Twilio started out as a simple communications service, offering an API that allows developers to enhance apps with the ability to send and receive texts or voice calls, and has grown its customer base to more than 40,000. One of its flagship customers has been Uber, where for a long while Twilio has handled all the phone interactions between Uber, its drivers and riders. But more recently, the company has been expanding the capabilities of its platform in order to broaden its appeal to enterprise developers.
Intrigued by the latest announcements at its Signal user conference last month in San Francisco, I took the opportunity to catch up with VP of Product Pat Malatack on a recent visit to London. Before we dug into the details of the new announcements, he started out by outlining Twilio's mission as a universal communications service, which he likened to the role of Amazon Web Services (AWS) as a universal compute service:
What AWS is for compute and storage, Twilio is that for communication. It's a programmable API that you can include as part of any application or service.
We look at the world from the perspective of, what are the modalities that humans and machines communicate with other humans? We want to make sure that we have all of those modalities on the platform — voice, messaging, both SMS and chat, and Facebook Messenger support. We have video support on the platform. Basically every modality of human communication, we have a cloud platform that supports those communication types.
In other words, Twilio takes care of all the mechanics of plugging into these various communications channels, and presents its services to application builders conveniently packaged as an API. Developers can then call on the Twilio API either to surface messaging inside their application — for example to provide help via chat or video in the middle of completing an action — or to use messaging to deliver information or functions to people without them having to fire up the application.
Yet another engagement channel
This second use case, of reaching out beyond the confines of an application to wherever the user happens to be, is something I've previously written about as part of a trend towards 'headless' applications, which the user never has to directly visit. But Malatack positions it as just another extension of the proliferation of engagement channels available today:
The number of channels that you need to communicate with your customer in, as a business, is exploding. A decade ago, it used to be you just have the toll-free number, folks could call in, and talk to your customer support center.
Now, you have to message them — you have the toll-free number, that didn't go away — now you have Messenger and other platforms you need to support, now you have places like Slack, you have places like increasingly now Alexa, Twitter is another great example. You have this explosion of points in which customers can interact with your brand or your company.
That ability to meet customers wherever they are has been pioneered by software-first businesses like Uber, Airbnb and others, but now it's becoming a mainstream requirement, he says.
Every business is becoming a software company, and so we transitioned over the years to power a lot of the software experiences for retailers like Dixon's to Macy's and Nordstrom in the United States, in the retail sector. In financial services, we power ING's customer support center.
We've expanded from the software-first organizations to really any organization that wants to engage with their customers via software.
While larger enterprises have been adopting Twilio directly, others take advantage of its services embedded into partner applications, including Zendesk, Topdesk, NewVoiceMedia, Serenova (formerly LiveOps), ServiceNow and Salesforce, which is also an investor. Now the provider aims to broaden its enterprise reach, with the launch last month of the Twilio Engagement Cloud, which bundles its APIs with pre-built business logic.
Patterns of engagement
What Twilio has done, explains Malatack, is it looked at how its established customers have been using its services, and then bundled up the most common patterns into pre-built components that make it quicker and easier for enterprise developers to incorporate those patterns into their own business infrastructure. While some of these components relate to security issues such as two-factor authentication and verifying the identity of new subscribers, others deal with patterns of engagement.
The three most common customer engagement patterns are with automated systems, with specific departments or functions, and with individuals, such as a delivery driver, an engineer on a call-out, or a dedicated financial advisor. The engagement cloud offers services for each of these three use cases:
- Notify handles system-level engagements with customers, providing alerts and notifications across multiple channels, including iOS and Android push, SMS, Facebook Messenger and Twitter DM. A unique capability, says Malatack, is its support for segmentation and orchestration, which means it can automatically ensure each customer is notified via their preferred channel.
- TaskRouter manages the workflow surrounding interactions between customers and specific departments. It keeps a record of various parameters so that, for example, it can automatically match an incoming customer service enquiry to an agent with appropriate product knowledge and language skills.
- Proxy maps communications between two individuals without either of them needing to know the other's number. This enables a delivery driver or service engineer to have a phone call with a customer about a specific visit while keeping both phone numbers private.
All three types of interaction will often occur within a single event or process, Malatack points out.
With systems, department, individual, what we see is that the customer journey moves across all three of these. It's usually not one or the other. If you think about a package delivery, you have the notification of the delivery, if there's a problem I want to contact DHL or FedEx, but if the package is out for delivery to my house, I want to talk to the driver in the car. All three at all points in time.
The advent of intelligent bots and conversational agents such as Alexa, Cortana and others will add another ingredient to the communications mix, he believes, rather than eliminating any these existing options.
The business problem isn't, 'I need a chat bot.' The business problem is, 'I need an efficient way to engage with my customer effectively.' Some of those are going to be automated, some of them are not ...
We see chatbots as a tool to help make those experiences great, but it's actually just part of that customer journey. Some things I want automated, some things I don't. Giving the customer the right experience at the right time is really the business aim here.
The newest addition to the Engagement Cloud uses machine learning to apply natural language understanding (NLU) to parse what people say in a voice call or a message to work out what they want. Developers will be able to plug this capability into automated call answering, as well as chatbots, to replace fixed menu systems with a more natural conversational interface. This will be as big a transformation for automated customer engagement as when finding things on the Web went from structured directories to keyword search, says Malatack.
Where it used to be an IVR, press one to talk to support, press two to do this, it's like no, just tell me what you want to do ...
It really changes the paradigm of what it means to interact with a company. You can jump directly to where you want to go in one line, one spoken word, right? That's sort of the beauty of search, sort of what search did to navigation. We're seeing the same thing happening in many of these other customer interactions.
And of course the system should also recognize the right moment to put a human being on the line, rather than continue with an automated interaction. It's just another example of how the technology is helping enterprises transform the customer experience, which is what Twilio is really all about, Malatack concludes.
What we're seeing is the companies that are going to win in the market are the ones that build great customer experiences at really the most critical moment that you're interacting with their brand — when you're actually trying to talk to them.
Everyone's talking about conversation. Walmart eCommerce CEO Marc Lore was recently talking up conversational commerce. Here on diginomica, Denis Pombriant has been exploring the notion of conversational CRM, while Salesforce recently posted some thoughts on conversational customer service. Meanwhile, as mentioned above, I've been having conversations about the trend towards conversational apps, while Kurt Marko's explored the emergence of conversational hardware.
Does anyone notice a trend here?
Conversation is becoming the new frontier of how people interact with computing, and that's likely to have a huge impact on every kind of enterprise application in the coming years — not only the customer facing ones, but also those used by an increasingly mobile workforce (especially in those roles where people tend to work with their hands doing something other than tapping on a keyboard).
Twilio is thus well placed to play a big role in this emerging trend, having secured its place as a leading provider of on-demand communication services. Its introduction of the Engagement Cloud platform is well timed to help consolidate that role.