Dealing with digitisation means getting the basics right in order to allow the organisation to drive forward customer service by embracing AI and machine learning and thrive in a rapidly changing competitive landscape.

Financial Services providers are working in an ever more complex and faster changing world. Since 2008 the British Banking Association calculates that there have been over 80 separate pieces of legislation and significant rule changes for banks and other financial institutions to deal with.

This is a challenge on its own, but regulation is also varying more between different geographies.

Recent months have seen moves in the US to repeal or at least roll back two key pieces of banking regulation – the Volker Rule and parts of the Dodd-Frank Act which will mean big changes for bank IT systems.

Britain’s departure from the European Union will bring further challenges for British financial institutions.

While traditional banking systems have focussed on reliability, compliance and the ability to scale, they now have to deal with creating systems which can be built and changed at a much faster pace than before.

Alongside these challenges, IT departments must also deal with ever growing volumes of data and customers demanding improved multi-channel service provision.

Where once this just meant phone banking, customers today want mobile apps, intelligent agents they can ‘talk’ to via chat or text and payment using wearable devices.

The IT department must also protect the organisation against well organised cyber attackers who are already embracing automation and artificial intelligence to break into networks.

Within the organisation developers are demanding more resources to push the business forward and require support for new applications and a shift to constant updates and iterations of key systems rather than traditional waterfall approaches.

Financial institutions are also facing more challenging competitors than ever before both from start-ups and from established organisations leveraging their brand into the market.

Building from the ground up

But in order to satisfy all these varied demands you cannot not just bolt on new systems or functions to existing infrastructure. It actually requires getting the basics right first.

Ensuring that the core infrastructure is right will allow the organisation to act flexibly whatever the demands from customers and business units.

In order to take advantage of new technologies like big data and intelligent decision making you need to get the foundations in place first.

Making sure that data is available to applications across the organisation means breaking down silos which might be reinforced by existing infrastructure.

But it is important not to forget the human side in all of this.

Tearing down silos is not just about ensuring technical compatibility. It is also likely to bring together disparate groups of people who have not previously seen themselves as co-operating colleagues. Dealing with these issues are just as important as dealing with the technology challenges.

Once you’ve created a truly flexible and agile infrastructure, in terms of both technology and people, then you are ready to deal with whatever the future throws at you. Many projects fail because they go straight to the end point – new mobile applications, big data analysis, artificial intelligence or improved and automated customer relationship software – without getting the foundations right first.

By ensuring core systems are ‘match fit’ it becomes far easier to bolt on extra and new functionality.

In fact you will have created an environment of constant innovation and the ability to change on the fly. Because the reality of digital transformation is that it is not a single process with a beginning and an end. Rather it is a constant process of change.

Financial services rely on providing the best customer service to differentiate themselves from the competition, that’s why they are at the forefront of new digital services like automation, chatbots and new ways of communicating with customers.

Using AI in the real world

IBM has helped the Royal Bank of Scotland improve services by creating an intelligent agent to help deal with customer queries.

Key to getting this right was the agile development process the partners adopted. The agent, Cora, began with quite simple aims – beginning with just 60 tasks or intents. But by using machine learning the assistant is getting better every day and learning from every single interaction with a customer.

Naresh Vyas, head of solutions at RBS, said: “you don’t program a chatbot as much as you teach her–it’s really a new kind of systems development practice, where iterative learning shifts the development paradigm.”

By keeping a sharp focus on performance IBM and RBS can measure Cora’s success and fine tune her abilities.

The team constantly check Cora’s ‘containment rate’ – the percentage of calls which are successfully concluded with no help from a human agent. Cora can currently conclude up to 40 per cent of calls without any human assistance. For corporate banking this figure is an impressive 80 per cent. The agent can now perform over 200 tasks compared to 60 at launch.

Core relies on agile development processes but the speed of adoption is staggering. At the time of writing Cora is dealing with about 100,000 calls per month but that is expected to rise to two million by the end of the year. This makes call centre staff far more efficient and frees them up to deal with complex and value-add services and calls.

Cora next stage of development will see the service extended to other platforms like Amazon’s Alexa.

The bank has also created a whole new profession to help fuel Cora’s development. RBS employs human agents, known as ‘conversational analysts’, who constantly monitor Cora’s interactions with customers to help push constant, iterative improvements in how she listens to, and talks to, customers. By 2020 RBS aims to be using Cora to have natural, meaningful conversations with customers on any platform they choose.

Changing technology means changing processes

The best technology projects can radically change how an organisation deals with business processes. The example of Cora shows that not only is RBS radically changing the way it deals with a growing percentage of customer queries.

Thanks to Cora RBS has accelerated how fast it can deal with queries while also reducing the cost. But the system is also providing a wealth of new information on what customers want which can push forward adoption of intelligence-led decision making for future strategy. More intimate knowledge of your customers can allow you to start predicting what they’ll want in the future, not just analysing how well you’ve performed in the past. Automating time consuming processes can completely change your organisation’s cost base while also freeing up staff to focus on high-value, personalised customer interactions. Of’ course all of this is more than just an IT project.

It needs buy-in and support from the board downwards and an acceptance that it will continue to change how the organisation and its people, processes and technology work together.

