The Digital R&D Lab serves as an internal incubator for ideas and plays host to digital projects carried out in partnership with non-Radio-Canada groups.
By Tracey Lindeman for Nieman Lab
Up in the tower of Radio-Canada‘s Montreal headquarters, overlooking the Jacques-Cartier Bridge and the old Molson brewery, is the nervous system of the 80-year-old national public broadcaster’s digital shift.
This particular meeting is centered on chatbots — namely, the chatbots being built by a small team of engineers at Radio-Canada, the French-language arm of the Canadian Broadcasting Corporation. Digital research and development director Thomas Le Jouan’s Macbook display is up on the TV; in the left-hand column of his iMessage window are texts from chatbots at ABC, BBC, CNN and the Guardian.
It’s still early days for Radio-Canada’s chatbots; the project only began in earnest at the end of March. The team has just one full-time engineer (and two part-time engineers borrowed from other departments) working on it, looking like a startup within the much larger Radio-Canada — an organization that counts about 3,000 employees, most of whom are in Montreal. (CBC/Radio-Canada as a whole employs just over 7,200 people, according to last year’s annual report.)
They’re able to borrow support staff from other parts of Radio-Canada, and occasionally hire freelancers to help with things — like VR video stitching, for example — that fall outside of staff’s expertise. The two of them are working on multiple projects at any given moment, from chatbots to developing a Radio-Canada Google Home app to a machine-learning partnership with Montreal-based deep-learning specialist Element AI.
“We know we can use technology to enhance what we’re currently doing as a public broadcaster,” said Maxime St-Pierre, the head of digital at Radio-Canada, and Le Jouan’s and Kronström Richard’s boss. He’s been at Radio-Canada for a year, after many years in the private technology sector. Inspired by virtual reality efforts from The New York Times and innovation projects at the Washington Post, his mission has been to make an octogenarian public broadcaster more digitally agile and creative.
“It’s not digital first,” he said. “It’s digital always.”
Teaching an old broadcaster new tricks
Kronström Richard founded what today is RC Lab five years ago, as a way to establish a community within Radio-Canada that was interested in news industry innovation.
At the time, he worked under Le Jouan as a community manager in the TV news department. He came up with the idea to merge his work and his personal interest in technology while after attending SXSW. (“Xavier came back from SXSW and came into my office and said, ‘We have to do an RC Lab!’” Le Jouan recalled.)
Kronström Richard set up social channels for the Lab and began posting. RC Lab remained a social media–driven community for its first two years, with the occasional event or workshop for people to attend in person. But as more global media outlets began experimenting with digital storytelling and operational tools, Kronström Richard’s vision for the lab also evolved.
“I wanted to become a real lab, where we could experiment ourselves and not just talk about other media outlets’ great ideas and initiatives,” he said.
Expanding RC Lab from an employee’s pet project to an official, integral part of the Radio-Canada digital reinvention took some effort.
Kronström Richard and Le Jouan had an idea for an internal accelerator that would enable Radio-Canada employees to launch digitally minded projects and products — but they knew they couldn’t do it without their larger organization’s financial and organizational support. They spent a year and a half putting together a proposal and budget request; at one point, their pitch deck was more than 60 slides. The Ideas Accelerator was born in February 2015.
Radio-Canada employees submit ideas to the accelerator program that need to be digital, prototype-able, and more broadly useful to Radio-Canada. After a selection process by jury and then a public vote, the finalists are paired up with developers (which the R&D Lab pays for) and given three months to make their idea a reality. The Lab currently has enough funding to finance four prototypes a year (each of which is allocated $20,000 CAD). Some of the funded (and more widely propagated) projects have included an anonymous information-drop site, a podcast series called “Journal intime” and radio-frequency identification technology to track lost or stolen TV equipment. Each prototype is evaluated on its own set of metrics, according to Kronström Richard.
The Accelerator is now entering its fifth season at Radio-Canada — and its second season at the CBC, after the English mothership adopted the idea.
“The Accelerator came from the ground up,” Le Jouan said. “We would never been able to do the accelerator without management on board. It would have stayed a nice little initiative.” The work done through the Ideas Accelerator secured recurring funding for Le Jouan and Kronström Richard, and eventually led to the formal establishment of the Digital R&D Lab.
Innovating inside any large and established (and bureaucratic) company can be difficult, and a crown (state-owned) public broadcasting corporation — one that saw many significant budget cuts in recent years — is certainly no exception. It’s only been a year since Kronström Richard and Le Jouan’s endeavors officially became the Digital R&D Lab. Says Kronström Richard, “Now we can finally dig in.”
The funding allows RC Lab to run several internal and external projects and partnerships. The Ideas Accelerator and chatbots are two of about nine ongoing projects. Another is a 360-degree VR app developed in partnership with a group of Concordia University software engineering students.
On a summer evening at Chez Roger, an airy bistro pub in the Montreal neighborhood of Rosemont, the din of clinking glasses and jovial laughter served as the perfect backdrop for the experiment to come. People had congregated to witness a special live taping of La soirée est encore jeune, a popular radio program at Radio-Canada.
The Concordia software engineering students have spent the past 10 months working on their capstone graduation project: An interactive 360 virtual-reality app to bring the popular show into people’s homes. One of the students, Olivier Brochu Dufour, had come up with the idea after failing to secure tickets to previous La soirée live broadcasts.
“We found a solution, which was livestreaming the show in VR,” he said. “It was as simple as sending an email — ‘Hey, do you guys want to do your show in VR?’”
Brochu Dufour and his classmates — Rahul Malik, Joseph Atallah, Roberto Ruffolo-Benavides, Wing Long Chung and team leader Ihcène Cheriet — made the app in partnership with the Digital R&D Lab.
“Whenever an obstacle came they were there to help us,” says Atallah of Kronström Richard and Le Jouan. Adds Ruffolo-Benevides, “They pushed for the project to succeed.”
But there’s a problem. When the broadcast begins, their app doesn’t work. Kronström Richard’s brow furrows as he pecks at his iPhone, to no avail. Over Slack, the team discusses that the app has blown through its YouTube request cap. In the end, only a small handful of people got to experience it. The students were deeply disappointed, Kronström Richard told me a few days after the failed attempt, “but failure is a part of our work.”
The students have a shot at redeeming themselves this month at another taping of La soirée est encore jeune. The experience of collaborating with students was still largely positive, and Digital R&D Lab intends to continue bringing on innovative student projects as well.
Ambitious projects like these help Kronström Richard and Le Jouan secure more money and resources needed as they work toward not only helping modernize Radio-Canada, but putting it on the cutting edge of news and storytelling innovation.
“The ideal for me is to know that if I and Xavier left tomorrow, that the program would still exist,” Le Jouan said.
This story was originally posted on the Nieman Lab website, and is posted here with the editor’s permission.