Speech recognition for kids draws $6.3M

A specialized voice technology startup in Ireland just closed a €5.8 million ($6.3 million) round, and executives say they are ready to aggressively market their software.

SoapBox Labs Ltd. makes cloud-based automated speech recognition specifically for children three to 12 years old, with a privacy focus. SoapBox sells the application for education roles and for toys. An article in The Irish Times reports that the company has signed about 30 global licenses.

A year ago, the company partnered with Microsoft Corp. in a pilot project aimed at turning around growing childhood illiteracy. SoapBox software has been hosted on Microsoft’s Azure cloud to make the app more scalable and to reduce the cost of getting it to educators globally.

The artificial intelligence and machine learning algorithms underpinning the speech application can deal with significant behavioral, physical and even environmental differences between children and those who are older.

Environmental massaging is important because the settings that kids are in most of their waking hours are noisier and more chaotic than are adult surroundings. SoapBox apps also can deal with the unpredictable way children speak.

The company operates in a sensitive area when it comes to privacy. Executives maintain that their business complies with COPPA, the United States’ Children Online Privacy Protection Act, and the European Union’s General Data Protection Regulation, or GDPR.

Its software today is available as an online API or online as a software development kid. By the end of the year, according to SoapBox, it will be delivered by embedded solutions.

The investment had two leads: Elkstone Ventures, a unit of the Dublin-based financial-services firm Elkstone Capital, and Astia, a San Francisco-based venture capital firm looking to fuel women-organized business ideas. Ireland’s Disruptive Technologies Innovation Fund also participated.

Article Topics

AI  |  artificial intelligence  |  biometrics  |  children  |  funding  |  machine learning  |  speech recognition