What’s happening in the world of speech? Lots. I recently attended TMC’s Speech-World event in Dallas and learned a great deal about the future of speech technology. At some point over the horizon, I am convinced we will all have natural language-enabled assistants that will be able to both access vast amounts of information and answer questions about virtually any topic. As I mentioned at the show, the biggest threat to offshore outsourcing is speech. This won’t happen this year, but at the end of 2006 we will start to see more mainstream acceptance of speech in businesses. 2007 and 2008 will likely be the years during which speech technologies will come into their own.
Customer Interaction Solutions’ sister publication is Internet Telephony magazine. During the mid-1990s, I observed three growth cycles of VoIP. First, everyone had to have IP telephony. A year later, no one seemed to want it. I saw upturns followed by market drops in 1996-1997 as well as 1999-2003. Now, 2005 is the year of VoIP.
The point is, we should be used to technology slowly becoming ready for prime time and going through ups and downs until finally the arrival of an acceptance period followed by a budgeting period. My intuition tells me that 2007/8 will be the right time for this to really happen in large numbers in speech. Yes, there is market research out there that may be more optimistic than me, but I trust my intuition more than the research, as the fickle nature of purchasing decision makers is tough to take into account via typical research methodology.
In the meantime, while we wait for everyone to have speech in their enterprise, I am seeing applications that handle the low-hanging fruit. For example, Cisco announced at the show that it will have a speech recognition product named Galileo available soon. Cisco’s Vickie McGovern gave an incredible presentation on her vision for speech and how speech will redefine communications.
The core message of the keynote was that speech can reduce the complexity associated with business communications. She touched on a variety of key elements including virtualization; Rich Media (focusing on personalization preferences); natural language processing (for such applications as directory services, e-mail and voice mail access, password services, communications management, and calendar management); and modality, presence and context (a good example: someone calls me while I’m on another call, I respond via IM, and the response is read back to the caller utilizing speech technology).
Not to diminish any element of Ms. McGovern’s presentation — the whole keynote was compelling, and she had my attention throughout — but perhaps the highlight was the demonstration of Cisco’s upcoming Galileo product. This was the first such public demo of Galileo, components of which will begin shipping this fall. On the surface, the application reminded me of an application called Wildfire.
McGovern used a standard desktop phone to call into the system, and using voice commands, she navigated through her e-mail inbox, listening to messages from, among others, Cisco President and CEO John Chambers. The demo also showcased Ms. McGovern accessing her appointments from her Outlook Calendar.
I’m definitely looking forward to seeing more of this application in the near future.
McGovern closed by declaring a new world of communications as we enter a phase she called “communications experience convergence,” during which we will see an increase in business process efficiencies. “Speech is redefining communications,” she said, “affording anywhere anytime access, and streamlining communications. Those are the key elements that speech brings to the table.”
Aculab’s Take On Speech
Aculab’s Mike Matthews, head of product marketing, followed McGovern to the stage. Matthews played up to his company’s position as a provider of DSP resource boards and other tools for the development community. “We don’t do solutions,” he said, “but rather, facilitate them.”
Matthews started off with a thought-provoking assertion: “Speech will become redundant,” he declared.
His talk ranged across a wide variety of interesting topics, including the fact that often in technology, the most interesting and lasting technical developments are frequently unanticipated byproducts. He gave a good example of how Teflon was developed as a heat-resistant element for NASA’s space shuttle program, but the chemical coating has become best-known for its non-stick use on kitchen pans and utensils.
How many commissioned technologies actually become successful, and how many were unanticipated successes? Matthews prodded the audience, “What about speech?”
He gave a wonderful example of how technology that was used in a prison to enable inmates to have greater access to making calls to the outside world (based on a series of privileges, earned credits and good behavior, etc.). Some of the technologies needed to enable secure prison telecommunications, such as speaker verification (for authentication) and word spotting (for monitoring the content of the calls), have found their way into mainstream speech applications, proving Matthews’ assertion that “Specialist partners working together with speech technology elements can result in innovative business solutions.”
The presentation closed with his discussion of addressing the reality that experience drives different approaches to old problems. For example, first-generation products are tougher to develop than subsequent revisions. Second-generation development builds on past experience, leading to more successful deployment of products.
“We need to keep working to improve the perception of speech-enabled solutions,” said Matthews.
The Innovators At Spanlink
I had a chance to meet with Spanlink at the show. They are doing some amazing things with IP contact center technologies. I wrote about the company over a year ago and continue to be impressed with how they have adapted to the changing world of IP contact centers. Ten years ago, we wrote about Spanlink in this magazine, as it was one of the first companies to enable “call me” buttons on Web sites. At the time, this was novel and the company was way ahead of the industry.
Spanlink employees briefed me on the growth the company is experiencing, and we then came to discuss its biggest customer, a company called Grainger (www.grainger.com), which I can best describe as a Home Depot-style hardware store with a focus on industrial specialty items. The company has 450 branches, and when customers call, they like the ability to speak with someone with a similar accent who is from their local region. People call the store because they have questions about products and whether those products are in stock. The company sends out voluminous catalogs — the online catalog is more than 3,000 pages. I am sure there are many questions that potential customers have about the catalog, as the items are fairly technical. For example, a random look at page 3,481 of the online version shows a pneumatic receiver controller, whatever that is (please, no letters; I am sure I will never need to know what it is).
The company tried to consolidate and, at one point, a centralized call center was implemented to answer customer questions. Guess what? Customers didn’t like it…they felt they had lost the personal touch of a local store.
