Watch Ronak Halani, our group COO, as he shares his insights on the ‘Applications of AI in Fintech’ at the FINTECH WEEK LONDON
Here we share a snapshot of the discussion.
The Application of AI
Presenter: Define what you see as the application of AI in Fintech for anyone new to the industry.
Ronak: I believe in the context of FinTech AI can be defined as follows:
Machine Learning Level 1 – This is where we use the RPA robots to observe patterns in user actions. This forms part of the process discovery where the robot is installed on a machine and is learning from human actions. This is how the machine learns how to mimic the user’s actions.
Machine Learning Level 2 – We feed a huge amount of relatively structured data and let the robot process it. This helps to identify patterns and follow pre-defined scenarios and process flows. Once the robot learns the patterns it will forever continue following the same path. Thus saving time and energy from human input.
Machine Learning Level 3 – Here a robot processes relatively unstructured data and identifies unknown patterns and scenarios that humans are usually not able to see without experience. At this stage, we can truly say that the machine is learning from its own past experiences rather than from humans.
Deep Learning Neural Network – There are plenty of theoretical examples where a machine could consciously make decisions. However, the practical uses of this technology are still unproven.
Agile Automations and FinTech
Presenter: What are you doing with AI that reduces cost and improves the customer’s experience?
Ronak: We design and develop bespoke Robotic Process Automation bots. These follow the same ML1, ML2 and ML3 (described above) for our clients in the FinTech sector.
This includes bots for:
- Transactions monitoring and AML
- Regulatory reporting and risk mitigation for FATCA and tax
- Accounting and bank reconciliations
- Procurement and ordering
- Customer onboarding, service, and relations management
- Lending and debt processing
Implementing AI in FinTech
Presenter: In your opinion what are the most prominent challenges?
Ronak: Challenges range from educating clients about the benefits of AI to more complex ones like cleansing data and choosing the right platform.
Other challenges include:
- Lack of awareness and education
- Lack of Agile Solutions that the business owners can try as a proof of concept
- Stigma and worry around AI, for example, “will it break?”, “will it replace workers?”, “will we all be out of jobs?”
- Legacy systems and data storage
- Lack of structured data to feed to the bots
However, by partnering with an end-to-end software consultancy like Agile Automations we have the right experience to deliver AI projects.
AI and regulations
Presenter: How well regulated is this area and what type of regulation is missing and why?
Ronak: Currently, the same set of regulations apply to applications of AI in FinTech, as any processing within FinTech. This is the same whether it is done by humans or by established software.
However, in recent years, we are seeing more structured regulations specifically designed for AI coming in.
Examples include:
- April 2021 – European Commission laying down “Harmonised rules on AI”
- In the UK there is currently no AI-specific legislation. This is because laws, by their very nature, have to be technology agnostic to ensure that future technology will still be subject to an overarching legal framework – Read an article by KPMG here
- UK Govt published a report in Feb 2020, laying down recommendations on the “Impact of AI on Public Life”
The important stuff
Presenter: What would be your key piece of advice when applying AI in Fintech?
Ronak: Businesses should proactively seek to implement and adopt bots with a trusted partner, such as Agile Automations.
This will make scoping and implementation much easier as we are proven experts in the sector.
Introducing Robotic Process Automation (RPA) carries a minimal investment for high ROI. We are always keen to take on new projects and provide proof of concepts.
We want to showcase the amazing benefits and versatility of custom-coded RPA!
Unsure where RPA can make a difference in your business? That’s ok, let us show you instead.
Briefly describe any manual process involving data below, along with a contact number, and we will build you a demonstration proof-of-concept project.