AI changes how we build software in financial services

It’s exciting to see organisations move beyond thinking “chatbots” when they hear GenAI. While the tech has advanced quickly in the last year, the design implications — how to sympathetically, meaningfully and safely integrate GenAI into digital experiences — still often pose a challenge.

This is especially true in heavily regulated industries like financial services. However, early examples of GenAI-enabled experiences provide a fascinating window into how the technology can be used. For example, we love Cleo as a demonstration of how AI can introduce personality, personalisation and contextualisation into personal financial planning.

Take Cleo’s “roast mode”; when enabled, it will shame you with GIFs and harsh burns about how much you’ve spent and how little you’ve saved in the last few days, weeks, months and years, hopefully helping millions of people avoid their overdraft, build credit, and budget better. Roast mode isn’t for everyone — but if it were, it wouldn’t be personalisation!

Building an intelligent product like Cleo is a giant leap in designing, building and operating software. It takes a fresh approach to strategy, design, technology and testing — not to mention rethinking organisational practices, processes and culture. Ask yourself: would your organisation be prepared to release a mobile app knowing it has the freedom and autonomy to say almost anything it wants to its users?

To understand this shift, its implications and why making this bold, brave move is worth it, we must reflect on what’s gone before…

Real-time human intelligence (pre-digital era)

Before software, businesses depended upon real-time human intelligence to provide their services. For example, your friendly, local bank manager would know you, listen to you, empathise with you, and work with you to get mutually agreeable terms to fund that innovative new business idea you’d been nurturing. However, this human-intensive approach took a lot of work, effort and investment to scale.

Design-time human intelligence (digital era)

In the 1950s, human intelligence shifted from real-time to design-time as BAs, designers, developers, testers, and others predefined complicated rulesets into computer-executable instructions. Software brought automation, reduced costs, and increased scale and reach at a price. It wasn’t adaptive; it couldn’t understand a user’s context, vary its execution, make exceptions, or adapt and learn. Our software-based products have no intelligence; they’re just complicated rulesets.

Run-time digital intelligence (intelligence era)

We’re now moving into what I call run-time digital intelligence. By sympathetically integrating intelligence into software, AI enables us to design and build experiences that behave more like humans than traditional software. Experiences that can “understand” the nuance of a user’s natural language, problem-solve, learn, adapt and interact in ways that suggest empathy and emotional intelligence.

Rethinking roles for the AI era

So, what must change for us to build intelligent software?

• Leaders must rethink the role of digital products for their users. Think about building long-term relationships rather than grabbing attention. Provide users with expert support rather than just handy utilities.
• Analysts must rethink how they define solutions—foster interactions that are natural and intuitive for the individual rather than forcing a predetermined path. Use enabling constraints rather than governing constraints.
• Designers must rethink interfaces intended to constrain user input. Instead, promote freedom, agency and the discovery of possibilities beyond predefined options and journeys.
• Testers and approvers must rethink how we test, review and build confidence in whether these solutions perform satisfactorily. We must move beyond a deterministic “if this, then that” mindset.
• We must all move towards rapid, low-cost experimentation that allows us to derisk AI-enabled innovation and improves our learning rate. That’s where the giant leaps will occur!

Conclusion

It’s becoming clear that we are not just witnessing an AI trend. Instead, we’re facing a profound evolution of how digital technologies enhance our lives. Early examples like Cleo show how AI can infuse our digital products with personality, personalisation, and contextual understanding. It offers a glimpse into a future where technology amplifies our human qualities rather than diminishes them.

However, the journey towards AI-enabled businesses requires more than technology; it demands a fundamental shift in our organisational mindset and culture. It challenges us to reimagine roles, from leadership to product teams, and to foster an environment where innovation is not just encouraged but is a cornerstone of our businesses.

At Waracle, we’re acutely aware of the responsibility and the opportunity this represents. Our mission to build the next generation of intelligent software is not just about enhancing our clients’ efficiency or profitability; it’s about creating digital experiences that resonate deeply with users, understand their needs, and speak to them meaningfully and impactfully.