NEA’s Tiffany Luck says enterprises are still figuring out their AI ROI
TL;DR
TechCrunch’s Equity podcast centers on NEA partner Tiffany Luck: after the tokenmaxxing push, enterprises are asking harder questions about whether AI usage produces measurable value. The mood shift has real examples: Uber reportedly burned through its annual AI coding budget by April, Microsoft revoked some Claude Code licenses, and Meta shut down an internal usage leaderboard.
Nauti's Take
The next enterprise wedge is not another chatbot, it is the unglamorous accounting layer underneath. Whoever ties token spend, usage, and real workflow gains into one hard dashboard becomes the CFO translator for AI.
Without that layer, adoption is just expensive enthusiasm.
Briefingshow
AI usage is moving from culture signal to finance problem. Teams can no longer treat license counts or token burn as proof of progress; they need metrics such as cost per completed task, defect rate, rework, and time to customer value. That creates room for new vendors, provided they deliver more than polished dashboards.
Sources