SAP and Enterprise Trends Podcasts from Jon Reed (@jonerp) of diginomica.com

Jon Reed of diginomica.com uses this feed to argue about enterprise software, not to admire it. Several recent recordings were made at industry events, and you can hear it. One is a 'halftime recap' from Oracle AI World in New York, recorded while the day is still under way. Another is an end-of-day debrief at a Sage analyst gathering in Atlanta that wraps up when the vacuum cleaner comes on. A third comes from the close of CRM Playaz IRL. Other episodes are panels or live streams. The guests give the talk its range. Rebecca Wettemann appears in two separate recordings from Oracle AI World: one about her CX research, published on diginomica, and one with her quick take on Oracle's AI direction. The research conversation keeps returning to data, and Walmart dropping OpenAI for commerce because its bots lacked accurate inventory information stands as the cautionary tale. Her stated next step is to study where AI is already working, because that, in her words, shows 'where the data bodies are buried.' The Oracle recap describes the company's pairing of infrastructure and Fusion apps as among the most coherent enterprise AI strategies, with agents that inherit existing security, approval rules and compliance. It also notes that market watchers are focused on data center build-outs, OpenAI and layoff news instead. Brian Sommer, introduced as a usual suspect, joins Reed at Sage Analyst Day 2026 to talk about ERP value, pricing pressure and why the trust built with finance leaders and accountants cannot simply be recreated with AI coding tools. Esteban Kolsky, who gave the closing keynote at CRM Playaz IRL, talks through AI literacy with Reed, and their exchange ends on a warning about becoming dependent on AI providers. That same event ran a satirical commercial for a law firm offering to sue companies over AI failures. SAP gets its own opinionated, unscripted review with returning guests Josh and Jeff. In it, one panelist argues that teams spending most of their time keeping the lights on should give something like a day a week to experimenting with tools such as BTP, and should accept that some attempts will fail. The live stream titled 'Enterprise month in review - agentic AI gut check time - with Andreas Welsch' compares quick-build AI to the spreadsheet era, which was cheap and fast but thin on audit trails and heavy on technical debt. It also asks who maintains agents once the people who built them leave. The jokes come easily, from a Princess Bride line in one cold open to a worry about getting arrested for staying past office hours at the end of another. None of this softens the core question of whether enterprise AI claims survive contact with governance, security, data quality and cost. The result is enterprise commentary without the vendor polish.