Artificial intelligence spending in financial services is easy to count. Licences issued, tools deployed, training modules completed: all of it produces a number, and all of it can be reported upwards as progress.
What that spending has changed is the harder question, and it is the one the contributed piece below is concerned with. Its argument is that the measurement gap is not a reporting inconvenience but a source of risk in its own right, and that learning and development teams, rather than IT, are the ones who have to close it.
Harry Chapman-Walker is chief executive of Kallidus, a learning and talent management platform used by regulated and industrial organisations. The article that follows sets out his opinion.
Financial services firms globally are operating amid economic and geopolitical upheaval, evolving regulatory expectations and market disruption from waves of digital-native start- ups. In response, they must innovate, change and adapt; all eyes are on IT and L&D teams to ramp up investments in and usage of new technologies, especially AI. Yet increased adoption does not automatically result in increased value.
Before asking how to innovate, it is vital to ask why. What is the desired outcome? How will the investment improve customer outcomes, strengthen operational resilience, or help meet regulatory responsibilities? Only when the firm has a clear idea of what it is trying to achieve, and clearly defined measures of success, can the appropriate L&D training, pathways and solutions be put in place.
Outcomes matter
Global insecurity is forcing businesses to rethink operations and accelerate change. But the outcomes are far from clear. In the rush to innovate and mitigate risk in a volatile operating environment, financial services firms risk failing to match automation and technology investment with desired business goals. Even worse, where those business goals are not clearly defined and communicated in the first place, the value of the investment can be undermined and create confusion throughout the workforce.
In short, firms are falling into ‘the activity trap’. Coined by Kallidus, this is where businesses measure effort, uptake and visible action rather than meaningful progress. When AI adoption is judged by the number of tools deployed, licences activated or training sessions completed, businesses risk mistaking momentum for impact. Activity may create the impression of change, but unless it is tied to defined outcomes, it does little to reduce risk, improve performance or build the capability needed for long-term transformation.
The pace at which firms are looking to scale AI adoption is a case in point. The plethora of AI tools available is exciting, but AI is no silver bullet. It is not the outcome a business wants to achieve, but rather a tool to achieve that outcome faster or more efficiently. Businesses therefore need to clarify the expected value. Will the investment reduce risk by helping to detect fraud earlier? Will it boost staff retention or deliver quantifiable bottom line improvements? Unless the business can measure specific, predefined outcomes, AI, or any other innovation, has added nothing.
Well, not nothing. The evidence is clear that firms risk adding rather than reducing risk through a lack of outcomes-led investment. Incidents of damage to both bottom line and reputation due to AI misuse are growing and visible. In addition, the hidden dangers created by inconsistent adoption and understanding are quietly undermining business value and creating very real business risk.
Driving adoption through usage targets or incentives will backfire if employees are not confident or do not understand what they are being asked to achieve and why. The FCA‘s recent Mills Review illustrates this clearly in retail financial services. As AI moves from recommending actions to taking actions autonomously, it says people’s roles will increasingly shift towards setting boundaries, granting permissions and overseeing outputs produced by AI on their behalf. That transition requires more than familiarity with the technology; an
effective human in the loop must have the skills and confidence to challenge AI and intervene when needed.
A new mindset
Innovation is essential and AI is without any doubt a fundamental business tool. So how can firms embrace an experimental mindset and achieve change while also minimising risk and, critically, ensuring compliance with fast-evolving global regulations?
Companies require a completely new outlook. Every business problem, operating model and approach must be reconsidered under the lens of AI, and vice versa. If an investment in AI is not aligned with key business pressures, it is not only failing to add business value; it risks becoming an expensive distraction. Whether the priority is improving performance, enhancing accountability or mitigating operational risk, expanding automation and scaling up AI should only be considered with a tangible return on investment goal.
An outcomes-led approach is key to ensuring the use of AI is aligned with the company’s core values. One of those fundamental values will be compliance with the changing regulatory landscape, from the EU AI Act to the UK’s AI regulatory principles and ISO 42001, alongside relevant rules and guidance like the Vulnerability Guidance and Consumer Duty. This understanding is required not just by the chief information officer or the IT team. Every single knowledge worker, from the board onwards, needs to understand the risks and
rewards of AI, as well as the importance of mapping AI adoption to business goals. Ensuring AI ethics and governance are embedded throughout the company will mitigate risk. It will explain the importance of balancing AI and critical thinking, and reinforce the importance of handling sensitive customer and commercial data to avoid any exposure.
It must also establish what meaningful human oversight looks like at the different levels of AI autonomy, and who is ultimately accountable if something goes wrong. Framed this way, compliance is not about training delivery. It is about making sure people are genuinely capable and ready to use AI responsibly and with a common goal.
The learning and development shift
Company-wide AI ethics and compliance training will also provide a platform for the essential outcomes-led discussions between departments and teams, if the right mindset is in place. And that is a big if. Business transformation is tough. It demands attitudes and skillsets that are often completely different to those previously required. For individuals with a fixed mindset who are content to do things the same way indefinitely, albeit extremely well, AI is unsettling and uncomfortable. The pace of change is overwhelming and, understandably, many individuals do not want to engage.
This shift is hitting L&D teams hard. How does an individual who has calmly and carefully curated course after course for many years suddenly segue to a completely new model, one predicated on responding to business outcomes? How do they evolve from measuring course attendance to tracking business performance gains? And how do they partner with the executive team to support this new paradigm?
Being honest, not everyone will have the skills or the desire. Change is not for everyone. Many highly intelligent and motivated individuals will feel uncomfortable in this new world and move on. But that is also fundamental to successful transition. To effectively reconsider every aspect of business operation, and determine how, where and when AI can be applied, requires a new mindset from every single person.
For L&D, it represents a shift in focus that professionals have been grappling with for a while. Many are stuck in the activity trap, but the move from measuring activity to measuring business performance is now more urgent than ever. In the case of AI, where does it improve performance within the firm and what is required to deliver that change? Who are the individuals struggling to achieve this shift, and can they be supported to unlearn what they have done in the past and relearn what they need for the future?
In an uncertain operating environment, decisions can be rushed and mistakes made. A change of approach is essential, and L&D should be the driving force. Defining and prioritising outcomes shifts the narrative. It enables firms to move on from an unfounded perception that AI and automation are the solution, and to build a strong foundation for effective change.
The opportunity for financial services is significant. In retail financial services, the Mills Review points to better customer support, earlier detection of fraud and better firm governance. But it also warns that AI could amplify fraud and introduce new systemic risks. Whether AI ultimately improves or undermines outcomes will depend on whether employees are equipped to use the technology effectively and to keep solid oversight of what it is doing and producing.
Aligning both technology adoption and usage strategy with clear business goals enables firms to define and deliver the required behavioural change. It ensures individuals across the entire business understand objectives and have the knowledge required to embrace new ways of working. It provides a foundation for measuring the impact of innovation, allowing companies to rapidly scale the most successful outcomes to drive additional value, and to drop any that fail. With the right outcomes-led approach, firms can both achieve positive change and accelerate success.
The post AI in Financial Services: Activity is Not an Outcome appeared first on The Fintech Times.
