In October 2005, I published The Voice 2.0 Manifesto. The Manifesto’s theme was the marriage of voice to the web, and all of the accompanying technological and business shifts that might occur as a result. Five-and-a half-years later, some – indeed I would argue many – of the predictions made then have come true, or at least partially true. Voice minutes cost next to nothing, voice applications are creating real value beyond just carriage in some segments, and open programmable architectures for voice services have emerged.
A lot has changed in five years, however. Skype, which was an early stage start-up with no revenue model, is now the dominate supplier of international long distance calling minutes worldwide. Smartphones finally exploded onto the scene, and in the process dashed the old mobile models to pieces. And the internet has morphed into the “semantic web” as increasingly device to device communication becomes as important as device to human.
Voice 2.0 was about how the internet intersected with the voice network. Voice 3.0 moves beyond Voice 2.0, as the voice network becomes the Voice Web – the amalgam of voice, and the internet and its intersection with real world problems to create new forms of value, useful in everyday life. By extension, Voice 3.0 is therefore also about how voice applications incorporate the best of the internet to increase the options and opportunities for developers, and ultimately create new value for users, corporations, and investors.
Like Voice 2.0, Voice 3.0 is a user and developer focused view of the world. It’s “all about me”, about how the focus is shifting from a supply driven world managed by the some of the largest corporations in the world, the telephone networks, to a demand driven world managed by users of those networks. We’ve already seen dramatic evidence of this momentum in the mobile world, as customers flocked to devices supplied by Apple and Android partners, primarily because of their ability to be infinitely customized via mobile applications to the customers’ needs and desires. In the Voice 3.0 world, we see the same dynamic as the large suppliers of voice services are beginning to recognize that innovation comes from users who find and adapt tools to new uses, and from the developers who build applications to “productize” those new uses. Just as in mobile, the winners will be those who successfully unlock the demand driven model.
Let’s start with the ubiquity of voice, or how voice gets embedded into the fabric of ordinary digital life. As Martin Geddes points out, we humans have really only got three modalities of communication available to us – gesture, voice and the written word. Despite the rhetoric around video, voice, which represents at least 1/3 of our communication palate, isn’t going to disappear at any time in the near future.
In fact, voice is being adapted to many different uses than before. For example, my own children communicate with their friends more often on the xBox network using spoken words than on the telephone. And, in fact, when playing multi-person internet games on a PC, they will turn to Skype as a substitute for xBox voice.
There are two billion voice “input terminals” in existence on the planet today, including telephones and computers. Many people now suggest that an explosion of voice enabled devices is imminent as all kinds of devices become voice enabled. Just as the internet has become embedded in billions of devices, the Voice Web will also be.
So what does the world look like when 10 or 15 billion devices are now voice enabled? When two or three voice enabled devices exist for every man, woman and child on the planet?
Two obvious possible outcomes are that:
- The personal voice terminal, or the telephone, disappears. There will be no need to hold a microphone to our faces when one can simply speak to the ether, and make communications connections.
- Voice control as user interface becomes much more prevalent. When you can speak to your automobile, television, computer or home appliances, then interaction models change dramatically.
Perhaps the largest promise of the ubiquity of the Voice Web is that the over 1.5B people who are functionally illiterate can join suddenly get off the sidelines and begin actively participating in the emerging commerce model of the voice enabled web. Until now, much of this population has been unable to participant online. In India, for example, a great deal of commerce is conducted by illiterate women en masse via dumb phones. This type of commerce is equivalent to the dark web today — important, but off-limits. With the Voice Web, phones don’t need to be smartphones to connect people to the larger network. Global economics will change when the participation of those previously sidelined becomes active and at significant scale.
The second big theme of Voice 3.0 is the programmable Voice Web. Programmability has been a topic in voice circles for years, beginning with the old-fashioned computer aided telephony right through to today’s hot topic, Communications Enabled Business Processes.
