SoapBox Labs CEO Patricia Scanlon discusses the company’s new partnership in the US and its plans for the future.
Throughout 2019, Dublin-based speech recognition technology company SoapBox Labs has made some major announcements.
Just this week, the start-up’s founder and CEO, Dr Patricia Scanlon, was named Digital Hero 2019 at the Spider Awards. In March, the company announced a major deal with Microsoft, shortly before adding former TechIreland CEO Niamh Bushnell to the team as chief communications officer, and announcing that former Shazam CFO Colm O’Carroll would be leaving his role at Apple to join the start-up.
Keeping the momentum steady in the final quarter of the year, SoapBox Labs announced another major partnership a few weeks ago, this time with a US research centre to develop and improve children’s literacy assessments.
This multi-year partnership is with Florida State University’s (FSU) Florida Center for Reading Research (FCRR). We spoke to the Scanlon to learn a bit more about the news and the plans for the future.
A whole new challenge
“I founded this business back in 2013. I’ve been working in the speech recognition space for adults for 20 years now,” Scanlon said.
“Back in 2013, I saw the need for speech recognition technology for children while watching my own daughter interacting with technology. It was an app teaching her to read at that emerging preliterate stage of reading. Phonics, blending and decoding. I noticed they had no way of assessing whether she had actually learned anything it was supposed to be teaching her.”
With a background in speech recognition, it’s unsurprising that Scanlon realised her area of expertise could be key to solving this issue.
“Children’s behaviours are very different and any speech technology that was out there was doing a very poor job – particularly when you’re talking about smartphones and tablets and real-world noise environments,” she said.
Much of the speech recognition technology aimed at young children requires them to wear headsets and sit in a quiet room. But anybody who has ever met a child under the age of eight knows how unrealistic of an expectation this is, she added.
Building a presence in the US
The first hurdle for SoapBox Labs was to access large volumes of real-world, representative, diverse data. Since 2013, the company has been collecting that data to fix, create, build and innovate around children’s speech, designing algorithms and using deep learning to develop an accurate solution for kids as young as three.
The company has built a presence in the US already, partnering with the MIT Media Lab and its Personal Robotics Group. SoapBox Labs is also involved with the Chan Zuckerberg Initiative, which is how it closed the deal with the FCRR earlier this month.
‘Edtech in the US is, without a shadow of doubt, one of the biggest markets for us’
– DR PATRICIA SCANLON
“We were visiting the Chan Zuckerberg Initiative, who do a lot of stuff with education,” Scanlon said.
“They were funding this project, which isn’t just run in Florida. It’s part of a bigger Reach Every Reader project run by Harvard, MIT, FSU and FCRR. They are the leaders in the US on literacy learning for children. They’ve got a hugely reputable research institute on this particular problem – young kids learning to read.
“What they identified was that in order to do their research they needed speech recognition for young kids learning to read. Diverse accents, very young children with unique voices. There was nothing out there that would work accurately. It was the Chan Zuckerberg Initiative that suggested us to them.”
Working with the FCRR
After a visit to Ireland to meet SoapBox Labs, the leads on the FCRR team decided that the Irish start-up was a perfect fit for the programme.
“It’s a five-year project overall, with the initial stages being spent on validating the accuracy of the speech recognition and the approach of using speech recognition in order to identify children who are at risk of having reading difficulties or dyslexia,” Scanlon said.
This allows for early intervention, she added, noting that it’s possible to identify a child with reading difficulties before they can even read by using this technology.
It also takes pressure off schools when it comes to diagnosing conditions related to reading, as these types of assessments have traditionally been completed using pen and paper tests, which take time, resources and qualified humans to correct.
“It’s a cost-effective and scalable way of doing these assessments for children, globally,” Scanlon added.
The size of the US market
The edtech market in the US is the perfect place for SoapBox Labs to scale and move into other areas, such as designing toys, the founder added.
“Edtech in the US is, without a shadow of doubt, one of the biggest markets for us. We set the bar really high for ourselves, we went after young kids first, and the young kids are much harder than the 10-year-old with clearer speech. Education is a really high bar. Getting it wrong is quite detrimental for a child, compared to getting it wrong with a toy.”
The US represents a valuable market for SoapBox Labs, with adoption of edtech and smart toys much higher than in Ireland. Why is this the case?
“From what I’ve seen in the US, decisions are made on a district or state level,” Scanlon said. “In Ireland, decisions are made on a school level, so there’s very little coordination. No coordinated thinking, no coordinated buying. Levels of access vary greatly too.
“Some schools have complete internet access, others only have access in a handful of classrooms. Some schools have iPads, others have fifteen iPads that are shared between the entire school. It’s very different to some districts in the US where every student has an iPad.”
While Scanlon said that there are positives and negatives to both situations, especially when we consider how important the thoughtful introduction of technology is, rather than the indiscriminate application of tech in the classroom.
With that said, she added: “I think on this side of the ocean we have a lot to catch up on! It’s not just Ireland, but the UK as well. They’re definitely further ahead of us, but not by much.”
As well as having opportunities to scale in the US market, SoapBox Labs is also looking at English language learning in Asia, with a product already on the market with a partner company in Asia.
The start-up also has some other languages in the works, including Mandarin, Spanish, French, German and Portuguese, which aren’t ready for the market just yet, but are still in the pipeline.