I’ve just run my second briefing on intelligent assistance. Much happened in the few months between sessions. This time, the second half of the day was dominated by stories about bots and their use cases on messaging platforms.
It also included the amazing things now possible via automated voice which Amazon’s Alexa Challenge exemplifies. Meanwhile IBM’s Watson continues its conquest of carbon life forms with the trashing of an extraordinarily talented Go grand master. Starcraft fans with their refined skills in deceptive war strategies are apparently the next target for combat.
Meanwhile, a whole string of new brands are emerging in bot design, deep learning and personal assistance. Heard of MyWave, Mindmeld, creativevirtual, myclever and x.ai? Happily there are UK brands in this list!
I bumped into a friend on the tube this morning. His organisation has just installed a hotel bot using interactive text response. Having used chapter 11 to leave their legacy contact centre business behind, they are looking to remake their fortunes in the brave new world of open-source forking and API style design.
In fact, last week I was treated to a behind the doors session with a well-known UK high street brand. The big surprise was that they are now writing their own software for those very same reasons. It seems like only yesterday when the smart response was to sympathise with over budget, underwhelming in-house IT deployments. No more it seems. Legacy brands want to code as well as the fintechs.
Watching The Next Phase In Computing Unfold Is Awesome
Both times I’ve run the intelligent assistance briefing, the collective response (including mine) has been in awe of what’s under way. Looking through the lens of a conversational interface, whether rooted in text, as is the template for messaging which billions of consumers already understand, or via voice, which is still the fastest, hands-free version of communicating yet invented, the possibilities for re-animating how customers and brand engage in a self-service context are ‘bigger than big’.
The not so secret driver behind bots of course is our umbilical addiction to smartphones. In this world of frictionless expectation and instant 24×7 attention to our needs, browsers and human intervention are giving way to messaging platforms and automated workflows. These are revolutionising the mundane tasks in our daily lives.
- Ordering lunch – look out for Tacobot on Slack
- Tracking ad performance – check out Zoey on Messenger
- Booking a hotel room – see Expedia’s early experiments
- Being served as a hotel guest – experience Edward at the Radisson Blu Edwardian in London
- Remembering to get things done – look out for Jarvis on Messenger
- Scheduling a diary appointment – ask for Amy over at x.ai
In terms of what else could benefit from this conversational style of self service, we are still at the start of the gold rush.
Of course, any wannabe bot service will live or die based on the design skill used to craft the interactive dialogue. Some will also have the ability to get smarter over time. It is this talent which is maybe the greatest strength of messaging-based solutions.
Zuckerberg pointed out at F8 this year that globally we digest 60 billion messages a day. SMS peaked at 24 billion apparently. So, if so-called deep learning/machine-to-machine needs input to become smarter, they are few better examples of truly big data sources than messaging platform communities. In other words, the naïve bots of today are likely to become a lot smarter pretty fast based on such closed-loop learning.
Bots In Context
However, as with any element within an omni-channel ecosystem, bots have their strengths and weaknesses. Any commercial value is predicated on understanding how and where to best deploy them. Repetitive, routine tasks are a good place to start.
Also standard messaging platforms might not always be the best option for their foundation, as booking.com discovered. Facebook’s current inability to introduce a third party (the hotel) into a conversation between customer and a booking.com bot means it cannot be used to link these two groups. So booking.com is leveraging its own expertise in machine-to-machine learning to build an in-house solution for that use case.
KLM, on the other hand, is blazing a trail in their massively popular use of Facebook messenger to get their customers on and off their planes. Relative to email or even dedicated apps, it’s redefined what low-effort support should feel like for a passenger.
I’d also like to remind any bot evangelists reading this that ‘complex’ and ‘emotional’ interactions are probably not best tackled via an automated text-based medium. Instead, early recognition of such customer behaviour should result in seamless escalation to live assistance, probably voice. Or if the existing or potential customer value matters sufficiently even video engagement might be justified.
Before This Generation Of Bots
The point about bots, though, is their heritage. They are fundamentally tied to their messaging platforms and that means a mobile interface. Although messaging+ mobile enjoy the zeitgeist right now, browser-based engagement is not yet down and out. Nor is laptop-based viewing for that matter. Indeed, the relative failure of tablets to become a viable laptop alternative has resulted in some resurgence in laptop sales.
In these use cases, the classic format of an empty search box comes into play. We all still understand its power. Google taught us that. Natural language, machine-to-machine and semantic search combine to provide ‘under the bonnet’ smarts to make consistently high levels of first contact resolution possible.
Since a high percentage of the inbound volume in many contact centres consists of relatively simple customer issues, effective intelligent assistance has a dramatic impact on how live assistance is utilised once much of that demand is diverted to intelligent assistant self-serve.
Some centres might choose to maintain headcount and invest more time with customers who still require human intervention. That’s a solid contribution to the CX agenda. Others will bank the savings. Maybe more so now that there is talk of recession post Brexit. So expect a resurgence of the ‘more for less’ demands as service budgets come under further pressure.
Voice Challenges The Keyboard
Then of course there is voice. The original conversational interface for humans. Maybe a strange one for people who equate a progressive digital strategy with a text-only approach. But in truth, the vendor momentum sitting behind the ambition to make voice an alternative to the keyboard is massive.
Think the combined budgets of Apple, Amazon, Microsoft and Google (Facebook is behind the curve on this one) and you get the point. Moreover, speech recognition and natural language processing now hit a standard of real-world effectiveness that means we are seeing a strong if recent uptake in automated voice engagement.
Once Amazon’s Echo arrives in the UK, hopefully this autumn, the power of this tech will become much clearer. If you can’t wait that long then check out the challenge I mentioned earlier on. Hugely impressive examples.
Organisations will need to use these voice gateways to allow customers access to their own products and services. Global brands are likely to choose more than one vendor ecosystem to maximise their visibility to customers.
A Random Heading To Fit In Stuff Yet To Be Mentioned
Of course there is much more.
For instance, the crucial role of knowledge management in providing correct answers. It’s still a human-plus-machine partnership to keep that working as an ongoing investment in time and effort.
The new imperative for personalisation demands effective identification and authentication (ID&V). Biometrics is bound to be part of any modern solution. USAA uses voice, face and touch. Is that overkill or a smart response to customer choice?
Virtual assistants such as MyWave, VIV, Siri and the rest are by definition personal. They want to get to know you as part of their way of becoming more useful. That demands trust. And privacy. Not something all vendors have a strong reputation in. So is it time all organisations to publish and adhere to a customer data usage policy? Especially since imminent EU law is going to encourage higher standards around customer data.
Will bots turn our smartphones into the scam-infested channel that landlines have become? How will we triage messages from friends and family from the volumes generated by the bots we subscribe to? Will bots become a back door for hackers? As bots boom, so will the headlines.
A Final Shout Out And Call To Action
For those who know me, they will recognise that intelligent assistance has become my new next-generation topic. I’ve become passionate about the whole thing. It’s a great ringside seat to watch the next phase of the 21st century unfold.
As much as I’ve had fun putting this material together and running the briefing, I must point out the tireless and professional work put in by the Opus Research team of Dan Miller, Amy Stapleton and Derek Top, who I consider the original pioneers in this space.
If you are States-side then check out their September conference in San Francisco. We just enjoyed our version here in the UK, so keep your eyes peeled when they return next year. There is a ton of great insight and reports on their site which any fan of this topic should be locked into.
Meanwhile, I’m catering to UK audiences who feel Brexitted and unable to justify a plane ticket to California! Next briefings are due in late July, then early September. Just get in contact for the details and agenda.
Thank you for reading.