Zenith claims it has taken a “significant step” in the journey towards automated media planning with the launch of a new, artificial intelligence-powered service.

The Publicis Media agency has revealed details of an “industry first” new approach, developed over the past six months and trialled with UK-based insurance brand Aviva.

It feeds data from across digital channels into a proprietary machine-learning algorithm, attributing the credit for conversions throughout the user journey, and then uses its algorithm to optimise media plans in real-time.

By pushing conversion scores into a demand-side platform (DSP), it claims there is no need for human intervention, avoiding the need for agency teams to handle “complex and time-consuming” aspects of digital planning.

According to Zenith, its test run with Aviva saw the brand benefit from a 6% cost-per-quote (CPQ) improvement on car searches, and a 10% improvement in CPQ for display advertising.

Aviva’s head of marketing (trading) James Turner said he was “delighted” with the results, and said such attribution modelling will help the advertiser to improve ROI and create “operational efficiency”.

Ian Liddicoat, global head of data and technology at Zenith, told M&M Global that the technology will be “incubated” within Zenith on a client-by-client basis, before being extended across all Publicis Media brands.

“This is ground-breaking stuff from a technology perspective, and it is the beginning of a fully automated digital planning capability,” said Liddicoat.

“With programmatic, addressable media will ultimately include TV, and the logical extension of that is there will be too many permutations for the human brain to cope with. It is essential there is some kind of machine learning, and is a significant step on that journey.”