You can tell Comcast what to do on its Xfinity TV voice remote

Customers are barking orders at their remotes. And Comcast is listening.

The nation's cable giant expects to process four billion TV-remote voice commands this year, taking consumers closer to a futuristic era of interactive, hands-free television.

With its early successes, Comcast now plans to broaden TV voice remote's functions to make the commands more responsive and human-like. The company also aims to spread the voice commands to more Xfinity Home features such as alarm systems, lighting, thermostats, and in-home cameras. Most of these voice-activated features are now available to some extent and Comcast says more capabilities are coming.

Comcast's voice development, with teams in Philadelphia, Washington, D.C. and Silicon Valley, lags tech giants Amazon and Google, which pioneered the latest advances. But the company says its voice offering is competitive with the industry and it believes the voice remote has made for a more satisfying experience for its TV customers. Many of them have complained for years about the complexity of the much-maligned push-button remote.

Message boards generally give the Xfinity voice remote good reviews, though some complain about problems that could require set-top box reboots.

The upshot is that voice is going to continue to grow. "It flattens the whole user experience, so they can just jump to whatever they want," Chris Satchell, the cable division's chief product officer, said of voice controls, which Comcast launched in mid-2015. "Voice has got to the point now so that you can make it the new interface."

The goal, Satchell said, is "how do you have a natural conversation with X1" – the cable company's set-top box – "and get what you want. We are not trying to be the best in e-commerce. We are not trying to be the best in general knowledge. We are not trying to be the best in jokes."

Comcast has kept its voice focus on TV and movies, music, sports, customer service, and smart home features. One concern is that TV remote commands directed at the television aren't then translated into commands for the smart home, or vice versa.

Demonstrating its adaptability at the Comcast Center, Satchell asked a TV voice remote for all the Dwayne "The Rock" Johnson action movies. Almost instantly, the movies appeared on the big flat-screen television behind him, with names and thumbnails. The Comcast voice remote for subscribers with X1 set-top boxes also can respond to spoken requests for Pandora searches for music and YouTube videos.

Tony G. Werner, Comcast's president of technology and product, said "today it starts and ends with the customer experience. Voice is so intuitive because everyone knows how to talk."

Voice also is booming as a business and a platform, popular with millennials and driven by the ubiquity of smartphone applications and Amazon's Alexa.

At the same time, recent advances with voice-recognition software and WiFi connectivity have enabled voice products to seamlessly connect to the internet. People verbalize desires and voice platforms respond, creating a new convenience.

Google, Apple, and Microsoft, in addition to Amazon, are recognized leaders. South Korea's Samsung also recently introduced its voice platform Bixby.

Satchell maintains that "when it comes to our domain, we are as good as any of them."

Adam Marchick, the co-founder and CEO of VoiceLabs, a software company, said that American consumers are expected to buy 25 million voice-controlled speakers, such as Amazon's Echo or Google Home this year.

"In five years, every U.S. home will have a voice-enabled device" — that's not a smartphone – "and the next question is what multi-billion-dollar industry will be disrupted," Marchick said.

Targets are likely to be smart lighting, grocery shopping, prepared-food delivery, home security, movies, and television, he said.

Marchick proposes a scenario: If Amazon makes it convenient to order a movie on Saturday night by just saying it out loud, will that make you more willing to rent – or buy – a movie from it or from a cable company with a push-button remote?

"If you buy an Amazon Echo, you will buy more on Amazon," Marchick said. "There is data on that. If you buy a Google Home, you will do more Google searches. I believe that if you use the Xfinity voice remote, you are more likely to be a loyal Xfinity customer."

Voice's resurgence seems counter-intuitive. The technology first boomed in the 1990s with voice prompters in customer call centers – not always a satisfying experience as the prompters many times routed callers in the wrong direction. Then nothing happened with voice, until Apple released Siri in 2011, and Amazon followed with Alexa in 2014, experts say.

Dina Abdelrazik, market research analyst at Parks Associates in Dallas, said that in recent years, "voice took the market by surprise. There are other manufacturers that are entering the space to offer voice remotes for a friction-less [TV] experience. But it takes a lot of sophistication and resources to build that capacity."

Voice remotes are differentials now for pay-TV companies but someday will be everyday offerings, Abdelrazik said. Satellite-TV operator Dish Network and Roku also offer TV voice remotes, according to their web sites, but not yet Verizon FiOS. "The end goal is to be everywhere, and you see this with [Amazon's] Alexa," Abdelrazik said.

Amazon is the market leader. But Apple's Siri may have the strongest hold on the U.S. consumer, because of the iPhone.

Parks Associates research shows that 64 percent of consumers using voice-enabled devices ask for information, 42 percent seek directions, and 31 percent request music.

Comcast has distributed voice-control TV remotes to about 12 million homes or slightly more than half of its residential Xfinity TV subscribers nationwide, with Comcast engineering teams ironing out the kinks. The Comcast teams are divided into four areas: voice recognition, artificial intelligence, meta-data, and hardware.

This is basically how it works: an Xfinity subscriber speaks into a remote and the snippet of audio is transported to Comcast though the set-top box. The audio is whisked into the Comcast cloud and translated into a text string representing the spoken words.

That text is analyzed by Comcast computers running artificial intelligence, which has been developed using a giant database of commands from millions of users. The Comcast X1 cloud – basically 50,000 separate computers – interprets what the subscriber wants to do and performs it, sending the directions back to the home set-top box and television. This all takes a second or less.

Many things can go wrong. The text string could be garbled with accents or mumbling by the subscriber. The words could be interpreted correctly into text string but contextually mismatched – someone may be looking for Steven Spielberg movies and not the HBO Spielberg documentary, or vice versa.

Jonathan Palmatier, vice president of product management in voice control in Philadelphia, and Grace Thornton, voice product manager in Washington, said the voice team anticipates some of these issues, using the company's database of language patterns.

Now located on the 35^th floor of the Comcast Center, the voice engineering teams are likely to relocate to Comcast's second high-rise in Center City when it opens in 2018.

But there are some things that Xfinity voice remote cannot do.

More than one Comcast subscriber has asked their Xfinity remote for a pizza – as in, I'd like a half-pepperoni, half-plain, say the engineers who monitor the deluge of TV voice streams.

No, you can't today order Domino's with your Xfinity remote. But stay tuned.