Artificial intelligence may sound like a phrase from a Sci-Fi film, but this is a technology that is becoming more accessible to almost every industry.
The advancement of AI technology has allowed us to talk to our smart phones, log into our banking apps with facial recognition, and communicate with interactive voice responses when we call a helpline. Moreso, this integration of AI into our lives will only penetrate deeper, especially in industries that have direct contact with the end-customer.
It’s important to note that AI is an incredibly broad term, representing three key fundamentals: process automation (the ability to repeat common processes), cognitive insight (the ability to pattern match and surface them for inspection), and cognitive engagement (the ability to take those patterns and act on them). All of these are necessary to understand language patterns and respond to a customer’s needs, though they must be considered separate to reduce confusion and clarify their use cases.
What is Robotic Process Automation?
Robotic Process Automation is the best introduction to automation for most end-user businesses and perhaps the least intelligent. It’s the cheapest and simplest to deploy, making it far more accessible and quickest to provide ROI, thanks to the repeatability of the solution. However, there are still challenges.
A common RPA scenario is the integration of front-office and back-office systems – ensuring integrations don’t muddle data, confusing what data is considered the absolute version to be synced with other systems via integration. Known as the ‘source of truth’, this should be a massive consideration in any RPA solution.
In complex RPA setups, there is no single source of truth. Only the source of truth that is relevant at that time, or for that particular field, to make jobs such as responding to a customer question easier, quicker, and more accurate.
Looking more closely at our target market – SMB contact centres – AI has massive potential, not only in RPA, but in cognitive insight and engagement. On average contact centre agents can deal with 200 to 300 interactions per day – that’s a LOT of data that needs to be interrogated for pattern analysis. So, there are clear opportunities around cognitive insight, enabling AI to surface relevant knowledge assets for agents while they are mid-conversation.
Using Automatic Speech Recognition
In voice-based customer interactions, before any analysis can be done to support agents, AI needs to understand what’s being said. Hence the need for ASR (Automatic Speech Recognition) and Speech-To-Text (STT). But it’s not just about knowing what’s said – it’s understanding how many people are talking, disregarding background noise, and only then can it identify who is talking!
Once that’s understood, the AI needs to be able to identify what is being said and that’s no easy task with language, local accents and dialects all being critical in its accuracy. Finally, once you have an accurate understanding of who is saying what, we then need AI to turn that into text. And all of this in real-time too!
You may also assume that this is easily overcome by finding an expert in ASR and partnering with them, but where the data (voices) has come from to train the models will impact their ability to perform specific use cases.
Training data on US accents will produce less accurate results when the model is used on UK’s multicultural accents and dialects. Therefore, it’s important to find a partner that understands the underlying ASR APIs and efficiently customises or trains them to the specific needs.
Finding an AI Partner
When looking for a partner, resellers need to know the end-user’s problem and then match the tech to a partner that has been proven to overcome that challenge. Not overpromising to the customer is vital too – consumers are starting to believe anything is possible with AI, but the technology is still maturing. So, make sure you ask any potential partners about their roadmap and vision for the future and consider whether that lines up with your customers’ expected challenges in the future.
The future is full of possibilities for truly amazing customer service through the proper use of AI and the connectedness of the things we use in our daily lives. For example, consider an Electric Vehicle – when the batteries start to degrade and hold less charge, an IoT sensor can pick that up, and send a notification to the contact centre. The agent can then call the customer to book in the replacement or inform them, and check-in to ensure they’re satisfied. It’s also a great way to reach out for any upselling opportunities.
Once that blend of IoT, AI and customer service is realised, the opportunities for individual businesses to differentiate themselves through slick, seamless interactions and exceptional customer service will be endless.
