NEA’s Tiffany Luck says enterprises are still figuring out their AI ROI | Equity Podcast

The 'mental load' opportunity

Next-generation personal agents from companies like Town and Ali aim to manage the logistics and domino effects of tasks across personal and professional calendars, not just complete single actions.

Current usability gaps

Existing agents still generate friction by requiring extensive teaching and verification, functioning like 'clueless interns' that fail to reliably reduce user workload.

Privacy and security prerequisites

Mainstream adoption requires robust security standards, potentially through certification bodies like AIU (AI Underwriting) that establish AI-specific safety benchmarks analogous to SOC 2.

Token spend accountability

Enterprises are currently 'token maxing' and exceeding AI budgets, creating a market opportunity for startups that help companies track and measure actual return on AI investments.

Unit economics under scrutiny

Sustainable AI business models require addressing compute cost structures and infrastructure dependencies before companies can achieve the free cash flow metrics public markets demand.

Early deployment phases

Despite massive private market valuations, most enterprises remain in the 'first few innings' of determining how to generate measurable business value from AI implementations.

Liquidity events approaching

OpenAI's confidential IPO filing and Anthropic's expected public debut will test investor appetite for high-growth AI companies that have reached unprecedented scale and valuation levels.

Converging market segments

While OpenAI currently skews toward consumer applications and Anthropic targets enterprise, the distinction between consumer and business AI use cases is rapidly blurring.

Infrastructure dependencies

A significant component of AI company valuations stems from guaranteed access to compute resources, whether through ownership or strategic partnerships with providers like xAI.