Today’s state-of-the-art is to marry a voice channel to a processing channel to create new kinds of applications. The earliest interactive voice response systems simply allowed individuals to press keys on the telephone keypad to alter a call path – press 1 for sales, 2 for support, and so on. More advanced systems today can automate different kinds of business processes, by separating business logic from the voice channel itself. For example, using analytics tools to track incoming customer response to advertising on voice channels as Ifbyphone does, performing database lookups in response to incoming caller ID information for customer relationship management purposes, or creating outbound reminder and polling systems.
In all these examples, the voice channel is treated as a relatively linear audio object which is only loosely coupled to other systems. The user communicates with a standalone IVR that exists solely for the purpose of providing the application. It’s as if today’s state-of-the-art voice systems are the walled garden AOL’s and MSN’s of yester-year, and the world wide web hasn’t yet happened. The result has been successful unified communications systems which give a great experience in totality, but the learning curve is high if customers choose to switch service providers.
The internet, however, grew well beyond those old systems. It evolved from the hyper-linked world of the early web to container objects supporting media and programmable objects, with sophisticated tracking systems, scripting, offline execution, mobility and more. The web went from being a hyperlinked text library, to the largest programmable application on the planet, fuelled by open standards, lightweight communications infrastructure, standards which allowed content to be separated from logic and presentation, and an explosion of end-user devices, including today’s mobile devices.
Voice is on the cusp of the same revolution – a revolution that will be defined by letting the customer define the business logic of the application, not the service provider. Imagine a “hyper-talk” protocol, where the voice servers of today evolve to become more akin to a web server – a hyper-linking voice application with the ability to autonomously download and execute other data and voice objects. In that world, for example:
- Voice mail wouldn’t be the linear audio message it is today. It would be more like HTML email, with live links and buttons embedded. Made a reservation at a restaurant? A day before, the restaurant’s reminder system might leave a voice object in your mail box: “Press 1 to confirm your reservation, 2 to cancel or 3 to speak with an operator”. Your voice mail box would know how to correctly execute the scripts embedded in that voice object, process the button presses, and inform the restaurant’s reservation system of your responses.
- Caller ID spoofing would be a thing of the past. If your bank phones, an authenticated caller ID system, akin to HTTPS on the web, would give you the confidence to deal with them directly.
Not only is this world coming, it’s required if voice and the phone is to stay relevant as a communications medium. Phone is so far behind today that it’s losing traction for business, in favour of the impersonal web where identity and programmability already exist. And yet what could be more important in business than personal contact, and more personal than a voice conversation?
Ultimately, we’ll build systems where communications result in artefacts that can be consumed by services that have not been pre-specified. Think, for example, of the role that RSS played in the syndication of content, and imagine a similar world for voice. Tool chains will be created that will allow people to participate in building these services, and an explosion of new applications to consume these voice artefacts will be built.
And that leads directly to the third big theme of Voice 3.0, which is the arrival of the semantic voice network. In the world of the internet, massive businesses have been built around the collection and management of huge databases of content – think of businesses like Amazon and Google for example. On the Voice Web, these databases will exist, but more likely as the proprietary property of enterprise. In their quest for accountable, measurable and actionable IT assets, management will turn an eye to voice—the last opaque data object in the organization.
We’ll see the advent of:
- Voice as an Asset — The combination of Voice Analytics used to help organize and categorize voice content, metadata surrounding the original voice object to assist with relevancy, weighting, and prioritization. We’re already seeing this in the post-Enron era, as corporations struggle with making voice an auditable asset, and these efforts will only become more sophisticated over time.
- Voice as Relevant — Voice content will be findable. Just as Google is a query mechanism that sources and displays relevant text (mostly) content, voice will have its search engine unlocking the value of those stored voice assets throughout enterprise.
- Voice in Context — Voice content will live side by side with other data types within enterprise transactional systems like CRM, and ERP. Voice will be used to provide the emotional context around the other data objects to help provide richness to the inquiry. Imagine searchable sales calls in the CRM next to the account information and text logs for a customer.
As voice becomes a “big data” asset, databases of conversations will be built, cut into snippets, decoded, analyzed, and added to the enterprise knowledge base. The Voice 3.0 tool chain will provide APIs that give access to these snippets, both privately and on the web. And ultimately, the content, context, and meaning of audio conversations will become a key input into business processes.
