Rise of the machines: How AI is changing unified communications

Artificial intelligence is now moving into live UC environments at scale. We look at the impact it will have and what lies ahead

There is more and more talk these days about how unified communications (UC) is set to be transformed by artificial intelligence (AI) – with the topic getting plenty of airtime at this year’s UC Expo in London in May. So is AI ready to come out of the lab and into live UC environments? And to what extent is it having an impact already?

One strand of activity was presented in a neat example from Cisco at the 2019 UC Expo. Amy Chang, senior vice-president of the firm’s collaboration technology group, demonstrated on how “cognitive collaboration” can improve a virtual meeting by giving participants useful information about others in attendance, with intelligence gathered by bots from social media and other online sources. So that’s one real AI-derived benefit that appears poised to land.

The UC space has also changed substantially in the past two years. With the likes of Microsoft fully engaged, it is now about so much more than voice over IP (VoIP), video conferencing and instant messaging. Technology such as interactive whiteboards, team spaces and document management all sit beneath the UC umbrella – and we can add AI-driven chatbots to that burgeoning list, with a remit to help users who are increasingly likely to need help with an excess of data.

“The intersection of UC and AI is all about helping people do their jobs better by letting machines interpret data and analyse information faster,” said independent network analyst Zeus Kerravala in a keynote address at UC Expo 2018, and that still holds good in 2019.

Julian Harris, head of AI technology research at knowledge network CognitionX, set some of the context for AI in UC, whether today or tomorrow, when he noted at the company’s recent CogX 2019 event that there are currently at least 700 conversational AI technologies in the marketplace.

But beyond this emerging capability at the margins, profound shifts are visible in the mainstream. Leading the way here is the fact that phone calls are in decline as a communication medium. So even if conversational AI is still in its infancy in terms of capabilities (and the experts agree it is, despite the dramatic take-up for voice-activated services using Alexa and the capabilities of Google Duplex, for example), the landscape of how individuals interact and communicate is always changing.

AI-driven voice is just one nascent element among many in a complicated, shifting space.

How is the communications landscape on the move? Today, we can say there are declining phone calls, a growth in messaging, a parallel explosion in video calls, and – in the customer engagement space – growing call centre automation. One measure is found in WhatsApp messaging volumes, which have risen exponentially, from one billion a day in 2011 to 65 billion a day by mid-2018.

There are now more than five billion registered users for messaging platforms worldwide, with WhatsApp and Facebook Messenger the most used.

Bots and the corporation

So a lot of change is happening with communications in the wider world. But let’s now turn back specifically to UC, and AI’s application within it.

Roy Lines is an independent digital transformation specialist, with DevOps and cloud services among the tools and approaches he applies. He works with large corporates in verticals such as fast food, consumer goods, FinTech and gaming, and says that when it comes to UC, AI is already having an impact. 

“This is an interesting time, because chatbots can already add so much value by supplementing human communications,” he says. “They can make carrying out crucial tasks simpler and they can add to conference calls or make follow-up actions after virtual meetings easier by providing relevant information to individuals just when they need it.”

If that’s the promise, in practice Lines has already begun using messaging and collaboration platforms such as Slack for information-sharing in more than one of his large corporate clients. By using automated bots on Slack, you can add to conversations by monitoring and delivering supporting information right on cue, says Lines.

Even more strikingly, it is straightforward to use Slack as a channel to simplify actions in a business by using AI, he says. 

Read more about AI

  • People in the Nordic countries are more welcoming and less fearful of the advance of artificial intelligence than those in other regions.
  • Artificial intelligence workloads impact storage, with NVMe flash needed for GPU processing at the highest levels of performance – and there are other choices right through the AI data lifecycle.
  • AI and machine learning are being introduced into corporate recruitment to save time and money, but employers need to understand the legal risks.

