Three tech companies made similar moves in the same week. Google, Cordial and Twilio each announced ways to give AI tools more access to customer, product and conversation data. The common message, reported by CX Today, is that AI is only as good as the information behind it, and businesses want to see how its decisions get made.
What changed
Businesses are asking vendors to connect trusted data to real-time decisions without losing oversight. An AI agent can only personalize responsibly if it understands what a customer signal means, sees the right context at the right time, and knows when to hand off to a human. That raises the stakes for data quality, shared definitions and rules about who can see what.
Google and NIQ
Google is working with NielsenIQ (NIQ) to make AI shopping recommendations more accurate and measurable. Google's Shopping Graph holds more than 60 billion product listings. NIQ adds data on more than 250 million products and tracks stores and sales across 90 countries.
The goal is to help AI judge whether a product is relevant, in stock nearby and actually bought. NIQ is also adding Google's open-source Meridian tool, which estimates how marketing spend affects sales, to its Cadence platform. The open question is how well brands can check the recommendations and sales credit that sit between discovery and purchase.
Cordial and Twilio
Cordial now lets enterprise customers pull live audience and customer data into Claude, ChatGPT, Microsoft Copilot and custom agents. Its customer data platform, which keeps customer information in one place, is meant to help marketers find higher-value audiences and get suggestions for campaigns, segments and offers. Better data means better suggestions.
Twilio is building what it calls a conversations layer, so voice and messaging agents keep context as customers move between service and sales. It also put an identity offering into private beta. The aim is to let businesses confirm who an AI agent represents. Twilio reported an eight-figure deal with a health-insurance provider, but did not break out revenue by product.
What to do
If your business uses AI for sales or service, start with your own records. Check these points:
- Customer data is clean, and teams define terms like "active customer" the same way.
- You know which outside AI tools can see which data.
- A person can step in, and there is a record of what the AI decided.
- You have a plan for fixing mistakes quickly when automation gets it wrong.
Tara DeZao of Pega told CX Today that a bad interaction should be corrected in real time. That is easier when the data and the controls are already in place.
The takeawayBefore connecting AI tools to customer data, clean the data and decide which tools can see what.