6 min read
The power of AI across industries (part 2 of 3)
In Part 1 of this series, we explored the evolution of AI from traditional roots to Generative AI (Gen-AI). Part 2 delves into AI’s impact on financial services and the public sector.
Financial services
Banking and insurance are set to be among the sectors most transformed by Generative AI worldwide. The key areas of impact include sales, operations, software engineering, and risk/legal/compliance. This mirrors what we’ve observed through our initial engagements and use cases across the financial services sector.
- Front-line experiences and digital concierges, often referred to as co-pilots, are game changers. Leading banks have already integrated advanced mobile app services, and it won’t be long before these evolve into conversational digital avatars that engage with you in your preferred language and style
- Customer personalisation in app and web channels is now more dynamic than ever, tailored to the style, language, and demographic of each individual. We’ve showcased next-gen state-of-the-art capabilities in making pictures speak, mimicking audio, generating video, and combining these elements to create unique experiences
- Product recommendation and upselling of new or enhanced services. We’ve partnered with a leading insurer to create a system that automates recommendations, like an Amazon shopping experience. Offering next likely actions is valuable, but doing so with greater intelligence and personalisation elevates the service’s effectiveness
- New product creation and strategising competitiveness is an interesting area we’ve been working on with a customer. AI can quickly and efficiently compare existing products against competitors in terms of features, legal terms, and benefits. Our early insights suggest that this speed advantage could lead to more dynamic pricing. Drawing on our experience with traditional projects, we’ve built pricing engines that can adjust in near real-time based on risk factors
- Automated process and claims handling or fulfilment. With foundational large models trained on vast amounts of data, the analysis of images and audio is more powerful and accurate than ever. A prime example is Lemonade, a disruptive insurer that leverages AI to automate core processes previously deemed impossible. This capability is essential for improving claims handling, underwriting processes, managing large volumes of data, including risk details, risk scoring, and standards comparisons
- Legal Terms and Support to accelerate M&A and due diligence. In investment and equity organisations, legal search and due diligence acceleration is a significant benefit area of Gen-AI. We’ve worked closely with one organisation to build tailored Small Language Models (SLMs also sometimes referred to as Vertical Models) trained on specific fields and document types to accelerate accurate search and summarisation
One final key aspect that has notably emerged is how to manage the regulation and risk of Gen-AI. We’ve helped a leading bank implement their responsible AI board and continue to manage this with them with a multi-faceted team of experts from across the business and our organisation. We look at each use case to determine compliance with security and ethical guidelines and policies of the bank. This is ever-changing with areas like the EU AI Act now coming and will be impacted in companies outside of Europe also.
Additionally, the risk of data leakage, or potential for IP loss, both being equal for some of our customers means we’ve had to develop fine-tuned foundation models hosted privately. These at scale also have the advantage of being great at narrow tasks tuned and trained to the specific elements we need but also much cheaper when deployed at scale with thousands of uses every day.
Public sector and government
There is real opportunity to prioritise and accelerate legacy modernisation through Gen-AI freeing up some legacy investment to fund the new. The process of continuous and evolutionary transformation, effectively repurposing costs. Within a year, making this cost neutral at a minimum, is a great approach to take as many of the Gen-AI projects are relatively fast and offer significant returns quickly.
For Governments, this opens opportunities to accelerate desired programs of work, particularly in areas that currently impose significant costs:
We’ve already proven that Gen-AI can address some of these challenges:
- The use of language in Citizen Services is key, with numerous nationalities and languages it’s now essential to provide services in multiple languages. Gen-AI can do this dynamically and is easy to adapt and build into solutions programmatically, so there is no excuse for not having in addition to English, good quality Welsh, Urdu, Chinese, or many other languages. Over time this will reach the option of every language on earth
- Accessibility is a significant beneficiary of Gen-AI, vital for ensuring that all individuals, whether they have a visual or audio impairment, neurodiversity, or dyslexia, can effectively engage with and utilise your systems. Our collaborations with organisations like the Dyslexia Association Ireland and the National Autism Society have resulted in tailored solutions integrated seamlessly into existing applications, simplifying the comprehension of language and complex text for users. Furthermore, we recognise the potential for regulated service providers to champion accessibility excellence within their processes, enhancing user experiences across various demographics
- Using Gen-AI for stylistic use of language is vital to Citizen Services. Imagine being able to seamlessly tailor communication to match the preferences and demographics of each individual. From formal and intellectual with a lexicon more appropriate to an elderly person, then with little adjustment in the application code itself (easy to maintain), apply the same engagement to a teenager digitally savvy who uses contemporary emoticons and phraseology. We’ve done this numerous times and it’s critical to work in areas like education and social care. A powerful example of this is in the summarisation or creation of notes for social workers or carers. The notes can be created and summarised for say a lawyer, a doctor, a child, a parent, or the social worker themselves using different styles but from the same core input, this would take hours or days previously per letter
- Harnessing summarisation and automatic creation of data is a key aspect holding the potential to significantly impact individuals’ outcomes. Gen-AI’s ability to transpose audio into text, summarise, make it searchable and give this to the care worker and system in near real-time is now a reality. This advancement transcends sectors—be it fire, police, social work, education, or health—ushering in a new era of care systems. With some of this data already being captured, it’s now a case of accelerating the integration and accessibility of this, reducing the overhead taking our carers away from front-line work all too often
We haven’t featured wider opportunities that stretch into the nascent and emergent fields such as video and audio generation but safe to say we’re already working on these multi-modal solutions with progressive customers and some departments. This is an exciting area that adds further to the list of opportunities above and some new use cases.
Conclusion
As we embrace this era of innovation, it’s imperative to navigate regulatory challenges and prioritise ethical considerations to ensure responsible deployment and maximise societal benefits. By harnessing Gen-AI’s potential conscientiously, we can pave the way for a future where technology empowers positive change across industries and communities alike.
If you’re interested in learning more from Brad about the transformative power of AI in the financial and public sectors, be sure to check out Version 1’s AI Webinar Series. Brad will be hosting two dedicated sessions focusing on the impact of AI in Financial Services and the Public Sector.
Stay tuned for the final instalment of this three-part series from Brad.
More from Brad
-
Blog
AI Safety Summit