TalkIQ Launches With $7 Million And Speech Recognition Software It Says Beats IBM Watson And Google
WiFi company Zenreach is used to winning at least 60% of the time when it goes head to head with competitors. But when the 150-person startup went back and studied the conversations its sales reps were having with prospects, CEO Jack Abraham was confronted with a challenger beating his team 3 times out of 4.
“We hadn’t even considered them a competitor,” says Abraham. “But when you can scan through all the instances a competitor is mentioned, it’s easy to scan through what you’re missing.”
The Zenreach sales team got that wakeup call courtesy of a new technology that launched on Tuesday called TalkIQ. Incubated out of Abraham’s venture fund Atomic, TalkIQ is coming out of the gate with $7 million in initial funding from name-brand investors and speech recognition software for sales, customer service and onboarding that the startup claims is 2x better than IBM Watson and 3x better than Google.
TalkIQ formed more than a year ago along the premise that if someone (or a machine) were to study every call made in a company, she (or it) could uncover patterns and insights to make the business run better—especially if you accept, as TalkIQ did, research that suggests that as much as 68% of customer interactions with company contact centers still happen by phone. So TalkIQ pulled together a team of voice recognition experts, natural language process experts and engineers to solve the challenge of transcribing voice calls accurately enough for them to be studied, then build the analytical tools to drive insights from the resulting text.
“Salesforce Einstein and the rest still can’t compare what they find to outcomes and all your other data,” says CEO Ben Coulter. Testing TalkIQ’s word error rate against IBM Watson and Google with a handful of companies, TalkIQ found it was twice as accurate as Watson and 3x as accurate as Google, Coulter claims.
One early advantage TalkIQ has had is that it’s been able to take advantage of Abraham and Atomic sister company Zenreach, which raised a $30 million funding round in July, to test its product more closely than it would be able to work with an outside company. One strategic question Zenreach faced was which social network to focus its attention on. Using TalkIQ to study its discussions on social with customers, Zenreach found that Facebook was mentioned 4x more than Instagram, and 6x as much as Twitter. “That put to bed the debate within the company,” says Abraham.
TalkIQ’s aspirations go far beyond sales reps, to customer service, employee onboarding and corporate strategy. The software could flag the most common complaints users might have with a company’s product or site; it could also extract best practices used by employees to share with new hires. And when companies decide which new features to prioritize based on customer demand, TalkIQ could help qualify the winners.
Investors Felicis Ventures, SV Angel, Polaris Partners and Cherubic Ventures led the initial funding round in TalkIQ. Coulter says that with their support, TalkIQ has no interest in a quick exit to become a feature inside Salesforce or another CRM player.
“The people who want to work on this look at voice recognition as their life’s work,” TalkIQ’s CEO says.