Voice 3.0 impact
Accompanying these three themes will be a radical shift in business model – from a supply constrained model to a demand driven model. Customers will want to be able to choose a dial tone provider, and applications independently. Innovation will accelerate as customers take these new services and use them in even newer and more different ways than anticipated.
Developer ecosystems will form a “virtuous cycle”, a self-reinforcing market that gathers momentum as more adopters flock to the platform. Platforms will try to recruit applications developers, customers will look to the ability of platforms to deliver the third party applications that they need and want as a key differentiator in the market, and as more customers adopt the platform, more developers will also, seeing opportunity.
Some applications will be pre-built, some custom built, and some modular. As a result we’ll see a thriving market for suppliers like today’s Ifbyphone, Twilio. and Voxeo You can think of Ifbyphone as the Ikea of voice, supplying modular components, and Twilio as the Home Depot of voice, supplying building blocks and materials. Both are important in the market. Recognizing the broad range of needs in the market, Voxeo plays both ends of the spectrum with services ranging from the self-service Tropo and Imified platforms, all the way to custom built applications with100% uptime SLA’s based on their Prophecy and Prism platforms.
The voice service provider of tomorrow will probably be much more like today’s SaaS providers – a hosted Voice as a Service business. “VaaS” will deliver a core managed and hosted voice service, decoupled from both the context of use and from the internet service provider. The service package will include not just voice, but detailed statistics, group management controls, and more. And it will bristle with API’s that will enable an ecosystem of other players to be built around it.
These in turn will unlock the value of stored voice assets, allowing the growth voice as an organizational asset. Iotum, for example, has stored every recording made by users of its Calliflower system since it was first introduced. Newcomer HarQen takes this to the next level, focusing exclusively on what do to with voice data once a call terminates. Both companies are in contrast but complementary to metered transport models, and both recognize that there is value in those voice assets beyond the termination fee charged during their creation. For example, HarQen has built a database of over 100,000 phone based job interviews that is not only valuable today, but as that database grows into the millions, may transform recruiting tomorrow. HarQen’s business model focuses on monetizing voice as a rich media asset that can be archived, which is how they’re able to turn a $.10 phone call into an $11.57 voice asset that customers are lining up to pay for.
Winners and Losers
For some time, Voice 3.0 is going to be a messy hybrid of delivery technologies. The Internet will be “reliably unreliable”. Telcos will stake out a corner. The “under the floor” players like Ericsson will take an increasingly large role in creating and managing voice-ready delivery networks. Some of the cloud communications and commerce infrastructure players like Apple, Google and Amazon may reverse themselves into owning delivery assets to just “make it work”.
Don’t be confused by the messiness.
Today’s large voice network suppliers – the companies that have always relied on constraining supply as their business model – must change or fail in the wake of companies focused on meeting demand. Ironically, the network asset that many telco’s market as their chief advantage may just be their largest liability. The technical rigidity of large telco infrastructure (and sclerotic business models) may just lock the incumbents out of the game, as they struggle to keep up with the massive proliferation of 10 billion non-telco voice nodes on networks. Moreover, the temporary solution of bridging to and from the PSTN actually weakens their ability to compete by artificially enlarging the network effect of upstarts like Skype.
Network effects in the Voice 3.0 world become even more important. Will an open standard emerge? Although many die-hard networking folks would prefer that scenario, it’s hard to say. We may find ourselves in a world where a dominant proprietary player like Skype controls the platform, as a result of winning the race to build thriving developer ecosystems, and the applications that customers use and want.
And new opportunities will explode onto the scene as businesses perhaps not even imagined today will discover how to turn a $.10 phone call, into a $10 voice asset.
Acknowledgements: Many people in our industry contributed their time and ideas to this posting over the last few weeks. I’d like to publicly thank Andy Abramson, Jonathan Christensen, Evan Cooke, Dean Elwood, Kelly Fitzsimmons, Martin Geddes, Thomas Howe, Erik Lagerway, Irv Shapiro, Dan York, and most of all Lee Dryburgh, who provided the initial impetus.