Author: admin

Adopting AI is wrapped up in policy red tape for some brands, has become a test-and-learn exercise for a few, is a rudimentary or revolutionary personal ChatGPT experience for a handful, and a pipedream for many others. Whatever the actual take-up, AI’s lure was firmly in evidence at yesterday’s Google Marketing Live event as 700 marketers and agency execs turned up to hear Google’s outline how AI will transform search advertising, creative development and measurement. Brands such as Les Mills, Officeworks, The Iconic, Canva and Virgin Australia were showcased for improving media ROI via Google’s latest AI-powered products and services. Heavy with hard commercial numbers, Google’s approach was a definite nod to the crunched economic climate and pressure all marketers are under to better prove out value and do more with less. The overriding message to advertisers is to stop waiting for change and start experimenting to make it happen.

Marketing teams around the world are both leading AI adoption in their companies but are equally restraining its deployment over increasing concerns about the security and scraping of their prompts and data to keep the large language models (LLMs) of AI developers learning and getting smarter. Salesforce Marketing Cloud CMO Bobby Jania says the Scarlett Johansson furore with Open AI two weeks ago over the likeness of her voice in the unveiling of “Skye”, typifies every conversation he has with a marketing team: it starts with concerns and questions of “where their data goes, who is going to have access to it, who learns from it, who trained off it”. The opaqueness on what data LLMs are ingesting is proving a boon for Salesforce and its position on retaining no customer prompts or data on any LLM’s plugged into its various cloud products. “The reality is right now it’s a differentiator for us because we’re able to talk about the fact that our business is not our customer’s data at all,” says Jania. “It should be table stakes for the industry. For a lot of solutions out there, their preference would be to continue to train an LLM using the data that goes in, which could be that customer data. And from who I talk to, every customer is very concerned about where their data goes”.