Telco use-cases for AI: A simple categorisation model

The coming years will see the application of AI technology across all sectors of the economy and life. The telecoms industry is no different. Although I’ve been commenting on telco-sector AI in the context of “TelcoFuturism” for some time (link), and co-ran a workshop on it in May 2017 (link), the last few months have seen a notable upswing in interest. I’d say that the public use-cases now seem to be significantly in advance of those for blockchain, in terms of potentially-transformational technologies.

That said, it can still be hard for many executives to grasp exactly what is likely to change, and when, for AI/telecoms combinations. This is highlighted by the surge in AI-related panels, presentations and even complete streams at industry conferences – although sometimes I see more interest from generalist AI people about the telecoms vertical, versus telecom specialists looking at what’s new. 

Both sides of the equation have large volumes of obscure acronyms, multi-layered technology stacks, and complex volume chains – which can mean that mutual understanding is often confined to narrow niches. AI covers machine- and deep-learning, language processing, machine-vision and much more. Telecoms includes vast realms of internal systems and processes that are unknown to most who are not insiders – domains like core networks, OSS/BSS, network optimisation, toll fraud and service-assurance are alien to those not steeped in the industry.

One of the ways I’ve been using to “set the scene” for describing AI/telecoms intersections is to simplify and categorise the use-case areas. I count three, possibly four, large “buckets” into which a variety of telecom AI impacts will fit. These buckets are not based on either specific AI or telecoms technology slices, but more on understandable business functions and roles:

• Dealing with customers
• Managing operations
• Creating new services
• (External risks)

 Within each of these areas, there are many, many sub-sectors – and also some overlap.

“Dealing with customers” can include everything from voice/text chatbots for customer-service, through to predictions of which customers are least-happy and may “churn” to competitors. Where telcos have retail outlets, it could incorporate various in-store technologies, or it could be about smarter web-consoles for B2B customers running complex managed services.

“Managing operations” is even more diverse – it could be fault prediction for network elements, optimising the 100s of configuration variables for radio networks, spotting fraudulent traffic to international premium-rate numbers, allocating engineering resources more productively, or protecting against hackers and malware. There are hundreds of possible uses here, which mostly overlay on top of existing operational/business support systems (OSS/BSS) See also my recent post (link)

“New services” also spans a range of areas, but broadly splits between AI-enabled and AI-enabling services. An AI-enabled service could be a local-language voice assistant added to a cable operator’s set-top box or remote control. Or it could be the provision of integrated “smart city” solutions including video-cameras and security analytics. AI-enablement could include offering “edge” servers for hosting local processing, milliseconds transport-time away from a device, or it could be the provision of anonymised bulk data for others to apply algorithms to. Telco opportunities with IoT+AI include both enablement and enabled services, in numerous manifestations.

The “risks” category includes a diffuse set of possibilities by which AI might harm the telecom industry, or dampen demand for services. Smarter devices (eg autonomous vehicles) will be able to host their own offline image processing & route-planning locally, rather than needing realtime connectivity at 5G speeds/latencies. Another threat could be customers’ smart assistants renegotiating price-plans on their behalf – after crowdsourcing millions of conversations to deduce how best to game the retention staff’s scripts and objections. (Of course in the latter example, the customer-retention team could themselves be bots). Numerous types of automated “least-cost X” and arbitrage engines are likely to emerge. Various security risks are also probable here too.

Clearly, using just these four "buckets" misses much of the fine-grained detail. But I find it helpful as a starting point, as most top-level industry issues apply differently to each. 

Consider input data, for example – for both customer management and operations, telcos have abundant historical records and ongoing data collection that may generate terabytes per day. But for the former, privacy considerations often come to the fore in terms of regulation and risk, while this is far less of a concern for internal operational data, for example on how the network is running. For new services, almost by definition the focus is on collecting/processing/transporting new data, rather than deriving conclusions from existing sets. 

This four-way framework is also useful for thinking about different types of ROI model - split broadly between impacting existing revenues, existing costs, new revenues and potential changes to underlying assumptions. 

