Your business wants productivity. Legacy vendors want control. (www.businessinsider.com)

Reltio (via Insider Studios) argues that the real tectonic shift in the future of work isn’t new dashboards but who controls the data behind them. As conversational LLMs, agents, and autonomous workflows can bypass traditional GUIs, legacy vendors are responding by tightening rules on data export, sharing, and use for model training. That’s a deliberate move to preserve vendor lock‑in: with data siloed inside CRM, marketing, support and financial systems, companies remain tethered to point‑and‑click interfaces even as front‑end LLMs could eliminate the need for them. The post cites productivity friction (an average 1,200 toggles/day causing ~4 hours/week lost) as the “toggle tax” enterprises pay to access distributed data. For the AI/ML community this matters because models and agents need unified, real‑time access to enterprise data to deliver accurate, contextual automation. Policy restrictions on bulk access and training data will limit the capabilities of agents, complicate fine‑tuning, and increase the importance of robust data engineering: real‑time, trusted, governed data layers, fine‑grained access control, auditability, and interoperability. Reltio’s prescription is clear — the winners will be organizations that treat the data layer as the product: open, scalable, and governed so LLM front ends can safely and effectively drive productivity without recreating vendor lock‑in.