Speech Technology Magazine

 

Speech Providers Embrace the Cloud

The forecast is bright for hosted solutions.
By Michele Masterson - Posted May 1, 2013
Page1 of 1
Bookmark and Share

The cloud has been a ubiquitous offering for many segments of technology for some time, and speech solution providers are no exception. Growth in this arena is continuous and offers many benefits, including scale, flexibility, lower barriers to entry, and cost savings.

Opus Research reported enterprise spending of about $1.5 billion on hosted speech in 2012. Many hosted contact center service providers have seen super-normal growth, reporting gains of as much as 35 percent higher than in the past.

"Every solution vendor with speech-enabled IVRs has added a cloud-based, virtualized, or SaaS [software as a service] version. Many of the contact centers in a cloud have an IVR option to their offering, plus there are the classic providers of hosted automated speech services," says Dan Miller, Opus Research senior analyst and founder. "Rapid growth is the result of enterprise spending coming out of a deep freeze, combined with the fact that it's easier to leverage investments made in Web-based marketing, sales, and support services, which are largely cloud-based."

The Case for Cloud: Cost Benefits

Companies are always looking at the bottom line, and the benefits of keeping costs down with speech cloud solutions are indisputable. There are usually massive upfront costs associated with on-premises solutions.

"The cloud makes a lot of sense from an economy of scale perspective," says Nigel Ferris, solutions architect at Voxeo. "There are certain services that we can provide in the cloud because we provide them for a lot of customers, whereas maybe it wouldn't make sense for a company to take its own initiative."

"Oftentimes, the cost to procure and maintain complex technologies, the variability of user traffic, and the need for agility will be the primary drivers toward hosted," says Dena Skrbina, senior director of OnDemand Solutions Marketing at Nuance. "With cloud-based solutions, companies don't have to incur the expense associated with purchasing [solutions] and then operate redundant configurations or complex infrastructures based on worst-case traffic scenarios that may or may not occur. They don't have to update or upgrade technology. Instead, they simply pay for what they use when and if they use it."

Cloud-based speech solutions can also be a great equalizer, enabling businesses' Davids to be on the same playing field as Goliaths.

"Not only can huge organizations have good applications for their customers without breaking the bank," says Judith Markowitz, president of J. Markowitz Consultants, "[but] you can have small and midsized companies do the same things in the cloud."

In addition, for some companies, the decision to go the cloud speech solutions route comes down to the basic premise of capital expenditures (CAPEX) versus operating expenditures, and whether or not it makes sense for the cloud to do all the heavy lifting.

"Sometimes it's a CAPEX decision that causes customers to look for hosted speech. Instead of the cost of buying a speech analytics solution deployed on-premises, of having to involve IT, having to work through all the logistics, creating indexes and files, and maybe having to add an analyst who can interpret the results or create new search patterns, the attraction is to be able to pay for a service that does all that for them," says John Bourne, senior vice president of global channels and alliances at Verint Systems. "Conceptually, [cloud] is similar to having a consulting service, where another firm does the work to uncover valuable insight and information."

Operating speech solutions via the cloud also offers cost benefits in terms of scale and staying on the cutting edge without racking up additional charges.

"You have a lot of innovation and scale of possibilities that you gain by being in the cloud," says Kelly Weinhold, product strategist at Angel, now owned by Genesys. "With the cloud, you have one product where you have constant innovation and updates. Customers get to have access to those features as soon as we release them, so that's a big pro for using the cloud. It's a seamless upgrade. The cloud also provides a lot of flexibility. We can integrate changes fairly expeditiously."

Consider, too, that the more technologies in the cloud are being used, the better they get. For example, speech recognition has seen boosts from compiled data in the cloud.

"Speech recognition is a technology that benefits greatly from data," Skrbina says. "In this case, data is the recording of the words and phrases spoken by the caller. As millions of callers speak to the system across thousands of applications, the data improves the speech models, which, in turn, improves the performance of the technology, which improves the caller experience. This can happen at a large scale and rapidly on a shared service."