I'll be covering these topics in more depth in various upcoming presentations and reports, as well as looking at other areas of telco-linked innovation such as blockchain, 5G and enterprise verticals. Please get in touch if you would like more detail, or are interested in internal workshops, external support through events or white papers, or are seeking ongoing strategic advisory support.

A challenge for new software-based telecom services NFV and eSIM: demarcation points

A couple of months ago, I had a problem with my home broadband connection, which intermittently cut out, or needed the router to be switched on/off again.

When I arranged a call-out from a BT engineer, I was told in no uncertain terms that if the problem was with my house-wiring, I'd be charged a fee. If it was a fault with the termination box or router, or external wiring, then BT would fix or replace it.

In other words, the little white BT-branded box and socket is the "demarcation point". It's where BT Openreach's responsibility stops, and the customer's starts. (BT Retail also provides me with an Infinity router, for Internet access, but that's technically a different service component, rather than network access).

Something similar is true with my mobile phone - the SIM card represents the demarcation-point for the connectivity part of Vodafone's service offering. It is essentially part of the network, rather than part of my phone. While I need operator settings and configurations to be pushed down to my device (eg APNs for Internet), that again is something separate.

Now consider what happens in a more software-driven, virtualised world.

For fixed connections, we may find either some sort of white-box termination unit (residential or business) down to which might be pushed virtual functions like a vRouter, vFirewall, vSmartHomeServer or whatever. Or alternatively, those virtualised functions might be located in a data-centre or local exchange server - there's a lot of talk about cloud-based vSetTopBoxes, for example.

And in mobile communications, we are starting to see eSIMs emerge for some use-cases, in which the manufacturer embeds a physical SIM chip in devices, and operator profiles get pushed down to it, or potentially switched between.

In both those cases, and various other forms of virtual/physical scenarios, we are moving from a world of clear and unambiguous demarcation points, to a world in which they are much less well-defined.

Today, if people switch SIM cards over to use a different network, or if a SIM stops working and needs replacing, or if consumers churn from one home broadband provider to another, then the lines of responsibility are pretty clear and obvious. There's very little finger-pointing that can occur.

But what happens with an eSIM? Is the manufacturer responsible if it stops working? Or is that the fault of the operator whose profile is working on it, the service provider which packaged and downloaded that profile, the software vendor(s) involved, the retailer where the device was bought - or worse, a second operator if the failure occured during a switch-over to a new profile? Who pays for the diagnosis, or for replacing the whole device if the eSIM can't be separated out? What happens if data is lost? Who is liable?

A similar set of questions apply with 3rd-party VNFs, or other software functions which drive underlying connectivity, or sit in the data path. We may be entering a world in which there is a "VNF AppStore" model - where the customer chooses between different software routers, or WiFi controllers, or firewalls. For businesses it may be possible to sell ethernet-style connectivity, and let the CIO take responsibility for those other connectivity-oriented functions, but that's clearly not an option for the consumer market.

This is different to higher-level virtual applications - if a game or a VoIP service stops working, but the network connection is still live, it's reasonable to assume that the connectivity function isn't to blame. 

Overall, there's a lot of opacity here - nobody really knows how to deal with network responsibilities, in a world where there are no clear demarcation points. It's a set of lessons that will need to be learned very soon - and which will probably involve regulators or other authorities in disputes.

A new "killer app" for WebRTC? Video-identity verification / know-your-customer

I've recently come across an interesting new use-case for WebRTC, which I think could be a major driver of adoption - because it has a very clear monetisation model. It may also be a catalyst for growth of video-based contact centres, beyond the Amazon Mayday-style customer support scenario.

I'm referring to the use of video-authentication for web/mobile transaction security. This is also called “Online Legitimation” and is already widely used in Germany, especially for major financial services transactions such as opening of accounts, or setting up loan agreements, which have “know your customer” requirements.

The user is directed to a short two-way video session with a customer-service advisor, allowing verification of the user's identity against a photo on file, or by holding an ID/passport up to the camera. A code is then sent via SMS for two-factor security. (The latter approach has been approved by the German Federal Financial Supervisory Authority as an alternative to conventional offline legitimation procedures). 

