Safely manage your Zendesk from the AI assistant you already use, via the Deltastring MCP. Beacon configuration platform
← Back to news

Google buys Spirit data for AI training, say executives

Google has acquired Spirit data for AI training purposes, according to statements from company executives. The move signals Google's continued investment in securing proprietary datasets to enhance its generative AI capabilities, particularly as competition intensifies with competitors like OpenAI and Anthropic. Spirit's data—likely comprising customer interaction logs, support transcripts, or similar operational records—represents the type of real-world conversational material that trains more effective language models. For CX teams, this acquisition underscores a critical reality: the data flowing through your support platforms, ticketing systems, and customer interactions has become a strategic asset in the AI arms race. The question facing many organisations is whether their data governance policies adequately address the possibility that platform vendors or their partners may leverage customer data for third-party AI training, even with anonymisation claims.

The implications for CX professionals are substantial and multifaceted. Teams already deploying AI-native solutions like Salesforce Agentforce or competing platforms should scrutinise their data processing agreements and understand exactly what happens to their interaction data post-collection. Smaller vendors and mid-market CX platforms face pressure to either secure their own proprietary datasets or risk becoming less competitive as larger players like Google accumulate training material at scale. This creates a widening capability gap: organisations using platforms with restricted data policies may find their AI features lag behind those with access to larger, more diverse datasets. The broader concern is whether CX teams have sufficient visibility and control over data usage policies, or whether these decisions are being made at vendor level without meaningful input from the teams generating the data. As AI adoption in CX nears 70% yet only 2% of programs reach centre of excellence standard, the quality of underlying AI training data will increasingly determine which implementations succeed and which plateau.