With sophisticated technology teams, comparatively large budgets and mountains of data, banks and insurance companies might seem prime territory for implementing AI.
But doing so successfully depends on customers and suppliers understanding which challenges and processes will benefit from the technology and how AI can work in concert with human understanding.
This does not necessarily mean developing investment strategies or fuelling high speed trading with AI.
For example, financial organisations are bound by regulations and associated processes that can cause delays and reduce revenue. A bank onboarding a new client means meeting “know your customer requirements,” which in turn sparks multiple streams of work around risk and identity.
While insurance firms rely on careful analysis of data to establish risk, that data is often siloed across multiple legacy systems. This again causes delays, but it also makes it harder for businesses to fully engage with customers across their entire product line.
The agentic approach
These are all areas where agentic AI comes into play, says John McNamara, director of solutions engineering at Salesforce, with “autonomous agents doing discrete pieces of work.”
Those agents can collaborate with other agents, both within the Salesforce ecosystem or with agents from SAP or Oracle systems, or with core banking systems.
In the case of customer onboarding, for example, multiple agents could be set in motion in parallel, checking credit rating agencies, analysing a customer’s financial transactions or establishing if they are exposed politically.
In insurance, agents can operate across those disparate legacy applications and tap into the associated data siloes.
As Salesforce’s senior solution leader for insurance Katie Sutton says, this means data becomes the oil that smoothly links various applications and teams together. This can speed up decisions on claims, meaning happier customers. But it also gives businesses a more complete view of their customer, driving proactive engagement and helping them grow business with individual clients.
Humans in the loop
But agency does not mean complete autonomy. Many of those processes still need humans in the loop to make final decisions.
That’s why Salesforces views “agents” as part of a digital workforce, working alongside and collaborating with humans, who are in turn freed up to focus on higher value work.
For example, says McNamara, in the case of know your customer, Salesforce’s Agent AI layer “is smart enough to know when to defer to a human and when to make its own decision.”
That’s a result of the agentic layer built using Salesforce’s Agentforce platform being tailored to the individual client.
Use case ideation is key, says Sutton. This pinpoints where the business is currently spending time on manual tasks, what problems agentic AI can solve and how it can deliver return on investment. This process also takes into account the specific regulatory and compliance regimes each client works under.
Different players might have different starting points and ambitions. Fintech firms might be keen to get started quickly, explains McNamara, while traditional banks and insurance firms might take a more cautious approach.
But, says Sutton, the ideation process typically uncovers multiple relevant use cases – which gives the organisation a roadmap to work through.
That means with the most successful engagements, once the first use case has gone live it’s a snowball effect, she says. “You start with one, and they just continue to build and build.”
Conclusion
The dynamic world of AI is delivering fast-paced change that can reimagine business operations and uncover new ways of generating growth. Businesses must carefully identify where they can take advantage, or they risk being left behind.