This is very different to a normal person-to-person videoconference or "see what I see" application. It is using a video session (not a "call", really) to compare the likeness of two things (as well as conduct other vocal checks through a Q&A process).

At least four providers have already emerged in Germany (email me for names), and a large number of banks and credit-card companies have signed up with one or other of them. One of them has a price-list suggesting prices of between €4-8 per transaction, depending on volumes expected by the client. As well as WebRTC (desktop and embedded in mobile apps), there are sometimes options for Skype, FaceTime, fallback to Flash and so on.

The regulatory approval in Germany has clearly catalysed a sudden market uptake – it will be interesting to see if this is replicated elsewhere. This has the hint of a “killer app” for video contact centres, especially in finance but potentially also for other sectors such as online casino sign-ups, contract-witnessing, notarisation or other applications where visual ID approval is needed. 

This also fits very well with the move of many services online/mobile channels. It would be ridiculous to require an in-person identity check or notarisation, to set up a new online-only bank account. 

The precise number of relevant transactions requiring "hard" identity proof is a little hard to estimate - but a rough number of a billion annually does not seem unreasonable as a potential addressable market. There is a good argument for WebRTC solution providers to consider lobbying finance regulators or other bodies, to get them to clearly accept video-based "know your customer" processes.

Disruptive Analysis is about to publish an update for its WebRTC Market Analysis & Forecasts report, including analysis of various new & emerging use-cases such as this. Please get in touch to receive details & pricing (or to arrange a private workshop or project on WebRTC) information AT disruptive-analysis DOT com

---

What's YOUR view of contextual communications?

In recent months, I've been drilling into the new "hot topic" of contextual comms. Martin Geddes & I are so enthused by the topic that we're running a workshop on June 15th in London (details here), and we're already considering follow-ups, maybe in the US later in the year.

We're combining both the "here & now" of context with a view on where we might be heading in the medium-to-longer term. Martin wrote a very forward-looking and provocative piece on the possible future recently (here).

I'm really interested in what "contextual communications" means to everyone else. There's no fixed definition at the moment, and I suspect that we're going to get an "Olympic Rings" multi-way Venn diagram. Some views of context will overlap, while others will be miles apart. For instance, I've seen or heard all of these described as Contextual Comms:

• Sending web-form info to an contact-centre agent during "click to call"
• Embedding video/telepresence into a robot
• Using mic & speakers on a phone to map out a room acoustically & tweak the echo/noise processing
• Use a media-server to analyse a caller's tone (eg angry vs. happy) or facial expressions, and adjust the experience or script for a salesperson
• Using a device orientation sensor to work out if a phone is flat on a table, or help to the ear, and adust the UI accordingly
• Using machine-learning and analytics to assess the best time to call someone
• Mechanisms for indicating the purpose of a call
• Embedding a call into a timeline or activity-stream interface for UC and collaboration, so it can be recorded, captured & seen alongside text commentary or speech analytics

I'm sure there are dozens more as well. I'm looking forward to distilling some sort of map or ontology, so we can collectively understand this new landscape a bit more clearly. Is it one thing with lots of variants? Or 5 separate trends with a little overlap?

Do YOU have a good example or definition of Contextual Comms? I'd love to hear from you, either via a comment here, or by doing an interview briefing.

And if you'd like to talk about it publicly, we're offering all the workshop attendees an chance to present or demo their view - basically an "open mic" section of the day to showcase their unique take on context.

If you'd like more detail about the event, or to get in touch separately about context, please comment,  see this page to book a spacea, or email information AT disruptive-analysis DOT com.

Quick thoughts from WebRTC Expo last week

I've already written about the telco aspects of the big WebRTC conference & exhibition last week.

This is just a quick summary of some of the other things that have stuck in my mind. It doesn't cover all the companies or trends I saw - a more holistic analysis will be included in forthcoming updates of my paid WebRTC research. Also, there were sections of the event (in multiple streams) that I couldn't attend personally - hopefully there will be other articles on healthcare apps, for example.