Bharat Bhushan, CTO, Banking and Financial Markets, at IBM, said: “Low interest rates, sluggish revenue growth and some of the recent regulatory changes in Europe have meant that financial institutions need to be efficient. They are driving efficiencies through automation and digitisation whilst keeping up with the pace of technology and customer expectations that are often set by other digital platforms they use.

“Depending on the organisation, digital transformation programmes can include broad ranging topics and technologies such as developing apps to content and CRM systems to APIs and Integration Strategy to Data governance and experimentation with AI.

Regardless of the detail, institutions are adapting open standards to avoid vendor lock-in, devOps to deliver to the business with agility and Design Thinking to understand exactly what user needs to deliver. And, during the process consider reusability – a bit like creating Lego bricks that other project teams from other departments can use to create new journeys in a relatively shorter time.”

Bhushan said that one of the changes in customer behaviour is that despite GDPR, the end customer will be open to sharing personal data if

(a) they get value in return

(b) the organisation establishes trust by being transparent about the purpose and process of using the data and

(c) their data is delivering some social good.

Bhushan also believes that it is a myth that only millennials want digital and mobile apps: “People are fundamentally the same. Every customer, whether they are 69 or 18 – they want to use digital services that are easy and simple, and they can use them whenever and however they want. Then there will be times when they would want to see or talk to a human expert; for example, retirement financial planning, mortgages or in exceptional situations such as fraud.”

Where challenges remain:

Financial institutions remain are inherently risk-averse and cautious, and their customers would not want it any other way. Bhushan observes that the majority of financial institutions tend to be fast followers but not first-adopters of cutting-edge technologies. He also sees a slowness around adopting AI and machine learning because such moves are often reported, wrongly, as meaning job losses. The reality is that such projects are often about creating new types of jobs.

Testing:

Adopting a more agile approach does not do away with the need for testing.

– Just because something works on a developer’s laptop does not mean it will work at scale, in a data centre and a live environment.

– Equally parts of a service can work perfectly but not work in the real world where they must all work together, if one component is creating exceptions or bottlenecks it can stop the whole system.

– Banks could adopt the processes some technology firms follow of offering a selected cohort of customers the chance to use a new beta feature of the service and automatically collect feedback as they use them to improve for wider rollout or, completely scrapping the feature.

The future

Looking forward Bhushan believes the majority of the incumbent banks: “will become truly digital organisations which happen to also have a banking license. They will create and participate in platforms and, ecosystems – allowing you to interact with them indirectly or, based on your data, recommend all sorts of products and services to you. For example, the best broadband or, entertainment package that might save some money without compromises. Ultimately, making everyone slightly better off financially whilst giving you nudges, hints and friendly coaching along the way to help you better understand your financial world.”

Footnotes:

BBA quote is from a BBA briefing: Reforms since the Financial Crisis.

The complete list is available below:

View list

Cora’s extension to Alexa is from here:

View extension

Financial services have additional burdens in making this work. They must prove their systems meet an ever changing, and ever growing, set of regulations covering everything from data protection to insider trading. They must ensure all activity fits within risk management profiles.

The cloud promised to end these problems. It promised endless, instant scalability at the touch of a button at lower cost than scaling up your own data centre. But the cloud’s dirty little secret is that there is a ceiling to this scalability both in terms of what is actually possible and at which point it becomes far cheaper to run your own private cloud instead of relying on a third party.

Even if the scalability is physically possible it often comes with a set of management headaches which make it effectively unworkable in practise.

That is why the majority of the big vendors are backing the hybrid cloud rather than one uniform environment. This allows organisations to exploit the benefits of different platforms while avoiding the pitfalls. Running your own cloud in-house guarantees information security for developers and let’s them optimise hardware without interference. Equally keeping infrastructure applications running on your own kit guarantees availability where demand is fixed or predictable.

But the public cloud can still be used for new releases and products where demand is harder to quantify and it can scale up to deal with peak demands. And it can do all this on a ‘pay-as-you-go’ basis without draining capital expenditure.

But building such a platform without it becoming a management nightmare of incompatible systems needs expert help.

Building a scalable cloud-based infrastructure does more than guarantee service levels for existing applications and business services.

It allows the organisation to be ready to exploit emerging technologies like Artificial Intelligence and machine learning which are increasingly available as cloud-based services rather than on-premises solutions.

This gives the obvious advantages of cost and scalability but also means that systems can be built which can exchange information and lessons learned from peer companies facing and solving similar challenges. Of course competition is not going away but collaboration is a better way to fix technical problems than attempting to re-invent the wheel. For financial services the challenge is not creating new services it is in scaling up to provide those services fast to everyone who wants to access them. The challenge from start-up fintech companies is that they have the potential to scale without being encumbered with legacy technology portfolios.

But despite this scaling up still proves an insurmountable barrier for many financial start-ups.

Arguably the single most disruptive fintech phenomenon of the last ten years, Bitcoin, is suffering from its own scalability crisis. The digital currency is struggling to increase the number of transactions per second which the network can cope with and arguments about possible solutions are ongoing.

For anyone building a more conventional scalable platform in the next few years it will likely be a hybrid infrastructure.

The secret is getting the glue right – a hybrid infrastructure is not just a random, mixed bag of platforms. It needs to be built on shared standards so that it can interoperate seamlessly and applications and services can moved around with the minimum of pain.

Finding the right partner to provide that glue is just as important as choosing the right providers for the actual platforms.