Artificial-intelligence startup NumberEight has raised a funding round after completing three trials of its technology with music-streaming companies.
The London-based company raised the undisclosed amount from investors including Beacon Capital, AI Seed, Ascension Ventures, 7% Ventures and angel investors including E100 angels from the London Business School
NumberEight’s software is designed to sit within iOS and Android apps, using data from the various sensors on the smartphone – from accelerometers and light sensors to barometers to understand the context of its owner.
For music-streaming services, this technology could be used, for example, to understand whether someone is at the gym or travelling on a bus, and adapt accordingly – from music recommendations to targeted advertising.
“We’ve completed three trials to date, with another one coming up,” NumberEight CEO Abhishek Sen (pictured above with co-founder and CTO Chris Watts) told Music Ally, adding that the three included a streaming startup as well as “some of the largest music-streaming companies”.
“In all of them, we delivered software or application demos where we showcased the benefits of how context can be integrated into the music recommendation and/or personalisation portion of the respective music startups,” he added.
“They lasted between 3-4 weeks and the feedback gave us the confidence that what we are building is indeed the right product and that there is a definite need in the market.”
The funding round will be used to hire more staff and continue developing NumberEight’s technology, with the company planning to release its first beta at the end of November.
The tech could be applied to a range of industries, but for now NumberEight is focusing on music, although Sen admitted that some potential investors had questioned this strategy when alternative markets (“like financial fraud detection”) might be more lucrative.
“Music may be a smaller industry when you compare it to financial fraud, transportation or fitness, but our take on it is that music is really interesting: it happens through activities and users get value throughout the day,” said Sen.
“You’re listening to music in your home when you wake up, when you get on the tube, when you go to work or the gym. That gives us a lot of insights into how we can deliver value, which can then translate into a lot of different things in other sectors.”
Sen sees NumberEight’s work as part of a wider trend around contextual awareness and ‘pervasive’ computing, with the data-crunching increasingly happening on devices rather than in the cloud.
He cited the concepts of ‘Federated Learning’ and ‘Edge Computing’ as well as Google’s work on image recognition as notable in this regard. NumberEight sees itself as part of a community, and hopes to publish an open source version of its software to get feedback from other developers.
Music-streaming services are the main customers NumberEight is targeting. “There are a lot of major players: Spotify, Pandora, Deezer, Apple Music, and Amazon is interesting because they also have video offerings and Prime,” said Sen, adding that streaming services in markets like India and China also offer opportunities.
While the obvious use for NumberEight’s technology might be even-more personalised music recommendations, he sees advertising as another important area.
“Advertising is where I feel money is left on the table,” said Sen, suggesting that currently, ad-targeting on streaming services tends to be very broad-brush, and based on signals from what someone is listening to, which may not always be an accurate reflection of their context. “I can listen to ‘party music’ while I’m at work,” as he put it.
This also applies to music recommendations, which Sen suggested have traditionally been based on people’s past listening and how they compare to other listeners.
“It’s been content-based recommendations or collaborative filtering, and context is always an afterthought. But what we want to say is that context is number one! And then you can bring the others in,” he said.