Top of mind themes emerging are:

• Customer service & call centre use-cases are everywhere, and definitely the most "commercial" part of WebRTC at present. The show featured a lot more integrated solutions, plugging WebRTC into existing platforms & industry structure - and adding in the power of the web in mashups for both agents & end-users. Various vendors had ways to analyse WebRTC speech, annote the agent's information via a browser etc. It all has an aura of "realness" around it, especially for finance and retail sectors.
• There's a whole range of other actually-deployed services, especially around distance-learning/training, video chat and conferencing. That said, a lot of UC and collaboration use-cases are more "now WebRTC-enabled!" rather than "designed from the ground-up for WebRTC".
• There's still a bit of a gap for innovative "pure WebRTC" use-cases that don't lean on traditional forms of communications, or just move beyond legacy telephony APIs. There's some fun music-jamming things, or a handful of games, but I haven't seen a (desktop) WebRTC-based equivalent of Chatroulette or anything clearly "viral". I'm also a bit perturbed that the adult industry seemingly hasn't bothered to turn up to the WebRTC party - usually it's first in line for any new web technology, and its absence is a bit of a puzzler.
• Where WebRTC innovation does turn up, it may well be in specific industry verticals. We know about finance and healthcare and real-estate demos and deployments - but I'm starting to wonder about government, media, utilities and so on. That said, an interesting chat at a separate event yesterday opened my eyes some possibilities around public safety.
• Many of the other use-cases are going to be heavily mobile-centric. This implies a need for 3rd party APIs/SDKs/platforms, as few consumer mobile apps use the browser, even on Android devices where WebRTC can be properly supported. On iOS, a 3rd-party approach is almost mandatory to embed voice/video into native apps. Purists complain that it "isn't proper WebRTC", but that's an irrelevant technical religious debate and doesn't detract from the opportunity. One company (Imagination) was even suggesting to rip out the Google/GIPS media engine from WebRTC and use theirs instead, as it's apparently mobile-optimised and can deal with the acoustic and silicon idiosyncrasies of smartphones better.
• Actually, the WebRTC purists (especially the "web" bit) were wincing a lot last week. Plug-ins are back with a vengeance, with at least 3 companies offering products. Temasys and Priologic had a bit of a barney about closed-but-free vs. open-source. I'm not a security specialist, but I can appreciate why tighter controls over what accesses the camera/mic are desirable. In my view, the purists have a lot more wincing ahead, as even if Apple & Microsoft adopt WebRTC (see below), both have millions of legacy version users who don't update.
• In a nutshell, RTCWeb (the IETF standard for nuts-and-bolts protocols) is looking ever-more important than WebRTC (the W3C API exposing all that stuff to web developers). Frankly, making the technology of nasty, complex bits of realtime communications infrastructure all work nicely, is rather more important than the precise way it's presented to developers. The "ends" of democratising voice, video & data into apps and websites and devices easily justify occasionally fudging the "means".
• For companies or developers that don't have longstanding communications expertise, the route to adoption of WebRTC is likely to be via one a platform / WebRTCaaS provider of some sort. There's a lot to choose from - some voice-centric, some video-oriented, some integrated with existing call centres or messaging, some more interested in QoS / scalability, some more around pre-built apps, some more mobile-oriented and so forth. Each developer will have to navigate a long list to see which are the best fits with their specific needs.
• Part of the problem is that some people read "MS & Apple don't support it" headlines and don't look beyond that. Most people I speak to don't realise what it is that all the WebRTCaaS API/SDK providers like Tokbox & Temasys & Requestec etc are actually doing. I think there needs to be a neater "category" and some serious evangelism - at the moment, I get the impression that the standards guys are a bit irked by their relevance and growing importance, just as they are with the SBC/gateway proliferation rather than P2P use-cases.
• The big discussion about silos vs. integration with legacy systems, via gateways, is largely confined to voice-primary uses, given that's what legacy telephony, UC and contact centres do almost exclusively. Video communications today is basically enterprise conferencing, plus Skype/FaceTime/Hangouts/consumer mobile apps. That's not to say that video-gatewaying won't be more important in future - and will likely need a lot of heavy-lifting infrastructure - but today, interop is about plain-old phone calls.
• Microsoft gave a presentation which (in a nutshell) said that (a) it's deeply engaged in this area, but (b) don't expect IE to support WebRTC 1.0 any time soon, except GetUserMedia APIs. There's a long list of other HTML5 bits & pieces waiting to be implemented by the IE team, so WebRTC has to battle for engineering resources against all of those, as well as undoubted pressure from Skype & Lync teams. However it appears ever-more enthusiastic about ORTC, which is showing so many signs of forming the basis of WebRTC 2.0, that various of the 1.0 advocates are getting a little irked at its jumping the gun.
• Various people were pessimistic/optimistic about Apple. "They won't bother", "But they joined the working group", "They'll wait for the standard" etc. In essence - nobody knows, everyone's got a theory. And we'll probably all be wrong. Happy to give you my own guesses if you want....
• Datachannel was around more as a "supporting actor" than as a headline act. Various use-cases had messaging or co-browsing or screen-sharing incorporated, but there seemed to be relatively few "pure" datachannel use-cases. I didn't see much M2M or IOT stuff around. A bit disappointing
• I think some people are getting impatient and expecting things overnight. It is only literally 3 years since WebRTC was even spoken about publicly - I remember it from the eComm in San Francisco in June 2011, the day before my & Martin Geddes' first Future of Voice workshop in Santa Clara. It's worth noting that with Snapchat/AddLive & the probable imminent arrival of WebRTC Hangouts, as well as various customer service functions, we're probably not far off 100m active users. Much as I hate  "Fastest growing web/comms thing EVER!!!!" articles, I could probably imagine writing one in a few months' time. Let's just put it this way - I expect there are already more active WebRTC users than RCS or VoLTE users, and it's certain to stay that way.

