As Managing Director at Visma Resolve, Martin Sommerseth has a deep passion for AI. He loves enabling developers to solve complex and challenging problems for customers. When asked where he sees opportunities to innovate within the accounting industry, he had a lot of insights gained from practical experience.
“The accounting industry has great potential to be revolutionised. Accounting offices are transitioning more and more towards advisory services rather than regular accounting tasks, such as bookkeeping and reporting. As accounting software providers, this shift triggers us to drive the change and identify the best possible opportunities.”Martin Sommerseth
Here are four of his biggest learnings and advice for companies looking to make similar moves.
The following insights were written by Martin for our eBook, Building products that users and developers love.
Spend time in the problem space before entering the solution space
Problem space versus solution space is a well-known dilemma in product development. The problem space is where customer needs reside. It’s where you learn about the users and their problems so you can determine what your product needs to do. The solution space is where the product and product representations are in focus.
There are a lot of examples of product teams spending too much time in the solution space, leading to solutions that aren’t solving the actual customer problem. Product teams are usually encouraged to be problem solvers – to provide solutions. So, it’s not surprising that product development teams are drawn more to building products.
To succeed with innovation, it’s important to spend enough time in the problem space to ensure that you understand your customers’ problems. In-depth domain knowledge – where you can see both the pain points and the opportunities – is crucial when using your creativity in the solution space. Without domain knowledge, you can miss the solutions that create the most value for your customers.
Fail fast through prototyping
A pitfall when working with product development is not failing fast enough. Innovating and disrupting an industry is complex and requires focus and rapid customer feedback. The chance of finding a solution that fits the market on the first shot is low. So, multiple iterations are often necessary before finding meaningful and valuable solutions.
Create organisational backing through a business case
Investing in innovation and aiming to disrupt an industry includes a lot of risks and often significant investments. Creating a business case to ensure that the work pays off and that you understand how much customers are willing to pay for the innovation is crucial. In a business case, it’s important to answer questions like:
- What is the value proposition for the customers?
- What is the market size and likely penetration rate?
- How should the innovation be priced and packaged?
- What does it require to implement the innovation?
By answering these questions, your organisation and investors should know whether the product is something to prioritise or not. These questions have to be answered long before the actual product is ready for production so as to not slow the process down. Working closely, and in parallel, on the technical and commercial sides of innovation is paramount to moving quickly enough to ensure success.
When the innovation includes methods utilising artificial intelligence (AI) and ML, it adds technical complexity to the development of the solution. To lower the risk of making bad decisions and strengthen the business case, it’s a good idea to conduct a technical proof of concept. The purpose of the technical proof of concept is to:
- Validate whether the innovation is technically feasible.
- Receive more feedback from customers on concrete use cases by running user tests on their data.
Productionalise AI through control and explainability
When innovating through artificial intelligence (AI) technology, it’s essential to avoid making AI a complete black box for the users. By letting users feel in control of the AI and by explaining the AI’s reasoning and choices, the software can easier gain their trust. Users who cannot control the service will likely not find the product or feature valuable. Multiple examples exist of product teams being too ambitious in how much an AI should decide on the customers’ behalf, leading to a low adoption rate due to a lack of control and trust.
Want more details on his team’s approach, including specific steps and results? You can read all about it in our eBook!