In the last two years, Harmoney has implemented key innovations in income assessment and credit automation, requiring a move towards an event driven microservice-based architecture and aligning its delivery teams.

Previously we had a traditional loan application process which asked users to enter their income and liability information so Harmoney could assess affordability, says Harmoney CTO Andrew Bates.

Invariably customers would enter some incorrect information, filing pre-tax or net income, or forgetting some liabilities, for example. That incorrect information required costly, time-consuming additional validation and manual intervention.

Utilising the customer’s credit file they could pre-populate their liabilities into the application process, avoiding wrong or missing information. This meant they could provide a richer experience that included third party logos. 

This helps increase borrower confidence that we have accurately assessed their information, says Bates.

He explains that by utilising customers’ bank statement data - which they were requesting but not automatically processing - they had enough information to implement an income prediction model. This vastly improves automation and reduces customer errors.

“We had to establish trust so a higher percentage of customers would use the automation, and we had to manage ethics around using this data to improve customer experience, whilst maintaining regulatory obligations,” he says.

He says the move towards a microservice-style architecture - incorporating technologies such as Apache Kafka (real-time data pipelines) supporting traditional batch and real-time processing of data - means they have made optimal use of available data, analysed and processed it in real-time driving insight and a much more dynamic, richer customer experience.

The driver for this was Harmoney’s goal to scale the business with technology and not headcount.

To improve customer conversion rates, improved automation of the credit decision process was necessary, he adds.

Our automated approach is market-leading, using data the banks have, but still do not use as part of their own processes, says Bates.

“As a result of this automation, we can provide a richer customer experience and, crucially, a faster customer experience, with loan assessments reduced to minutes.”

This journey has required innovation in the technology and processes used, and the organisational structures for the engineering teams at Harmoney, says Bates.

Previously, the engineering teams tended to be project oriented. But as Harmoney increased their automation and developed the platform, they saw the need for product aligned teams.

Thus, their teams are now built around their value chain (domain driven design) with monoliths split into microservices, and each team owning their own services. 

“We introduced DevOps, improved accountability and ownership and deployed self-service tools to manage infrastructure, deployments and releases,” says Bates. “We also improved our monitoring and observability including introducing metric driven development.”

Harmoney also recently implemented chatbots across finance and engineering. 

“To implement all these innovations, we have had to introduce new technologies around real-time data management, event-based architecture, machine learning, analytics, and time series data, he explains.

Startup to successful business

As head of technology in a fintech business, Bates says his role starts with setting the architectural vision for the company and backing that up by ensuring the executive understands the complexities and challenges of what they are trying to achieve. 

“I drive the conversations with the executive and the engineering team, on new technologies and key trends with financial services and fintech, to ensure our vision is relevant, progressive and achievable,” he states.

He also championed organisational changes for their teams to follow an agile organisational structure.

He got the support of the executive team at Harmoney to join the design council of a public sector health organisation so he can gain greater exposure to corporate governance.

He is also the proxy Harmoney representative on the Fintech NZ executive council and sits on several customer advisor boards of their partner vendors, and helps them set their future roadmaps.

Harmoney had just celebrated its fifth year and has grown from a startup to a strong successful business, says Bates.

“We have also continuously examined our organisational culture,” he states.

As CTO, he works with the head of people and culture and resource managers to maximise and align the capabilities of the business with their engineering teams.

By aligning their engineering teams around their value chain, the metrics that matter to the wider business at each step in the chain are owned by the team, creating buy-in and accountability, says Bates.

Operating a DevOps model and Self-Service/ChatOps services reduces hand-offs around releases, production support and infrastructure management. 

“Monitoring Driven Development ensures we have the right metrics in place before we start new features, and as our teams support what they build they can visibly see success via the metrics they own,” says Bates.

“My stepping back into a mentoring and support role, has created a ‘safe’ work environment in which individuals feel they can share any idea and the team increasingly learns from each other, rather than just from the top down.”