If you are in the process of investing in some form of intelligent assistance for sales, marketing or service reasons, then this is worth your time. I’ve interviewed one of the top experts in the field – Dan Miller.
Dan Miller heads up a great team out in San Francisco who have charted the evolutionary arc of this technology from the very start. With a freshly published report of the top vendors and a forthcoming London conference, I skyped Dan and tasked him with advising the ‘newly fascinated’ business leader on how to make the right investment decisions and transform their self service using Intelligent Assistance.
The result is a really useful set of tips to help you make the right decision. You can listen (the director’s cut 30 mins) or go for the shorter written version below. Or absorb both if you really want to squeeze the juice. Enjoy!
Martin: So tell us a little about Opus Research and your long-standing interest in conversational commerce.
Dan: I founded Opus Research back in the mid-1980s and immediately got fascinated with applications and services that combined the phone with advanced computing technologies. High-volume contact centers with huge interactive voice response systems that answered millions of calls were a natural and applications like automated directory assistance which captured spoken words and responded in nanoseconds were fascinating to me.
By 2002, however, the growth of wireless phones, the Internet, World Wide Web and VoIP transformed everything and we started looking at “conversational” technologies. These are the basis of today’s applications that incorporate speech recognition, dialog management, turn-taking and natural language understanding.
Martin: 2016 was an important moment for mainstreaming the IA industry. But in truth it’s been around much longer. How do you describe its origins and evolution up to last year’s excitement over bots.
Dan: Opus Research convened an event called the “Conversational Commerce Conference” back in 2012. We observed that the move to multi-channel, digital commerce would put classic sales and marketing at odds with customer support and technical assistance.
We knew this by looking at how the role of call centers was changing. The metamorphosis from phone-based “call centers” to multi-channel contact centers embracing chat, email, text messaging and social networks was well underway. The omnichannel contact center was conceptual and aspirational at the time because each customer touchpoint took place in a silo. Likewise, customer experience was dis-continuous and dis-integrated.
The grand irony is that when I coined the term “Conversational Commerce” I was thinking of “technologies that improved conversations between people and machines and people and their peers.” Today the term has gained traction because people think of it as “bots on messaging platforms” and they have a different set of success criteria.
They are not really conversational, they are “one-and-done” results oriented and rather than supporting dialogs, they support carousels and multi-choice buttons that actually discourage conversations and deep customer engagement.
Martin: So how does genuine conversational commerce offer opportunity across the engagement lifecycle and fit into being a digitally competent brand?
Dan: This question will be the crux of many of the sessions and presentations at the upcoming Intelligent Assistants Conference-London (May 4-5). What we call “The Botsplosion” has raised conversational commerce’s profile tremendously. Every brand has to have a “bot strategy” and we’ve seen hundreds of branded bots on Facebook Messenger and quite a few companies that have made customer care and order taking into “skills” for Alexa.
Financial services companies, airlines, hotel chains, retailers and many others have built bots, much like they built mobile apps. They want to be everywhere their customers or prospects might take them.
The challenge, of course, is to avoid building yet-another silo or creating another source of stranded investment if the uptake is slow on a specific platform.
As you’ve noted, Martin, The Blotsplosion has been a bit messy. There are hundreds (maybe thousands) of marginally useful, hyper-focused ‘bots’ out there giving doubters grounds to use the tired descriptions of the “trough of disillusionment” or “uncanny valley” where bot users are wandering in the wilderness.
Meanwhile, dozens (maybe hundreds) of implementations are proving that Intelligent Assistants are genuinely useful personal agents, advisors or intermediaries during our digital commerce journeys.
Martin: In your recent analysis of the market, you estimate the size and growth of what’s going on. Where are we in the growth curve. Still at early adoption or further on?
Dan: When we did our first market assessment a couple of years ago, we looked at 13 firms with tools, platforms or services that support development and deployment of enterprise-oriented Intelligent Assistants. We thought that they collectively generated about $225 million in revenue for initial start-up fees, licensing, activity on APIs and professional services.
By 2017 we see nearly 30 contenders for an Enterprise Intelligent Assistant ecosystem where annual spending will exceed $1.2 billion. From our point of view, we’re still in the super-normal growth stage. We forecast spending in 2021 to exceed $4.5 billion and it could be much greater.
We’re past the “early adopter” stage for sure, both from the point of view of end-users/customers and sophisticated brands.
Looking at the former, billions of people have smartphones with personal virtual assistants (Siri, Google Assistant, Bixby, Cortana) that are constantly improving their ability to understand an individual’s intent and even anticipate their next best action.
More importantly, with customer care and customer experience infrastructure moving to “The Cloud”, the speed at which companies can refresh their capabilities has accelerated dramatically.
Intelligent Assistant Solution providers now tout their ability to get brands up and running with a robust Intelligent Assistant in a matter of weeks.
Even if we’re in the early adopter phase, the speed at which we move to mainstream is impressive.
Martin: In terms of helping decision makers understand their choices, how you classify the very broad range of vendors out there? Are there any simple ways of grouping them?
Dan: I’m really glad you asked this. The answer is “yes.”