“In a food-delivery business, for example, sometimes a company won’t have capacity to take customer orders because of a kitchen failure – an oven is offline, say – or because of reduced delivery capacity,” says Lines. “To handle this, you can use a bot that allows Slack to be the channel to instantly reset customer booking capabilities so that they match the situation.

“The channel to do this could equally well be WhatsApp or even SMS messaging. With AI, action can be taken just through a simple text command from a phone – or else a voice instruction, potentially. Plus it provides that audit trail of actions taken – everyone in the company with access can see what action was taken, when it was taken, and what the trigger was.”

Lines says the use of AI-driven voice on conference calls is also available now, so bots can monitor a call and interject with useful supportive suggestions such as whether to circulate a particular document that backs up what is being discussed.

“There are corporates I work with that are using this capability now,” he says. “It can just be a fact or data point, too, that the bot supplies – not always a document for circulation. It might confirm some relevant sales figures on request, for example.

“Bots are a great way of democratising data across an organisation, as so many companies are still set up only for the senior leadership to discuss financial figures, for example. A Slack channel where the CEO ends up talking to developers, middle managers and more about myriad ideas across the business can be very powerful.”

AI and the consumer

UC doesn’t just refer to the internal interactions in a business, but also reaches into marketing systems and customer relationships.

Richard Robinson, chief commercial officer at data-science-in-marketing business Data Practitioners, says the wider communications play for UC embraces the customer experience – and AI is coming through rapidly.

“AI can boost the on-site experience for consumers and also enable personalisation – everything from visuals to chatbots to augmented reality to personalised website visits,” says Robinson. “But social listening is one area where AI is impacting things, enabling companies to be more targeted and responsive to customers.”

When it comes to call centres and automation, the potential to tie together communications across many channels is there but has only been partially realised, he says.

“Today there are so many ways to engage – call centres, online chat, email, via online advertising, via social media. Even today, these channels are often too siloed. But AI is helping companies to knit things together into something approaching a single customer view, with consistency in the interactions as the systems learn from each other every time a customer engages.”

At a relatively basic level, AI can serve up recommendations to customers, such as the way Netflix suggests TV and films to customers based on their prior choices and actions. The next stage would be to factor in other information, such as the watcher’s mood or the time of day or the weather or some other variable. The idea is for AI to make rounded decisions and recommendations.

“We work with a luxury retailer,” says Robinson. “Through machine learning, we have helped to identify a new segment of customers that the retailer has not yet engaged with. And we have used behavioural science to improve the chances of engaging well with the customers in this group – using the right colour palette, presenting the right offer, and with the right tone of voice. The outcome? A 68% increase in sales and a 45% decrease in acquisition cost. That is the kind of difference AI can sometimes make.”

But he adds: “It is early days still. AI is making a difference – with a long way still to travel. The opportunity is clearly there for better, more human, more informed engagement with would-be customers, just as there is the opportunity for better internal communications.”

One step at a time

If the promise of many AI technologies in UC is already clear, it is also important to sound a note of caution: with many of the technologies being used today, there is still much to do.

John Taylor, CEO of Action.ai, which has developed conversational voice tools for businesses, with a particular focus on sectors such as financial services, said in an address at CogX 2019: “Our technology is excellent, and our work to help businesses transform the customer experience is happening now. But the fact is that the promise of chatbots still outstrips the reality in many regards.”

Consider a well-resourced company such as Uber, which has a chatbot that is still dogged by one-star reviews. Why? Because it is hard for this stuff to work consistently. Human language is complex and people communicate in ambiguous, informal ways much of the time. Also, it doesn’t take much for any individual to abandon a voice-activated chatbot if it’s not working.

Taylor adds: “The work involved in managing dialogue so that it delivers and an individual can talk naturally and do things like employ jargon is complicated but also methodical. For voice to work in specific contexts, it takes depth, but it can be done.

“It is possible to build a great customer experience with a chatbot – it just needs thought and careful design. It is starting to happen now, and Action.ai and others are part of what is a step-by-step journey.”

Read more on Unified communications

Data Center
Data Management