The company turned to Spanlink, which built two distributed contact centers and a system to transfer calls to the branch closest to the caller. The solution was set up so that if no one was available to answer a call, it was sent to the next closest store after 20 seconds. From customers’ standpoints, it was everything they wanted — someone from a similar area to speak with about their questions. By centralizing databases, anyone at a Grainger branch can tell which store has what inventory. This is a perfect win/win scenario.
One of the reasons I chose to write about Spanlink is that the Chairman and CEO of Grainger mentioned in a statement to financial analysts that due to the new Spanlink system, the number of calls answered within three rings went from 92 percent to 98 percent! The CEO went on to say the company expects incremental sales of one to two million dollars in the first year after the new system’s installation.
So often, we think about IP contact centers merely saving money. In many cases, they can save and make you money simultaneously.
From a business standpoint, Spanlink has made a brilliant move by basing its product on Cisco VoIP technology. Spanlink enhances Cisco’s product and makes it better. Cisco, in turn, has a vested interest in seeing Spanlink do well.
Back To Speech
Other big news at the show came in the form of Premiere Global Services, which announced two new speech-enabled services to its portfolio. I spoke with Betsy Rahm, director of marketing communications, and Dewey Anderson, product management director. These two people were the perfect sources of information, as I learned about the current and future shape of Premiere’s services.
If you haven’t watched closely, you may have missed the fact that this company was formerly Xpedite, a leader in the fax solutions market. Premiere is now a global powerhouse in communications, providing messaging in every format from voice to e-mail to fax to SMS. The company has 46,000 worldwide customers and has broadcasted 14 million messages in a single day. I am told it benefited greatly from the political campaign season last November.
The company has also specialized in niches in the market, such as dispatching, where it has a solution called DispatchManager that automates and increases the efficiency of truck roll-outs to customers. The company also has a product in the mortgage space named, logically enough, MortgageManager. A more generic service is Collections-Accelerator. Who couldn’t use a turbo-charging solution in accounts receivable?
One of the areas Premiere Global Services sees as important to its growth is voice, where the company is honing the power of speech to increase its value to customers. A new application announced at the show uses the power of speech to survey. As part of an outbound campaign, an individual can be called and asked if he or she wants to be part of a survey. If the person responds in the affirmative, he or she is launched into an application that conducts a survey, requesting answers of “yes,” “no” or multiple choice answers to a series of questions.
There is also a new collections application that allows companies to both ask for balances due and collect the money on the spot via credit card. This is a very powerful product as, in my experience, people who owe money are generally embarrassed to speak with a live agent. Having an automated call followed by automated collection is the best way to increase payments. The next step for this application is to take checks by phone — something I am told is in the works.
Anderson and I also talked about the future of such services, and we discussed that the most immediate product we may see is a shopping cart application that allows a customer to receive a reminder call about a potential purchase; for example, contact lenses that need ordering, or printer ink. A customer answers the phone, and if he or she decides the product is desired, the customer responds regarding the item and quantities. There is certainly a convenience factor here that is a win/win for consumers and for companies using these services.
Another future service we will see, perhaps as soon as August of this year, is authentication applications that allow the home of a student who fails to show up at school to receive a call about the absence. Authentication ensures that a parent, not the student, gets the message.
I asked for examples of more future applications as I was partly intrigued by the possibilities of these services and partly ecstatic that these applications didn’t exist when I was in high school. A query about what could be next led to a wonderful discussion about a new concept for such services. Imagine finding a product you like — perhaps at a friend’s house. Let’s say it’s a pair of shoes that retails for $200. You don’t want to pay more than $150. You enter — probably via a Web site — the product info and the maximum amount you will pay. You then set a reminder to follow you as you travel. When you are within X miles of a location that sells the product you want for the price you want, you will get a call to let you know the price and location.
One would imagine eBay would be a natural company to take advantage of such a service, or perhaps Google’s Froogle service. Of course, many things need to happen before such a service could work. Infrastructure would need to be in place, etc. I did think of another application in which this technology could work today. Assume Apple comes out with a new iPod, or there is any hot consumer electronics device on the market that is in short supply. Using the Apple example, I could go to the company’s Web site and sign up to receive a call when the item I want is within five miles of me. Theoretically, Apple knows which stores have the iPods. Perhaps over time, consumers would get used to using such services, which could then grow in functionality. Remember that GPS is built into many phones, so this application could be built now.
Some of the most interesting services, according to the company, are those that combine broadcast functionality with backend transactions. I really think the shopping cart idea has legs, and spam has drastically reduced the effectiveness of e-mail. The phone, coupled with a speech recognition/transaction processing backend, is an inexpensive and effective way to communicate with customers.
Speech is far from the mainstream, and one of the hurdles mentioned at the show is the fact that speech solutions are still expensive to deploy and maintain. Accuracy is there, but it costs money. As prices drop, we can expect to see more widespread adoption of the technology. Perhaps hosting is the most cost-effective way to deploy these technologies today. The combination of IP contact center and speech technologies together needs to be explored, as every contact center can save money and increase sales when these technologies are implemented correctly. When they are used synergistically, look out!
Thanks go to Greg Galitzine and his blog on TMCnet for assisting me with portions of this column. To read Greg’s blog, which is updated daily and generally covers every important development in VoIP, please visit: http://voip-blog.tmcnet.com/blog/greg-galitzine/. CIS