Low-code/ no-code combined with OpenAI GPT-3 and the likes
The killer combo in finserv is already here
The buy-in for Advanced Natural Language AI models, like OpenAI GPT-3, has recently spiked. Even though the acronym GPT is being mocked by the French and most people don’t care what it actually stands for, the excitement is increasing around the huge potential from collaborations with OpenAI.
GPT3 = Generative Pre-trained Transformer 3
Even for all of us focused on Financial services, a traditionally less open sector, increasingly regulated and complex, and with legacy Saas systems, leveraging Advanced Natural Language AI models, these are very exciting times.
Today I will zoom into one of the many areas that are extremely promising, ready to explore, and not so visible as they are behind the scenes. The power and impact of Low code/No code in combination with Advanced Natural Language AI models in financial services has not yet been appreciated enough.
We live a world that is demanding increasingly more Digitalization, delivered faster, and in an unstoppable fashion. We got a glimpse of this kind of pressure during COVID when financial services providers had to come up with ways to facilitate the delivery of the $953-billion business loan program of the Paycheck Protection Program (PPP) as soon as possible to the people and businesses. EY used the low code Microsoft Power suite to create a Platform for both lenders and borrowers by combining EY`s and Microsoft domain expertise.
This kind of demand is ever more pressing because of the macro-economic conditions that require doubling up on any kind of efficiencies you can imagine. IT departments can’t meet these needs for more than one reasons. IT departments have limited human resources (and declining with layoffs), priority management issues result in conflicts, and Business needs and specs get often lost in translation with IT people.
Low Code/No code services combined with advanced Natural Language AI models, are coming to the rescue.
Low Code/No code services are already enabling enterprises to bring to market, apps and software in just a few weeks, and in several industries already they are empowering an increasing community of Citizen Developers. Citizen Developers are businesspeople that typically have little or no code experience and are using low code and no-code tools to build applications.
IDC claims that more than 500 million apps will be developed by 2023. Gartner reports that already in 2020, about 25% of new apps were developed by enterprises using low-code or no-code resources. This is expected to Triple — to 75% — for large enterprises by 2015 with both software developers and Citizen developers becoming users of these increasingly innovative tools.
Some of these low code and no-code tools are based on Drag & Drop actions and others are based on AI wizardry via text or voice. Think of the Excel and Lotus which were the first generation of low code, no code applications.
Now fast forward to today, and Microsoft has integrated several open-source programing language capabilities into their Power Apps ecosystem. Last year they integrated Power Fx an open-source programming language for low code and this year they have integrated OpenAI GPT-3. Power Fx and AI, enable “programing by example” making easy for business people to customize interfaces and dashboards according to their needs.
Microsoft`s goal is to democratize app development for everyone. Whilst anyone familiar with Excel and Power Point can get going with Power Apps and Power Fx, the integrations with OpenAI GPT-3, are reducing the time of learning, searching, and debugging complex formulas. As a result, the efficiency gains for are suddenly skyrocketed.
Rabobank — Conversational AI use case
Microsoft`s Power Virtual Agents is a concrete example of low code development combined with powerful AI, for both pro developers and for Citizen developers. Rabobank announced this November how they leveraged the Power Virtual Agents capabilities towards their strategic mission in Conversational Banking. Rabobank already had chatbots and an interactive voice response system (IVS is voice and or touch key activation for phone customer service), which is not uncommon in banking. In just one quarter, Rabobank revamped its Virtual agent solution using Microsoft`s Power Virtual Agents low code & AI capabilities.
Rabobank`s Virtual Agents journey
First and foremost, Rabobank integrated their Chatbot and their IVS. Customers still have both choices to interact (phone or chat) but behind the scenes they are integrated and training the advanced Natural Languages models. The virtual agents are increasingly capable of directing the customers to the right agent faster and even completing tasks without the need of a human agent. For example, the virtual agent can raise withdrawal caps without the need to connect to an agent. Within the first 2 months of the Virtual Agent deployment, there was a solid increases of Rabobank customers using the chatbot as their first point of contact (from 20% to 25%) and more importantly, 40%-50% of the business contact phone calls were handled entirely through automation.
Rabobank foresees that the improvement of its virtual agents will lead to 30% of customers using the virtual agents as their first contact point which means processing successfully 12 million conversations per year. The next step is to increase the virtual agent`s ability to serve more customer needs, for example to change the address of a business in the system without needing a human agent.
The Rabobank example is not only about the speed to market but also the continuous improvements in customer service (AI training) and the ability to serve more customer needs.
A large Canadian Bank — A unified customer view for Financial Advisors
Another great example of low code that empowers businesspeople across a large incumbent organization is that of a large Canadian bank that is empowering its financial advisors with a unified view of the customer. A dashboard with smart integrations of 13 keystone and legacy apps across different business divisions, was delivered by Avanade, a Microsoft partner, in less than 3 months.
A snapshot of the dashboard speaks for itself.
I say `Smart` integrations because these connections are done via APIs and critical data remains in the existing systems of record leading to rapid implementation, adherence to governance frameworks, and massive improvements in workflow and therefore the Advisor’s ability to serve the customer.
Low code tools integrated with Advanced Natural Language AI models, are gaining adoption. They empower agile/ scrum business teams developers within companies and professional coders to serve the business needs.
The use of low code/ no code & AI tools can enable quick wins and empower employees to serve increasing customer needs more efficiently.
In this macro-economic environment, customer behavior and needs are changing and requests around servicing loans, credit and payment issues are on the rise. Customization of existing software, quick integrations to create customized dashboards to serve clients better and more wholistically, is paramount.
Doing more, better with less human and technical resources is the future, and the future is here.
Low-code/no-code tools combined with Advanced Natural Language AI models, is a killer combination.
I am a Microsoft Ambassador and this is a Microsoft-sponsored post. I would like to thank Gil Brodnitz, CTO, US Financial Services and Industry Advisor for the valuable discussions we had on this topic. Opinions are my own.
📌 Twitter: https://twitter.com/efipm
📌 Subscribe to my YouTube Channel with my insights and industry leader interviews. New video every Wednesday: https://www.youtube.com/EfiPylarinou
📌 Spotify Podcasts. Follow here: https://open.spotify.com/show/5bRkZEYHSwPiGx7vTqylw6?si=Mg3hN5PDQ86K10GjeK52jw
📌 Linkedin: https://www.linkedin.com/in/efipylarinou/
📌 TikTok: https://www.tiktok.com/@efiglobal
📌 Web: https://efipylarinou.com/