You will definitely find a set of solutions providers that put an emphasis on “out-of-the-box” capabilities, speed to deploy and speed to results. They will have engagement models that, in essence, put a conversational front end on a static set of data, meaning that they will demonstrate how a company’s FAQ can serve as the raw material for a “conversational” IA designed to recognize that a customer’s question can be answered by the contents of an FAQ and act accordingly.
This can be accomplished largely through the use of a “rules based” engine and rudimentary Natural Language Processing (NLP) with the ability to “understand” what an individual is saying, even he or she uses highly imprecise language or dialects. The knowledge base is static and the IA either knows the answer or does not and, when it does, the results are fast and satisfying.
At the other side of the spectrum are firms that apply highly-advanced technologies – namely Deep Neural Networking, Cognitive Computing, Robotic Process Automation, Predictive Analytics or Conversational Analytics – to enable an IA to provide answers or suggest best actions based on a complex and dynamic set of data and metadata.
It has ingested (therefore “knows”) information about the customer, including location, transaction history, loyalty status and the like. Its responses are informed by a voluminous amount of information from a company’s knowledge base and from third parties. Often, the solutions from this category of vendor are heavily reliant on expensive professional services required to “discover” the most efficient applications of advanced technologies and support ongoing refinement of answers.
In true Goldilocks fashion, there are several solutions in the “just right” category, bringing a mix of core technologies, practices, protocols, development and reporting tools and professional services that promise speed to deploy, speed to results and ongoing refinement that conform to a firm’s requirements for service quality, cost control and customer experience. We won’t name names here, but decision makers should recognize that the answers are out there.
At this point in the evolution of IAs it is important to take note of the profound changes that IAs are making in customer experience. Their prime directive is to take quick stock of an individual’s context, status, preferences and, ultimately, intent in order to provide the best answers or propose the right action.
Martin: If I’m totally new to using IA for self service, what do I need to be thinking about to help me recognise the right type of technology and partner?
Dan: There are five basic considerations.
- What works out of the box? Meaning that it requires a minimum of expensive professional services and ongoing support.
- What solutions minimize incremental costs by leveraging existing knowledge management capabilities and the personnel responsible for providing the right answers?
- Who’s going to support my “customer first” strategy? This is not a science experiment. I’d like to see implementations that have tangible, measurable impact on CSAT, NPS and customer effort reduction.
- While we’re at it, make sure that the reporting and analytics capabilities are in concert with existing KPIs for contact center and customer experience operations – as well as existing compensation schemes.
- Then there’s Price and ROI. Just make sure you can build a business case around hiring an IA or virtual agent.
Martin: I’ve also noticed that this form of self service is being promoted by Forrester as relevant for B2B purchasing journeys. Have you seen any other examples outside customer service where IA is valuable?
Dan: Absolutely! Opus Research is tracking an entire category of “Enterprise Assistants” and “Enterprise Advisors” that apply NLU(natural language understanding), ML(machine learning), Analytics and related technologies to supply chain management, inside sales and sales qualifying and management, human resources (including onboarding, training, mentoring and the like), not to mention “bots” on enterprise collaboration platforms. Expect this report to come out after our conference in May.
Martin: Given your insight into the market from all the conferences and research you published over the last half dozen years, what are your favourite examples of great deployments?
Dan: This may be a surprise because it is slightly obscure but a telecommunications company called Windstream in the United States has employed Nuance Communications NINA technology in a way that is tailor-made to the IA business case.
As described in this case study at our 2014 Intelligent Assistants Conference by Sara Day, VP of Consumer Marketing, the company had huge financial incentive to provide high levels of customer care to its residential telco customers even as they were becoming a world of “cord cutters” and wireless customers.
Recognizing that it could anticipate the mix of calls received by customer care contact centers and that IAs could do a great job of expeditiously handling 90% or more of those calls, the company found that it could manage the expenses associated with customer support while simultaneously increasing measurable customer satisfaction.
Looking at the wide world of bots, I’m very partial to Edward, an SMS/text-based virtual concierge offered by the Radisson Blu Edwardian Hotel in London (and elsewhere).
Martin: Tell us more about the forthcoming London conference. What are the main themes and use cases?
Dan: At the Intelligent Assistants Conference-London we continue our tradition of combining thought leadership and industry-defining research with real-world case studies and use cases that provide insights into what it has taken to succeed IA strategies. Past gatherings have featured speakers from leading financial institutions, telecoms, packaged goods suppliers, insurance companies and travel/hospitality leaders.
This year, we’re happy to feature T-Mobile and TalkTalk (telecom), Enfield Council (local government), Motability (health related), BGL Group (financial services) and Amazon.com (digital retailing. Their key executives will talk about how IAs improve customer experience and customer care strategies, how they regard alternative vendors and what it takes to promote a unified IA strategy across multiple business units.
We’ll also be talking about:
- The balance between privacy and personalization
- How best to talk to “metabots” like Alexa, Google Assistant and Cortana
- Making the move to secure, trusted conversational IAs that support marketing, customer care, sales and transactions
Martin: Dan thank you so for your expert view on the market and all the free tips. I’m looking forward to catching up again at the conference in May. Meanwhile I hope we have inspired others to come along.
Useful Links
Opus Research
- Decision Makers Guide To Enterprise Intelligent Assistants
- London IA Conference
- Join Dan’s Intelligent Assistants, Developers and Implementers Linkedin group
