Call us: 0800 3167566
THOUGHT LEADERSHIP GUIDE
HOW TO UTILISE AI IN A FINANCIAL SERVICES CONTACT CENTRE
ASK AI – WHY OPUS?
Revolutionising customer experience with AI in financial services
Would it be fair to say that, for a long time, the financial services, insurance, and legal industries have been somewhat risk-averse to adopting modern technologies and AI?
This cautious approach can be attributed to the industry’s regulatory environment, the need for security, and the potential risks associated with handling sensitive financial data. We understand. Having worked with clients like First Central, one of the biggest motor insurance providers in the UK, we’ve seen first hand the transformation a modern contact centre can have on a business in an industry traditionally resistant to change.
This also means we’ve seen the reasons behind this resistance, the processes you must adhere to, and the day-to-day challenges you face and we’ve overcome them. As businesses like yours have advanced into cloud technology and virtual environments, it’s been proven time and again that it is possible for highly regulated and tightly governed industries to not just keep up but lead the way in technology.
Artificial Intelligence (AI), in contact centres and anywhere else, is no different. While scepticism exists among various stakeholders both within and outside of the financial services sector, this should not hinder your progress or success. Throughout this guide, we’re going to introduce AI-based or AI-assisted contact centre components that companies just like yours are benefitting from today. Along the way, our goal is to ease the pain and reduce the tension surrounding AI in a business-to-business (B2B) capacity.
The current view of AI in financial services
There’s a major difference between consumer-facing AI and that of AI in the business world. This stems from the governance applied by the vendors tasked with making that AI fit for purpose and safe for business. Take Salesforce Einstein, for example. As the first comprehensive AI for the world’s largest CRM provider, simply letting AI loose on your businesses wasn’t an option. Instead, Salesforce, a company valued at around £188bn and one that serves more than 150,000 customers, drip fed AI into its portfolio following considerable testing, quality control, and training. This is what makes the difference between consumer AI and B2B AI. In reality, when working with businesses like First Central, One Family and Hood Group we wouldn’t have been able to deploy any of the technologies they currently use in their contact centres if B2B technology reflected the level of security and privacy of consumer technology. Thankfully, the processes and tools built around base level functionality means businesses in these controlled environments can thrive using modern day technology and not get left behind. 
The use case for AI in financial services contact centres
If financial services organisations and insurance brokers are already using AI in their contact centre, what are they using it for? Here are four major use cases we see often when firms like yours wish to revolutionise their customer service operations.
1 – Conversational IVR
Conversational IVR for incoming calls You’re probably familiar with the term IVR. It stands for interactive voice response, which usually implies one-way conversation. Callers input their reason for calling via keypad presses on their phone or by saying the menu number out loud. More modern IVRs introduce the possibility for callers to say their reason for calling.
This is now basic technology. Instead, IVRs can now have two-way conversations with your customers. Using pre-selected answers alongside AI and machine learning, callers can say their reason for calling and get help directly from your IVR.
This is only a basic example and there may be other steps involved like identity and verification (ID&V) and key presses to enter payment details but what’s happening here is your IVR is having a conversation with your customer, meaning your human agents are free to handle more urgent or emotional queries.
This ensures that not only this particular customer, but all of your customers, benefit from reduced wait times and quicker response times. Your IVR is having an open natural dialogue with your customer, crossreferencing other systems (billing, CRM, sales, etc.) to recall data and information. The conversation is driven by the customer’s input and responses. Your IVR is simply responding to the input.
What if a customer needs to speak to a human?
We always advocate for configuring an option to escalate the call to a human. In fact, if the conversation is turning sour, you can use sentiment analysis to detect this and make it happen automatically.
Conversational IVR for web chat
IVR also extends outside of incoming calls. What about if you take the premise of managing inbound calls through automated and AI responses and apply it to web chat? Yes, it’s a chatbot but one with more intelligence than you’re used to.
The natural language understanding (NLU) tools behind conversational IVR mean that chatbots can now have two-way conversations, pick up on emotion and urgency, and provide more than yes/no answers and route customers to agents.
You can apply traditional question and answer logic and route web visitors to agents, but you can also update account information, process payments, and retrieve product information too.
Save human time on requests that don’t need human input. Benefit from natural language processing and conversational IVR to automate routine conversations with your customers.
2 – Identification and verification (ID&V)
Sticking with the theme of automation and efficiency, how many times per day do your agents ask customers to prove who they are?
“Can you confirm your account number, address, security question, and mother’s maiden name please?”
What if you could outsource the entire ID&V process to your contact centre technology? By building security questions and integrating your contact centre with CRM or other identity systems, humans no longer need to lose time asking the same questions all day long. In fact, your verification processes become more secure as you remove the potential for human error. There’s no agent cross-checking different systems or misinterpreting security answers. When left to machines, it’s black and white. Either your caller gets the answer right or wrong. No unauthorised staff can access accounts.
On occasions where you do wish to keep humans as the ID&V processors, lean on AI prompts to ensure everything gets checked. Leaving security, payment reminders, and compliance questions to memory doesn’t make sense when you have the option of flashing prompts on-screen to make sure agents don’t skip vital steps.
3 – Speech analytics
As conversations take shape, between AI or humans and your customers, there’s always potential for things to go sour. It might be a timing thing, a misunderstanding, or a genuine mistake. Whatever the scenario, customers get frustrated and conversations can get heated. Using speech analytics, you can better understand and assess agent performance with certain customers and query types. Armed with this information, you can identify knowledge gaps that inform agent training. Whether you need to improve error rate, call handling time, or attrition levels, the reports at your disposal mean you have reliable data showing you areas for improvement.
