Nearly a decade ago, Communications Platform as a Service (CPaaS) was relatively new— even as a concept. Communications services were driven by infrastructure that often took the form of software designed to run on a physical server or even proprietary hardware. Most businesses were beginning to think about how they could move some of this infrastructure to the cloud as they had been doing with other IT systems.
Unlike databases, email servers, and other software that lead cloud-migration projects, communications software is real-time, bandwidth-sensitive, ephemeral, and interactive all-in-one. These unique aspects of real-time communications mean it often lags other IT trends by years, but not indefinitely. Like other cloud technologies, CPaaS has gone through several stages of development.
CPaaS 1.0 - faster deployment for developers
The Platform as a service (PaaS) concept emerged as a way for developers to get their code to production faster. With PaaS, developers didn’t need to worry about server management and on-going maintenance and could solely focus on this task. CPaaS emerged to do the same for developers and their communications needs.
In this first phase of CPaaS, the communications interactions developers required, tended to be simple, traditional telecom actions—answer a call, play a prompt, respond to a touchtone key press, or send a text message. Previously, these capabilities only came as part of large, complex systems, so the ability to easily access specific communications was an enormous boon to developers. These early successes helped to cement new business models like the sharing economy that couldn’t have existed without that flexibility.
CPaaS 2.0 - going mainstream
As CPaaS started to prove itself, developers started asking for more communications channels, product features, supported platforms, developer tools, and global coverage. For developers, this demand also meant supporting additional programming languages and frameworks - React, Kotlin, Flutter, and Swift area few examples.
Voice communications also needed to span the traditional phone network (PSTN), IP-based SIP communications systems, and web-based communications with WebRTC. Messaging expanded beyond short message service (SMS) to various forms of IP Messaging that could happen independently of the phone network.
To make CPaaS more accessible, support was added for web and mobile platforms on a mix of languages and development frameworks. Businesses are global, so CPaaS platforms needed to be global too—with local phone numbers and specific regional support.
As this technology matured, businesses not considered “developer-focused” started to look at CPaaS to transform their communications capabilities and opportunities. CPaaS moved beyond a tool start-up used to become the new standard for solving communications challenges and driving business outcomes within larger organizations. Some examples of new business functions interested in CPaaS included:
- Product teams looking to improve their apps
- Marketing organizations needing more effective ways to reach their audience
- IT teams searching for better ways to enable internal communications
- Contact center teams looking for more innovative approaches to customer interactions
However, adding more features and new capabilities on top of the same, aging paradigms promised to become messy and brittle.
CPaaS 3.0 - communications transformation
CPaaS has now moved beyond a wish-list of nice-to-have concepts to a check-list of must-have functions. Meanwhile, the broader computer science community continues to make rapid advancements in applying machine learning to various applications. Tooling for the cloud has exploded, and CPaaS has been, is, and will be a beneficiary of these broader computer science trends.
Moving forward, three trends will likely dominate the evolution of CPaas:
- Development simplification
- Omni-channel customer experience
- AI and machine learning
Simplifying the development experience: serverless and no-code
New CPaaS features won’t add much value if the related apps are too difficult to develop. CPaaS focused on APIs in the early stages, which made sense as most developers had solely focused on getting something to work. However, over time these systems evolved into large-scale production systems, and the focus has shifted to on-going maintenance and quality. CPaaS now needs to not only help rapidly deploy new functionality but also efficiently maintain it.
To address these new needs, serverless and no code capabilities are increasingly important.
Serverless - the way to efficiently scale
One of the main product concepts to come out of the PaaS approach was serverless cloud functions. Products like Amazon Web Service (AWS) Lambda or Google Cloud let developers deploy small programs to the cloud with no infrastructure concerns. These serverless platforms automatically scale and remove the cost of ownership. To help with rapid development and easy debugging, they have also evolved into more comprehensive Integrated Development Environments (IDE’s), allowing developers to easily code, debug, and control deployment without having to leave the platform. Like the PaaS platforms that inspired them, serverless has found its way to today’s CPaaS market.
No-code - the new way to develop
As astute developers have realized over time, more code isn’t the best way to solve a problem. This issue became evident as the amount of code to be managed increased rapidly with CPaaS use through many applications within the same organizations. To help address this, no-code builders now allow both developers and less technical users to create programs by dragging and dropping visual blocks. These kinds of systems already existed for building basic IVR applications. However, modern no-code solutions go far beyond IVR needs and allow the development of multi-channel apps that quickly integrate and interact with external systems. The best no-code tools come with their own troubleshooting and debugging features. These systems appeal to non-developers and expand the market for communications platforms, but the best allow some mixing of no-code, low-code, and full coding to give developers flexibility where they need it while reducing complexity wherever possible.
