CDP Institute

It’s one year since Google started showing ads in its now-familiar AI Overviews. They’ve now added them to AI Mode, a newer option that handles more complex queries. AI Mode will also offer a bespoke shopping experience, with agentic AI to track price and purchase when it fits your budget.

CommerceIQ has unleashed the “industry’s first suite of agentic AI teammates to drive e-commerce and profit impact.” It’s not clear exactly what’s “first,” but your new friends include Ally Sales Teammate, which identifies reasons for underperforming sales and recommends actions to improve results; Ally Category Teammate, which tracks content, availability, and pricing across SKUs; and Ally Media Teammate, which identifies underperforming campaigns. The agents are trained on data from 1,400+ retailers.

Jamie from our Department of the Obvious was thrilled to hear that in-store kiosks for food orders are perceived as less “friendly” than human shop assistants, according to an Intouch Insight study. Jamie nevertheless offered some anecdotal evidence that he has met some unfriendly human staff in his time. He also noted the less obvious finding of higher levels of consumer satisfaction with non-human AI-powered drive-thrus and advance mobile ordering.

Publicis Groupe has acquired Capitiv8, the influencer platform that boasts 15 million creators, including 95% of all influencers with over 5,000 followers. This follows last year’s acquisition of Influential, itself a huge influencer network which works alongside Publicis-owned connected identity platform Epsilon. Adding Captiv8 gives Publicis the biggest influencer network in the world.

Cloud data management platform Informatica has unveiled a broad strategy for agentic AI. AI Agent Engineering is a new offering allowing users to build and deploy multi-agent AI systems with no-code orchestration across a range of eco-systems, from AWS and Microsoft to Databricks and Snowflake. Also launched: a suite of CLAIRE agents to manage complex data ops including ingestion and quality control.

We’ll believe it when we see it, but a new report from LUMA suggests that creators might evolve from influencers to paid media — and, surprise, it’s AI doing the nudging. AI offers improved creative review, probabilistic comparisons of outcome against traditional ad channels — and creators themselves are dipping into predictive outcome modeling. Creators are also now tending to span across a number of platforms rather than sticking with their niche.

Microsoft is offering a low-code tool that lets non-technical users train AI models on their company’s own data. Microsoft 365 Copilot Tuning will be in limited release this June. Other announcements at the company’s Build 2025 event include multi-agent orchestration and new agent-building tools for developers.

ChatGPT announced a beta feature that lets companies include their own data in its Deep Research results. Users can point to files stored in GitHub, Microsoft OneDrive and Sharepoint, and Dropbox. Note that including data in query results isn’t the same as using that data to train a model.

Since we’re all about data access today, let’s also report that analytics platform Sigma is adding a File Column Type that can store unstructured files such as PDFs and images in Sigma tables, bringing structured and unstructured data together. I knew you’d be excited.

A new product from Salesforce, Agentforce in Slack, will bring AI agents into Slack conversations as purpose-built “teammates” rather than general purpose assistants. This is a step in the direction of replacing “do-everything” AI with more specialized models. Salesforce is also offering new, flexible pricing models for Agentforce in addition to the price-per-conversation model.

Having been fined almost $600 million by the EU for preventing developers from promoting alternative payment systems within their apps, Apple is digging its heels in. Why not just warn users not to purchase using those other channels? Big exclamation marks have been seen alongside listings of apps which don’t use Apple’s “private and secure payment system.”

As many as two-thirds of audience-serving companies are failing to use their audience data for personalization, squandering opportunities to activate and engage, according to a survey of 119 audience professionals by Omeda. Progress is slowed by tech limitations and data fragmentation, even though 85% see audience data as a competitive advantage. While search traffic is slipping, the respondents look to events, newsletters and direct traffic to build audiences.