Sentiment analysis introduces real-time insight into what’s happening during live calls. With AI transcribing your calls in the background, automatic detection occurs when key words or raised voices happen. If a customer swears or becomes angry, supervisors receive notifications to suggest they intervene. Similarly, when positive language is detected, such as terms like “upgrade” or “renew”, triggers for cross-selling opportunities and contract information are activated, providing relevant prompts to the agent.
4 – Agent Assist
When agents first start out, they go through a process of observing a more senior agent. Next up, the roles swap and you let your new agent handle some calls with the senior agent sat by their side. When the new agent needs help, they have someone beside them to offer support and ensure they provide the correct information. Obviously, this isn’t a scalable solution for every agent all the time.
But what if it was? Agent Assist means that every agent can have the help of a machine plugged into every aspect of your business that has learned from all your customer interactions. Each time an agent fails to mention your current promotion, for example, they get a prompt on-screen. Or, if a customer has asked about discounts but your agent has forgotten what it is, it automatically appears so there’s no need to route around different systems or put your customer on hold to ask a colleague.
Agent Assist, you can build out entire libraries of prompts, scripts, and notifications. These can be triggered by any number of words, phrases, or rules. Off the back of each usage, AI tracks outcomes like successful upsells, retained customers and SLA adherence.
Gaining Trust in AI in Financial Services
There is no expectation that you switch on your new contact centre and send everybody home. In fact, we suggest you adopt these AI-friendly tools and processes before, during, and after your initial contact centre implementation.
Generative AI Studio
Instead of letting AI into the wild and saying whatever it likes, spend a few weeks/months in a sandbox environment putting AI to the test. You can create and deliver personalised, relevant interactions by tailoring prompts to specific use cases and using contextual customer data. Rather than using an off-the-shelf AI product, you get the chance to choose exactly what happens when. GenAI Studio also addresses organisational concerns around responsible AI with its ability to control the data that generative AI can access and share.
It’s low code by nature, meaning you don’t need constant help from IT or developers. Instead, contact centre managers and supervisors can program their desired outcomes for different scenarios. You get automated quality management, data testing, and monitoring of real-world situations so you ensure AI is treating your customers exactly as you wish.
Machine learning
They say what goes in reflects the quality of what comes out. That is, unless your AI is controlled and learning all the time. For example, just because an angry customer makes a demand, it doesn’t mean they get an angry response or AI buckles to that demand. Instead, backed by data and customer interactions, your contact centre AI recognises situations must be diffused and a mutual agreement reached. This happens over time. The more information you feed into your AI, and the more approvals and rejections you make, the smarter your personalised AI contact centre becomes.Exploring AI Applications: Who’s Using It and How?
Finding the right fit for AI in your business may take some time. There may be a period where you allow your conversational IVR to handle all calls from specific clients but reserve human agents for VIP callers.
This process could take days and then you find the perfect fit. In reality, you’re going to want to review how much AI and who gets it on a regular basis. We suggest this is often during your first implementation. The more you check, the more confidence you have in AI.
Over time, you can reduce the need to quality check AI’s work and automate this process too. One day, there might be a time where it’s completely autonomous but for now, it’s best to take a trial and error approach when it comes to drip-feeding AI elements into your business.
Third-party audits
Data and privacy are paramount in any organisation but there will always be bias inside your business.
One team, who advocated AI technology, may be too lenient. The other, who were cautious with their approach, may be too strict.
Instead, use a third-party auditor to regularly (annually) audit your use of customer data and information to ensure there is no change to your adherence since your adoption of AI. This, of course, is optional. For those more sceptical, however, this proves a vital step to the acceptance and adoption of widespread AI efficiencies.
Who’s using AI in their contact centre today?
First Central is one of the biggest motor insurance providers in the UK. Armed with 900 staff and receiving 3.75 million minutes of phone calls inbound every month, its legacy phone system was nearly a decade old and couldn’t integrate with any of the new strategies it was building. First Central needed the data continuity and agility that comes with having an omni-channel contact centre. It had to empower its employees and better engage its customers through their channels of choice.
As one of many financial, insurance, and legal businesses using AI to make their contact centre more efficient and provide better customer experiences, First Central undertook a number of discovery workshops and requirement gathering exercises to unearth the objectives of their contact centre refresh. Armed with this information, Opus put together the right blend of AI and human-style contact centre for First Central to gain exposure to the potential of AI.
As a result, First Central now has a cloud contact centre that:
- Keeps customers coming back
- Demonstrates innovation to customers and employees
- Is flexible in terms of the number of staff
- Reports on human and AI interactions
- Has a simplified call routing plan
- Removes unnecessary agent admin tasks
How to get started with AI in financial services without losing control
You can’t just turn on AI and hope for the best.
When tasked with bringing efficiencies to your contact centre, you need a trusted partner who’s been there and done it. While AI may still be a new technology, it’s something Opus have been implementing in our client contact centres for years.
Through our deployments for finance brands like One Family, Courtiers, and First Central, we’ve learned what it takes to future proof while remaining bulletproof.
Want to take your first steps into contact centre AI? Book a call with one of our implementation experts on 080 0047 3537
Download this useful guide now as a pdf
Discover the current industry views on AI within FinServ, the use cases for AI and take a look at which companies are successfully using it and how AI is delivering value.