Communicate with customers wherever they are
Originally, phones only had voice calling and SMS messenger. Now, the typical smartphone user’s screen comprises many communications apps—dialers, video conferencing apps, and dedicated messengers, not to mention various calling and messaging tools that might be embedded inside other applications. It shouldn’t be a surprise that users have grown accustomed to using multiple communications channels according to their personal preferences.
Businesses must care and think about the communications channels that best meet their engagement needs. That often means moving to a multi-channel approach where any channels can be mixed and matched - this refers to “omnichannel” in contact centers.
More than Voice & SMS
The public mobile phone network has been capable of voice communication and SMS for some time, so media support has become a prerequisite for any CPaaS device. Business communication infrastructure is largely based on technologies like SIP, so technology support is also required for any business-focused platform. More importantly, newer technologies like WebRTC have matured to become the de facto mechanism for newly developed communications applications. WebRTC continues to evolve, and modern CPaaS providers must continually invest in this technology to stay current.
Besides voice channels, WebRTC has made it easier to embed real-time communications into any web or mobile application. The user interface (UI) aspects of a video are different from voice and introduce new complexities for CPaaS. Support of different modes of video communication, including peer-to-peer for optional performance, multi-party conferencing media servers for larger groups, and real-time streaming infrastructure for live broadcasts, are all necessary to fulfill growing video requirements.
More Messaging Platforms
IP-based messaging is a critical service for affordable user-to-user communications within an app environment. While SMS is slowly maturing into a rich communication services (RCS), the majority of chat interaction has shifted to a new breed of messaging apps and users are downloading more than SMS. Each platform comes with a unique API and business relationship requirements, so the CPaaS must simplify these differences into a unified messaging product. These new messaging API’s may also have new or unparalleled capabilities. For example, the WhatsApp Business API provides a special interface for businesses to interact with their customers using that messenger.
With the growing number of channels comes an increasing number of technical identities. PSTN-based voice and SMS may have had a shared phone number, but now video calling, and IP messaging require unique user identifications. Many technical identities exist behind the scenes to support the various technologies driving a given communications platform, and users generally share a single identity with the businesses with which they interact. However, users also expect coherent interactions that refer to past communications had with the company - no matter what communications channels facilitated those discussions. Progressive CPaaS platforms need to facilitate the unification of these identities to provide the most useful omnichannel experience. This includes simplifying the mechanisms necessary to communicate with multiple channels.
Leveraging AI and machine learning
Advances in machine learning are bringing better forms of artificial intelligence to all kinds of services and products. For communications, this is manifesting itself in two major areas: voice interaction and bots.
Human-like voice interaction.
Speech recognition technology is commonly used in telephony to respond to spoken user input and provide automated analysis in call center environments. Mass adoption of this technology was once limited because of high costs, little language support, and poor accuracy. Recent advances in machine learning techniques have dramatically improved accuracy while lowering costs. These modern AI-based methodologies are also more adaptable—making speech recognition available for more languages and regional dialects. Besides transcribing speech into text, these modern platforms include advanced data structures for better use, options for improving transcription in specific environments, multi-lingual support, and complementary capabilities like sentiment analysis, and sometimes additional paralinguistic analysis.
Similar machine learning technologies have also improved speech synthesis. IVR systems that wanted to avoid robotic and annoying synthesized voices used to resort to tediously recording every prompt the system might need. Modern speech synthesis, like the neural models offered by Google’s DeepMind or AWS, sound incredibly lifelike. These systems continue to improve in performance and variety, with more distinct voice prints and broader language support. CPaaS should support multiple speech platforms so developers can best optimize them to their particular needs.
The science fiction of verbally interacting with a robot is now a reality. Rather than human-like androids, these robots take the form of smart speakers and smartphones. Juniper Research estimates voice assistants are available on 4.2 billion devices today, and that figure will double by 2024, exceeding the world population. “Thanks to these devices, users are growing increasingly accustomed to bot interaction, and they expect them to work well,” the company said.
To meet user expectations, CPaaS must make this technology available to all communications applications while also making it affordable for SMBs, not just enterprises. Fortunately, the massive global scale of virtual assistants like Amazon’s Alexa and Google Assistant has made their underlying bot engines economical. Now it is up to CPaaS to add these technologies to their voice and messaging platforms and continue to improve this support for the latest in human-like interaction.
Future-proof communications have arrived
No one can predict the future (yet), but many of the broader trends in computer science and human communications are clear. CPaaS has responded. New development paradigms like serverless and no code have started to redefine the category. Omni-channel communications expectations will also accelerate the need for communications platforms to adjust their interaction models. And finally, AI, speech, and bot interaction, once novelty features, will soon be mandatory requirements in a communications landscape that will demand more robust and cross-channel functionality than we could have imagined.
The goal and the result will be improved customer interaction on par with businesses’ necessary pursuit of efficiency to meet modern communications needs in a dynamic pandemic disrupted world.