The secret to successful AI
Chatbots and virtual assistants are transforming the way financial services organisations engage and interact with their customers. Open Banking has catalysed the development of new, sophisticated stand-alone Artificial Intelligence (AI) solutions and the
advent of ‘voice’ operated services has led to the increasing use of ‘conversational commerce’. A growing number of us now want and expect the convenience of a ‘chatbot’ – available 24/7 to answer queries and carry out even complex tasks - and we aren’t impressed
when the bank or insurance company can’t deliver the goods.
Many leading banks (RBS Group, Citi, Bank of America, Santander et al) have already delivered impressive examples of this exciting new type of AI. Automated voice services are rapidly becoming the norm with banks and other financial institutions looking
to deliver an intuitive and frictionless service that avoids the need to call the customer services department to resolve queries.
But – as with so much emerging technology – simply introducing a chatbot or virtual assistant is just the start of the significant journey to realising its benefits. Ensuring that a plan is in place so that your newly developed chatbot matures over time
incorporating fresh features and rich functionality to service and delight customers is more challenging than it seems to many at first, but the rewards can be significant and long lasting.
What, therefore, is the secret to chatbot success? It starts with taking the time to truly understand what your customers actually want, need and expect and not simply what your leadership may believe they want. Being rigorous about improvement will ensure
your chatbot stays topical, relevant and genuinely useful. The ideal chatbot interaction is easy, ensuring they deliver is not.
Success demands a relentless focus on customer needs, specialist CX (customer experience) skills, lots (and lots) of data as well as the time and mental energy to analyse and act on that data. It also requires a long-term strategy that enables the chatbot
to prove its worth and ROI through constant improvement over time. The most successful organisations view chatbots as a long-term investment in future success rather than a short term ‘quick win’ to replace skilled customer service teams with a robot!
During my career, I’ve focused on using innovative technology to optimise customer facing solutions. Working with Amazon and RBS on both the Alexa and Cora programmes respectively, I’ve had a first-hand view of how new AI technologies can deliver a vastly
improved customer services as well as an innovative product development strategy.
Here are a few insights I’ve picked up along the way that help to make AI integration as successful as possible.
Think CX (not IT)
Many banks which struggle to get their AI projects up and running effectively are the ones who are too-quickly frustrated at the time and costs involved. Why? It’s because their IT departments lack the skills and knowledge to nurture the project beyond its
first steps. Too many financial services organisations approach AI as an IT project with a set delivery point. Business leaders ask their existing IT department to investigate, design and launch a chatbot or virtual assistant without a comprehensive plan for
post-launch improvement – and that is the root of future dissatisfaction.
Developing a chatbot is NOT an IT project – or at least, not an IT project alone. In my experience, the best results are achieved when a chatbot is conceived from the start as a CX project. New technologies demand new ways of thinking and working as well
as different problem-solving approaches to address ‘bumps in the road’. Successful chatbots are not about IT – but about customers - what they want, need and think – and how they want to interact with you. It’s vital to bring CX experts and others into all
AI or Chatbot discussions as early as possible to ensure the right approach and mindset that defines what you want to achieve and the experience you are seeking from users. Launching a chatbot is less about how to reduce operations costs, but a chance to truly
explore ‘the art of the possible’.
Appoint a diverse team
Many organisations expect their existing staff to take on board new innovative technologies without considering the impact this can have on day-to-day operations, or what skills or mindset may need to be added to their base for the project to be a success.
Progressive ways of working often require a totally different skill and mind set. The most successful teams delivering complex AI solutions tend to be what I call ‘left brain/right brain’ confident. This means they understand how to interpret customer data
as well as how to turn it into actionable results using a data led design approach.
Using innovative technologies means thinking and acting differently – and with it comes new responsibilities. It’s vital to consider ethics and ways of behaving when looking to integrate AI. There’s huge potential for unconscious bias to be built into the
logic of a chatbot, disadvantaging certain groups of customers whilst only meeting the needs of small select groups of individuals. Business leaders should be mindful about the data strategies they employ when programming the chatbot to ensure they are being
open and inclusive.
One of the best ways round this is to ensure diversity, in both the design and development teams to ensure that the chatbot really does treat all customers equally and doesn’t favour certain ‘types’ over another. Your AI / Chatbot team needs to be made up
of subject matter experts (SMEs) in a broad variety of areas: Conversational designers, UX professionals, MI analysts, data scientist, technical support, customer journey managers, customer service managers, along with a support team to focus on Legal requirements,
risk, product specialists etc. So, appoint an inclusive and diverse team of ‘all the talents’ and invite your CX, SME and IT specialists to assemble a unified ‘chatbot’ development team – which sits down together, shares ideas and learns from each other to
create a unified and seamless strategy of constant improvement – delivered through the combined expertise on offer.
Dream big but start small
AI empowered chatbots – the latest generation of a user-oriented technology approach - remain an exciting new development area for many financial service organisations. It all seems great! Easy and quick to use – instant, frictionless ‘24/7’ service – and
– for many the true prize – a chance to slash ongoing costs. However, once launched, there is a tendency to rush to overburden the system with too many requests, often without considering whether it’s equipped to handle them. A clear rationale is cost – with
business owners wanting to create efficiencies and ‘do more with less’.
