Self-Service Analytics Stalled Once. Will AI Stall It Again?

Executive summary Self-service BI stalled at about 25% of employees for more than a decade because the numbers didn’t agree, not because the tools were hard to use. AI chat makes asking easy again, but if each chat works out its own definition of revenue, every answer becomes a new “spreadmart” (a term from the early 2000s BI era for a spreadsheet or desktop database a team built outside the official BI system, with its own definition). Salesforce found that 89% of data and analytics leaders with AI in production have already seen inaccurate or misleading outputs. Self-service analytics with AI works when every tool answers from one set of shared business definitions, so two people asking the same question get the same number. Self-service BI stalled because teams couldn’t agree on definitions, so fix definition mismatches before you add tools. AI chat brings the same risk back at a faster pace. Workers at more than 90% of companies use personal AI tools (MIT Media Labs NANDA, 2025), and each chat can produce a private number nobody else can check.…