Overall, I'd say the Expo was more "grown up", with real use-cases, but perhaps a dearth of truly surprising "Web whiz-bangs". And I still haven't seen a WebRTC-enabled drone, despite the cavernous hangar-like venue used for the event being perfect for one.

Disruptive Analysis WebRTC Market Status & Forecast Update - March 2014

Disruptive Analysis has been covering WebRTC since June 2011. And it is now just past the one-year point since it published the industry's first comprehensive analyst study and forecast covering the entire WebRTC value chain, in February 2013. Comprehensive update documents were issued to subscribers in June & October 2013. 

The third revision has now been published, covering recent trends in use-cases, standards evolution and WebRTC industry structure - as well as updated forecasts.

Over the last year, a few other analyst reports have been published on WebRTC. But none has really covered all aspects - enterprise, telecoms, consumer web beyond, with both qualitative and quantitative input. Much of what has emerged has stuck to the original 2011-2012 narrative about WebRTC as being just "like Skype but in the browser", rather than examine how WebRTC is changing into a much broader set of propositions. 

In the October 2013 update, Disruptive Analysis identified the trend towards non-browser WebRTC, especially on mobile devices. This has accelerated and is now an undeniable part of the landscape. Indeed, it is the core driver of recent headline service launches, including the consumer-oriented Tuenti and WeCam launches just this week, as well as the B2C customer-service interactions seen in Amazon's Mayday (which is part-WebRTC) and American Express' new iPad app enabled with video-chat.

These developments reflect both positives and negatives about WebRTC's evolution. At one level, there are issues with the lack of IE/Safari support, and continued debate over video codecs. Security and firewall/network middle-box traversal (in some instances) remain issues being addressed by IETF.

But what offsets these problems is the large and growing emphasis on "getting on with it anyway", in pre-standard form, and often using cloud platforms and 3rd-party SDKs to embed WebRTC into mobile apps, standalone PC applications and yes, even plug-ins. This is inevitably slowing down some use-cases, while speeding up others.

