Artificial Intelligence (AI) is already a big deal in the business world, with an increasing proportion of day-to-day processes and communications being assisted by machine learning and natural language processing (NLP).
And uptake of AI is bigger than you’d think. From tapping into a chatbot to ask a question, to receiving a “you’ve left something in your shopping bag” email, consumers are on the other end of AI processes all the time (often without knowing it).
Penetration of AI into customer services, and for the purposes of this article I mean Machine Learning and NLP, is vast. It’s expected that by 2020 80% of businesses plan to adopt AI as a customer service solution, and 85% of customer interactions will be handled without a computer.
So how can your SME use machine learning and NLP to achieve excellent customer services? It might not be as hard as you think….
A bit about machine learning and NLP
Machine learning is a subset of AI which allows systems to automatically learn from experience without being programmed by a human. The learning is then automatically applied to create an improved process. Machine learning relies on computers accessing data that they can learn and use for themselves.
NLP is another subset of AI and computer science concerned with the interaction between computer and human languages. It focuses on how computers can be programmed to process and analyze large amounts of natural language data.
Offer personalisation and recommendation
An area of machine learning that’s already supporting customer experience, is personalisation and recommendation.
Machine learning collects data about a customer’s previous browsing, geo-events, on-site interactions, purchases, referrals and other online behaviours. It then uses this data to determine a more personalised service in the future.
Personalised user experience is gradually becoming an expectation rather than the luxury consumers raved about in recent years. In fact according to Salesforce’s 2016 Connected Customer report, by 2020 57% of buyers will depend on companies knowing what they want before the first interaction.
This level of personalisation is pretty advanced, but there are ways in which your SME may already be using it.
Take, for example, email marketing. It’s likely that you’re already personalising salutations and subject lines. You may also have software to personalise the user experience on your website based on their location.
More complex personalisation involves the kind of machine learning that uses consumer shopping histories to recommend products or services, and also personally tailor customer support.
Look at areas of you would benefit from personalisation, and assess the products on the market. Issues to consider when shopping the market:
- What are your goals?
- Your ROI
- What issues are you trying to address?
- How tech-savvy are you?
- How complex is the software and does it provide decent customer support?
It’s estimated that 44% of consumers don’t realise they’re using AI. And it’s hardly surprising given how quickly it has seeped into our daily lives.
One example of this is how consumers use virtual assistants and chatbots as a go-to option for finding what they need online. According to Creative Strategies, 97% of mobile users are using AI-powered voice assistants, whether that be Siri on iPhone or Amazon Alexa in their homes or any of the other many and varied VA on the market.
Whilst shopping consumers expect to use AI, with 49% being willing to shop more frequently, and 34% spending more money when AI is present.
But in customer services, in particular, chatbots are also an expectation. According to Chatbots Magazine, 67% of people already expect to see or use messaging apps when talking to a business rather than picking up the phone or emailing directly. For 45% of consumers, chatbots are the prefered mode of customer service communication.
Google also offers pre-built NLP APIs which extract a person, place or thing from text, analyse sentiment to label messages with tags, all to help direct the right customer inquiry to the right support assistant.
So, customers don’t expect to speak directly to a human to find answers to their issues online. But chatbots create savings for your SME too.
Not only do you save time and resources manning phones or responding directly to each email, but you can benefit from:
- The ability to deal with high volumes of support tickets simultaneously
- Data insights into your customers behaviours
- The ability to deal with multilingual customers thanks to NLP
- Using automated support to free up agents to deal with more complex issues
- Multi-channel, 24/7 support
Start by doing a bit of research into the chatbot products on the market. Things to consider when you’re looking at solutions:
- Balancing your ROI
- Being able to dictate how, when, where and why a chatbot appears to your customer
- Does it enable multi-channel support?
- Does it include NLP and language support?
- Do you get multiple types of chatbots which take on multiple tasks?
Check out our recent article on how to use chatbots to improve your customer service. It might help you to look at the benefits (as well as some of the disadvantages) or investing in a chatbot platform.
Resolving product issues before they arise
Thanks to machine learning, it’s becoming easier for companies to avert the need for customer services in the first place. With the proliferation of smart devices, the IoT is allowing businesses to monitor their products after they’ve left the warehouse and have been delivered to the customer.
Take the automotive industry as an example. Connected vehicles are made with smart software which, once in the hands of the new owner, will send signals back to the manufacturer so that faults can be reported. The car manufacturer will then get in touch with the owner to let them know there’s an issue. While there are currently only around 64 million connected cars globally, this technology is expected to reach 1 in 5 vehicles by 2020.
But we’re already seeing this on a smaller scale in our homes, with some of our kitchen white goods using smart technology, or our security cameras, locks and thermostats.
While predictive customer service might be a long way off for many SMEs, there are some lessons to be learned.
Predictive customer services, based on data gleaned from your customers’ online behaviours, can determine the best course of action for a customer service ticket. This kind of response, known as “next best action” (NBA) is a way of using machine learning to present the right response to the right customer at the right time.
Look at the various options for Next Best Action software, and determine how predictive customer service might be of benefit in your SME.
Helping the customer along their buyer journey
A big part of customer services is done behind the scenes, making sure the customer experience is seamless and interconnected throughout their interaction with your brand.
Machine learning can help you to predict your customer journey and set up the right triggers and responses to deal with any potholes along the way.
An example of this is the classic cart abandonment email. With an average rate of around 69.23% abandon rate, around £18bn is lost in sales every year, a well-timed email can help to recover an estimated 5% of customers.
For your SME this could be the difference between thousands of pounds a year.
But there can be more subtle ways of prompting customers along their journey. Alerts and notifications can be triggered to remind consumers of items they’ve been browsing that are now back in stock, items they’ve left in their shopping carts, recent discounts, or shipping updates.
Obviously, consumers need to opt in to receive these messages, so you can’t rely on it to nudge all your customers, but it’s a relatively innocuous way of moving people towards a purchase.
Set up your workflows between your email marketing tool and your website and focus first on your cart abandonment email. Mailchimp has a designated section on it, demonstrating how integral it is to a successful customer service strategy.
Next, look at options for setting up notifications or alerts to keep your customers updated within their buyer journey.
Gather customer insights to improve the delivery/user experience
Machine learning is all about data, and whatever your strategy for using it for customer service, you’ll be left with data insights that you can then use to improve:
- How you deal with future customer support enquiries
- Your customer experience more broadly.
But there are external machine learning tools out there, such as Crayon, which gathers customer insights to help you improve your business processes.
Machine learning and NLP are worth watching in 2019, as they penetrate all areas of business and commerce.
Consumers are expecting more interaction with computers rather than humans, and are already using machine learning and NLP (often without realising) as they connect with businesses both on- and offline.
But companies are also seeing the rewards machine learning and NLP can bring. Not only can they bring savings both in man hours and in budget, but it can provide a more sophisticated and scalable level of support on a 24/7 basis.
The more machine learning processes businesses adopt, the greater the data insights they’ll be able to access, and then plough back into getting their customer experience right. If done well, machine learning can create a virtuous circle of customer satisfaction.
Having said this, the AI processes we’re talking about are still in their infancy, and SMEs can be forgiven for taking tentative steps towards adoption. The main takeaway is to start small, and being by incorporating elements of machine learning into your processes to help you cut costs and better achieve your goals.