Splendid Unlimited

A minefield. Well, not quite. But certainly very complex. There are 26 administrative regions, or Cantons in Switzerland, each with their own tax rules. And everyone in over the age of 16 MUST complete a tax return. The penalties are high. As is the cost of submission. – average around 1,500 CHF per submission. Unsurprisingly, the number of extension applied for each year is high, but of course then the fee charged by accountants to submit last minute returns are not for the faint-hearted. You CAN submit the return yourself, but of course it’s a very complicated process.

Our client Taxly was created with the bold aim of ‘Complete your tax return in under 10 minutes. Free of charge*’.

We began our engagement by understanding the problems. As well as the complexities of tax (yawn), there’s another, harder to counter problem. No-one WANTS to do their tax return. It has no upside. Other than no fines. So we also quickly uncovered a need to make compelling your tax return.. well,…. ‘Likeable’. If there can be such a thing. You can view the

So we took an approach of two parallel streams of work. The technical teams, business analyst, Swiss tax expert (client side) and experience designer set about unpicking the complex branching which exists in the process. You’ve got kids? Great. How many? How old? What benefits does each receive? How much does their education cost you?… This lead us to creating individual branches in each question, so the we could create a dynamic form – giving the ability to create completely bespoke journeys from the dizzying set of answers and branches. All of this meant that a customer really COULD complete the form in ten minutes. By only seeing what was appropriate to them.

A digital assistant asking simple, personalised questions, in just 5 minutes taxly’s AI TaxBot will digitally file and optimize an individual’s taxes 10x faster than any human tax advisor or traditional tax filing software.

There’s an added complexity. The 26 cantons have between them about 10 different form models. And we had to comply with each. Which meant balancing the processing of data on the front end (so we could be dynamic), and on the back end (so we could submit incrementally in accordance with canton rules). Once the flows were defined, we needed to look to UI, and figuring out how to ensure the user understands where they are in what is ultimately a fluid process. Delicacies around sub-categorisation of questions required a delicate, modular approach to what might at first appear as simple form fields and CTAs.

Underpinning this was the second stream of work – how to make the process ‘likeable’. As a startup, the brand themselves had no footprint, but that worked in our favour in that we were had no barriers.