On-Premises Versus Cloud

While there are many reasons for speech-based solutions to be accessible from the cloud, there are also considerations for making the case to stay on-premises.

"It depends on what the organization plans on doing with [a solution], how they plan on using it, upgrading it, and maintaining it," says Donna Fluss, founder and president of DMG Consulting. "If an organization plans to continue to innovate and take advantage of new releases, keep it current, and continually optimize it [that's a case for cloud]. A good IVR, for example, needs to be optimized on a quarterly basis. You shouldn't be putting it in and ignoring it."

When weighing the pros and cons of going from on-premises to cloud or having a hybrid model, the complexity of the IT environment comes into play, says Sean Murphy, director of product marketing at Utopy/Genesys (Utopy was acquired by Genesys earlier this year).

"Because so many contact center environments are heterogeneous and spread all over the world, it comes back to the same key values but in a unified way," Murphy says. "For example, as we know, a lot of contact centers outsource some of their calls these days, so providing a solution in the cloud really offers a unified analysis of all different areas of a business, even if the calls are outsourced. If it's on-premises, you can still get that done; it just takes a lot longer and there's more complexity involved."

In 2012, Voxware released what it says is the industry's first voice-picking cloud solution. In January, the company rolled out its Cloud Voice Management solution for the enterprise.

"We see a tremendous amount of interest in the cloud, and I think that the reason is that many companies have a variety of different IT environments and resources available," says John Casagrande, vice president of client services at Voxware. "We've found that with cloud implementations, they're generally faster, even faster than our standard on-premises implementations, and customers are able to realize benefits more quickly. For certain IT environments that have their own IT policies or have made their own investments in IT infrastructures that they already have in-house, on-premises can be advantageous. That said, I think more and more customers are becoming aware of cloud and the benefits it can provide."

Additionally, when companies add higher value applications that are more sophisticated, such as speech analytics, integrations with CRM systems, and other back-end systems, the case for cloud solutions becomes stronger, says Keith Dawson, principal analyst at Ovum. However, companies should take care to have a holistic view and look at all aspects of the cloud, on-premises, or a hybrid solution, and, for example, not just base a decision on cost.

"It's more based on a sophisticated analysis in the long-term compatibility of the vendor, the ecosystem in which the cloud vendors are working together, sometimes in collaboration with on-premises vendors," he says. "It's become a much more nuanced kind of decision. I think that in the long run, it tends to make for [a] better kind of planning decision rather than the kind of knee-jerk responses that say, if I want to save money in the short term, I'm going cloud, or if I'm looking to preserve my system independence, I'm going to go on-premises."

Mazin Gilbert, assistant vice president of technical research at AT&T, agrees that the consideration about whether to go cloud, on-premises, or hybrid is not a black-and-white decision.

"If you had spoken to me a few years ago, I would have given you a very different answer about embedded intelligence," Gilbert says. "Even though we've invested a significant amount of time and effort for decades on our cloud-based solutions, the way we're really developing our AT&T Watson speech technology capabilities is a hybrid solution, where the intelligence and the technology sit both embedded and in the cloud. The world is changing. We're betting on this notion of hybridization. As far as the customer is concerned, he just wants a seamless, user-friendly experience. If you're sitting in a car and you're talking to a virtual assistant, obviously you don't really care where the technology resides. What you want to make sure is that your transaction and information is executed. At the end of the day, we're trying to support a connected life experience."

There's also the issue of security worries, and the basic premise of taking sensitive information off site. While cloud is a "thousand times more secure than it used to be," according to Markowitz, some companies are still reluctant to put information into somebody else's hands.

"Banks or financial institutions may have cloud-based applications, but in terms of the repositories they have of sensitive information, they tend to keep it in house, premises-based. If you want to make sure that fewer eyes, hands, and systems access your proprietary information, that's an argument for staying on-premises or deploying a hybrid solution."