However, experience shows that taking a realistic and pragmatic approach to chatbot deployment pays dividends. When Amazon launched Alexa in 2014, ‘she’ operated within a narrow, highly specific arena – answering a limited number of questions – and learning
from the queries that she was asked. Amazon took its time examining the data it collected, developing a rich understanding of the vast and hugely differing requirements (and interaction preferences) of its customers. The company then took this on board before
it meticulously began to expand the services that Alexa could perform to help her evolve into the multi-functional and highly skilled ‘virtual’ assistant that exists today in millions of houses. The lesson of Alexa is for businesses to consider taking a minimal
viable product (MVP) approach to chatbot development. Try to resist rushing to a solution that attempts too much too soon. Let your AI learn to do a few things well before it attempts to become all things to all people.
Don’t abandon your legacy
Even in an article as enthusiastic about the potential for AI-enabled customer experience as this is – it’s worth reminding ourselves that inflated estimates and excessive excitement about the prospective gains from disruptive technology has led to much
disappointment, among customers as well as at board level. It’s important to set realistic targets for the solutions we introduce and in no way to neglect those legacy systems that have underpinned best practice for years. Treating an AI solution as an excuse
to eliminate the skills, intelligence and flexibility of your customer service function is a fast route to disappointment as well as additional costs which can outweigh any potential gains. Far better to take a harmonised approach, where existing and emerging
techniques co-exist together to inform and improve each other.
Not all your customers will welcome a chatbot; some will love it, others might have to be persuaded or ‘pushed’ to use it, others will always prefer a ‘human conversation’ and will resist all attempts at engagement. Remember, customer conversations are not
always linear – and even with specialist natural language recognition technology - your chatbot might be challenged to answer all queries. Rather than saving time, talking to a chatbot might lead to conversational dead ends or cul-de-sacs – delivering a frustrating
customer experience. If the customer experience matters the most, any AI chatbot should be designed to route to the best help possible as quickly as possible – even if that takes it back to a human representative.
Instead – expect your customers to use a variety of contact methods – making the chatbot part of your communication strategy along with messaging, social, email, video and voice. Let them know that chatbots will not replace valued and highly experienced
customer service agents - but will simply be another way to get in touch – and make that clear right from the start. Your chatbot can then develop over time to become the essential ‘go to’ service that you want and need it to be.
Expect greater contact volumes
Almost all business owners expect that a chatbot will reduce the burden on the customer services team – creating almost instant efficiencies and cost savings. However, in almost all cases, the opposite is true. The easier you make it for customers to engage
with you – the more they will want to do it. This isn’t a bug – it’s a feature!
Opening up a totally new engagement channels doesn’t always mean that conversations will be diverted from one channel to another – instead it means a greater number of conversations will take place.
Increasing the amount of ‘incoming traffic’ and making existing customer services teams work even harder may seem, at best, a sideways step for a business looking to embrace a more efficient future. Not so. Sometimes a chatbot shows you just how many customer
conversations you’ve been missing.
Customers often start out asking one question, but very quickly move on to asking complex multiple questions. With the seamlessness and always-on availability of the chatbot service, customers tend to ask questions more frequently, with raised expectations,
and become frustrated when they don’t get the answer they need. Instead of seeing this as a failure of the chatbot – CEOs should see this increase in traffic and additional data as a major advantage. After all, more customer data allows smart business leaders
to refine and improve the way they design and deliver services to customers.
Focus on the data – and ‘feed’ your CX team
Successful organisations are great at launching new solutions– but all too often, the data they generate with these solutions is left to decay in vast databases. What sets Google, Apple and Amazon apart from their competitors is that they constantly gather,
analyse and use data to improve and enhance their services. This has been a key component of their success as market leaders – and the keystone of their reputation for amazing customer experience.
An expert CX team is passionate about data – and want access to as much as possible. As well as constantly analysing all available data, your CX team will ensure formal governance and data privacy rights are taken into account, so you can be confident you’re
compliant with the very latest regulations.
So, start to get serious about the data you gather. Empower your CX team to collect and analyse data and ensure that CX-led design informs your AI/chatbot initiative from the start.
Look to the future
Change is the new normal – and AI technology is developing at lightning speed. New advances in Cloud services, machine learning, data uses and computational power, combined with the shifting social use of services, rising customer expectations as well as
the unending growth of digital means that any AI solution needs constant and ongoing evaluation to stay relevant and topical. It’s critical to make ‘horizon scanning’ the core element of your AI strategy to ensure you’re embracing the best and most exciting
technologies. Think of how innovative contactless payments were a few
years ago; now they’re mainstream and part of everyday banking. What’s the next ‘big thing’? Who are the companies to watch?
Ensure you’re close to what’s happening in the world of AI technology – or work with a partner that has their ear to the ground and can keep you informed and up to date.