This translates into some modestly-lowered forecasts for PC support and adoption of WebRTC in 2014, but increased mobile support, especially from 2015. That is underpinned by both a growing array of "WebRTC cloud enablement" providers such as Tokbox, Twilio, Weemo, Temasys and about a dozen others, as well as a sudden surge in native or browser-based support of WebRTC on Android. 

By the end of this year, a large % of new Android devices (phone and tablets) will ship WebRTC-enabled "out of the box" - some in multiple different ways.

In terms of use-cases, the high-profile emergence of Amazon’s Mayday customer-support button on its new Kindle devices has catalysed huge interest in replicating it on other platforms. While only a small proportion of call-centre agents will transition to video initially (mostly for complex high-value customers/products), the economics and business processes will slowly adapt over time. Elsewhere in the enterprise space, WebRTC is cropping up in more UC and conferencing contexts, although as yet, we haven't seen a Mayday-comparable "cheerleader" which moves the market almost overnight, but that may come later this year.

In the consumer space, it seems that mobile/social adoption of WebRTC is finally starting to occur, but primarily on mobile rather than desktop. To a degree this reflects both browser-support issues, but also means that efforts are concentrated around the hot-as-the-sun communications app space. The Viber and Whatsapp acquisitions have just turned the focus on the area that WebRTC was heading towards anyway. Disruptive Analysis expects quite a lot more consumer-centric WebRTC apps to appear this year. It would be unsurprising if at least one "went viral". This could bring 10's or even 100's of millions of users overnight, dragging 1000's more developers in its wake trying to emulate it. There are also interesting desktop uses of WebRTC, including video-chat, but also extending towards defined verticals, such as healthcare and other forms of online consultation.

As predicted, telco use-cases of WebRTC have been slow to become real, especially where network-integration work with IMS is concerned. Although 3GPP is working on standardisation and a number of vendors have announced contracts, it seems likely that “live” commercial services will suffer protracted development, testing and internal-process cycles before launch. Most mobile operators have enough problems simply getting VoLTE to work in normal fashion, and so WebRTC enhancement or extension is not a priority. RCS is another "problem child".

Those telco-related WebRTC efforts which have  emerged into the light of day so far – a total of 4 by Disruptive Analysis’ count – are solidly in the “OTT” camp at present, including developer platforms. That said, there is clear interest from the service provider community, based on announcements, conference and web appearances of telco representatives, and Disruptive Analysis’ own report sales and consulting engagements into that sector. Other service providers from background such as VoIP and web-hosting are starting to appear as well.

As well as the established use-cases, other WebRTC domains such as the data-channel CDNs discussed recently, as well as Google's Chromecast dongle, are also starting to get traction. It seems likely that the overall market will continue to expand in scope as well as scale.

Highlights from the new forecasts:

• Devices supporting WebRTC at end-2014 reduced from 1.7bn to 1.6bn
• Devices supporting WebRTC at end-2016 increased from 4.2 to 4.7bn
• Mix of devices supporting WebRTC skewed towards mobile from PC, and from browser-based to non-browser
• Active WebRTC end-user base (individuals) at end-2016 increased from 1.7bn to 1.8bn

Additional detail and methodology/assumptions is available to the report's subscribers in the full document. 

Charts are available for companies wishing to use the data in presentations or marketing material, subject to sourcing to the Disruptive Analysis WebRTC Report. 

New customers buying the WebRTC study will receive both the original report and this March 2014 update, plus an hour's conference-call to discuss the current state of the market. 

Payments can be made either by card/Paypal online below, or by invoice/bank-transfer. Please inquire via information AT disruptive-analysis DOT com    

Additional ongoing subscription packages and/or briefings and workshops are also available. 

WebRTC report & Q1'14 update (PDF) 1-3 users

Non-UK/EU $1,500.00 USD UK/EU (+20% VAT) $1,800.00 USD Non-UK EU (only if you send EU VAT# via email) $1,500.00 USD

 

WebRTC report & Q1'14 update (PDF) Corporate Licence

Non-UK/EU $2,500.00 USD UK/EU (+20% VAT) $3,000.00 USD Non-UK EU (only if you send EU VAT# via email) $2,500.00 USD