However, as cloud-based solutions have matured, so has their security, Fluss points out.

"The last time we did the numbers, three years ago, [more than] fifty-five percent of IVRs were already in the cloud, which included a lot of large financial institutions. It's not typically an issue, because the customer's information in the IVR is commonly not on the cloud, but at the end user's site," Fluss says.

The perception that the cloud is not as secure as an on-premises solution is something that Angel/Genesys is working hard to change.

"We're very cognizant of security, and we put a lot of time and energy into ensuring that we are as secure, if not more secure, than on-premises solutions," Weinhold says. "We're PCI [Payment Card Industry] compliant [and] HIPAA complaint, and we're audited every year on both of those, so we're aware of how big an issue security is. We feel that the cloud is a secure solution if companies put the proper resources toward it. As that becomes more well known, it will become more expected that the cloud is a secure environment."

Making the Case for Mobile

As new channels come into play, more voice and speech technologies are being integrated into the mobile Web. Some channels have very high traffic patterns and volume, which is where cloud scalability comes into play.

"Consumers today have heightened expectations for mobile access to customer service," says Kim Martin, director of marketing at Voxeo. "Mobile Web portals, text applications, and social networking apps can result in high volumes that are more reliably supported in the cloud."

When speech recognition occurs locally on a mobile device, the speech utterance is recorded, translated, and acted upon using only the processing power that exists on a mobile device, explains Michael Morgan, senior analyst of mobile devices at ABI Research. The general benefit, he says, is that the device does not have to send out a network request and wait for a response before it can move to display results.

"When speech recognition occurs in the cloud, the recorded speech utterance is sent over the network to a data center that will then process and decode the utterance using [its] massive power," he says. "Because the cloud approach uses all of this power, it can process large utterances in a matter of seconds. After the cloud deciphers the utterance, it will send the decoded statement back to the device for use."

In 2012, Angel launched Lexee, a self-service solution that voice-activates mobile applications.

"We're providing a solution that helps companies connect with their customers in a new way by [enabling them] to have conversations with their mobile apps," says Angel/Genesys' Weinhold. "Lexee allows companies to access ASR and TTS services from a mobile app. If we attempted to insert these services inside the app, the app would be too large to use. Instead, we use the cloud to host these services, and the app contacts the services to access the technology."

There is also the widespread Siri effect, which essentially puts customers right where the technology is. Weinhold says that the company is generating a great deal of interest across a wide spectrum of industries.

"Because of Siri, there are now businesses looking at how they might incorporate mobile voice," Weinhold says. "They're interested in being a first mover in developing things like virtual assistants for their company, taking a Siri model and porting it to a customized version for their own company."

More end users are also interacting with voice in natural language on a greater number of devices; a single user might have a smartphone and tablet, or use voice in his car or in the living room on a TV. Consumers' preferences for self-service and their use of speech with devices have advanced. What was once a leading service for speech-based IVRs has evolved to a multichannel natural language platform focused on helping enterprises deliver a more natural, virtual assistant–style customer service for IVR, as well as for mobile and the Web, all using the power of a single shared platform, Skrbina says.

"We build solutions for all these devices," adds Vlad Sejnoha, chief technology officer at Nuance. "We build different portions of our technology stack that reside on a device, but they all communicate with our cloud-based technology to offload computational intensive tasks. More importantly, we do this so that we can provide continuity of experience across those devices so that any learning or adaptations of the users happens on one device, so it's there when the user connects from another device."

With certainty, he says, "The cloud is here to stay."


Staff Writer Michele Masterson can be reached at mmasterson@infotoday.com.


Page1 of 1
Learn more about the companies mentioned in this article in the Speech Technology Buyer's Guide:
Learn more about the companies mentioned in this article in the Vertical Markets Guide: