Interactive video advertising provider Innovid and Cisco are building on their previously announced partnership, announcing an initiative to have ads dynamically delivered to second screens using voice-based metadata from TV programs. Innovid's CTO and co-founder Tal Chalozin showed me a demo last week of Cisco's cloud-based technology analyzing words spoken in TV programming to generate keywords and context. The information was then passed to Innovid which delivered a relevant, interactive ad to the viewer's second screen within a few seconds.

Importantly, Tal noted that the idea here is not to create another second screen app, but rather to leverage apps from pay-TV operators, at least initially, which the user might have open while watching the linear TV program. In the event that the user doesn't have that particular app open, Innovid could send a push notification or a tweet, which might include a specific promotion (e.g. a limited time promotion, etc.) prompting user engagement.

For the pay-TV operator, the key benefits are to sell higher value ads in their inventory allotment (pay-TV operators typically get a few minutes of ad inventory each hour on cable TV networks, hence the local car dealer and pizza shop ads) and drive engagement with their apps. For Innovid, the initiative provides more opportunities to create and deliver personalized video ads. Cisco will offer this through its Videoscape Unity platform.

Cisco envisions this working on both its own Scientific-Atlanta set-top boxes, but other brands as well, along with connected TV devices. Tal said the plan is to pilot this with a cable operator in 2014.