advisory.
Since then, time has passed so much faster than in any year of my life before. Working for Taxy.io, i.e. living Taxy.io, feels like living on the fast lane. Our baby’s birthday these days shall be enough reason to pause a second and review what has happened so far.
In the beginning: Finding the team and develop the first concept
Do you like taxes? What do you know about all those rules, judgements, laws regarding taxes that affect your daily, private life as well as any professional transaction?
Taxes are ubiquitous. Everywhere, where money flows, tax issues are raised.
Simultaneously, nobody in this country knows everything about taxes. Even tax advisors are forced to constant researching and learning and I have deep respect for this cognitive challenge. However, at one point in our career, we were frustrated about how little we are able to understand about tax regulation and how cumbersome, expensive and prone to human error tax consulting can be.
“What if”, we founders prior to Taxy.io discussed via Whatsapp, “we can teach a software to “read” tax laws and and then just give us an answer based on our current situation?”
Asking a software seemed faster, more correct (a software doesn’t forget and “reads” much faster than a human) and (after development) cheaper than the 9 bn. USD tax consulting industry (GER only) indicates. And with a software taking over analytical tasks of tax advisory, tax advisors will have more freedom to consult their clients individually and proactively. Hence, AI would make tax advisory more human.
So we created the first concepts for Taxy.io, first a business model canvas, then a business plan, eventually a comprehensive paper that we used to apply for a governmental grant (EXIST startup grant). During that time we extensively spoke to and included feedback from practitioners, tax advisors, investors, accountants etc. to readjust our business model constantly according to real life needs.
In parallel to the acquisition of first funds we completed the founding Team Taxy.io. I’ve always been a fan of the 3-roles framework for startup founders (as described here).
In short it says that a successful founders’ team should cover 3 roles (not necessarily 3 people):
– someone who can sell (i.e. acquire clients)
– someone who can deliver (i.e. develop the software and bring it to clients)
– someone who looks after the money (i.e. acquire funding and manage costs)
In the beginning of 2018 SvenP, an exit-proven serial founder, (someone who can sell) and me, a former venture capitalist, (someone who looks after the money) completed our founding team with two friends we knew since many years through university: Steffen, a computer science grad plus MBA and former Harvard University researcher in machine learning, and SvenW, also computer science grad plus MBA and former manager at T-Systems for large platform as a service/cloud projects (both the deliverers).
After we were granted the EXIST grant by the Federal Ministry for Economic Affairs, everything was set: we had the concept, we had the team and we had the first funding.
And we had no excuses anymore to not “jump into the cold water” and start up.
From concept to MVP
During the first year we validated (or falsified) as many hypotheses as possible to create MVPs of our products — always our product vision in mind (link).
We pivoted quite quickly away from any B2C related plans and focused purely on B2B applications as we found less competition there, but more demand and higher complexity to solve.
After a couple of months we acquired the first paying customers for our NLP based “answer engine” on taxation issues. During that time our machine learnings were quite limited and so was the quality of our software. But thanks to the fascination of our clients in the possibilities of tax automation we could convince our first customers to stay with us until today.
I can’t stress enough how important it is for a starting company to find innovation-oriented customers to begin and grow with.
With a growing customer base we hired the first employees. Taxy.io now consists of ca. 20 people in 5 cities. Since we saved costs as much as possible we didn’t have a real office and worked from our kitchen tables in the beginning.
Retrospectively, it was a great training for remote working, which was still unusual in 2018 for a tech company, especially if you talked to VCs. Flash forward to 2020 and we are thankful to have made this experience because it allowed us to move the complete team to the “home office” within one day during the Corona pandemic without any technical problems. Disclaimer: We’re looking forward nevertheless to come back to our beloved office spaces in Aachen and Cologne, as soon as the crisis permits it, as a real team culture can hardly be fostered remotely (at least I don’t know how yet).
Develop a team culture in parallel to innovative software was also the reason to settle down in Aachen’s Digital Church by end of 2018, a co-working space in a former church. During the first year we were greatly supported by the Aachen ecosystem’s main players, e.g. DigitalHub and RWTH Centre for Entrepreneurship.
How do you create a fuzz without any marketing budget? We attended many startup competitions (because each competition comes with a free press release and other publications) and applied for corporate sponsoring programs (NVIDIA Inception program, Telekom TechBoost) to get our name out there and receive cash-like support (like discounted GPUs for our model trainings from NVIDIA or cloud IT infrastructure worth 100.000 € from Telekom)
In addition we checked out other foreign markets to incorporate expansions opportunities into our long term strategy as soon as possible. Especially USA seemed to make sense since it’s a large homogeneous market (at least language and business wise). Thats why we were very keen on investigating “the valley” first hand after being selected by the German-American chambers of commerce and the economic promotion department of North Rhine-Westphalia (NRW).
It became evident that after the first year of public funding our revenues will not be high enough to finance a team that is necessary to further develop our AI based automation solutions for tax firms. So end of cash was coming closer and we decided to go fundraising!
With early versions of our product in the market that proved the existence of demand, strong technical partnerships in the background, a growing, highly motivated and qualified team with a clear vision, we created suitable “traction” to qualify for a seed financing round.
This episode of Taxy.io’s story was also portrayed in the Wickeltisch Podcast episode 5 (in German). In the next chapter of our startup story I cover our seed financing round and how we won the first enterprise customers.