Okta’s COO says companies are in denial about the hardest part of the AI revolution: redesigning work itself

Okta President and COO Eric Kelleher said the company is pushing managers to budget for both human labor and AI agents, shifting from workforce planning to "work planning." The article centers on a broader enterprise AI adoption gap: Cognizant said 93% of jobs are already disrupted by AI six years ahead of its prior 2032 estimate, yet humans still need to remain involved in 90% of analyzed tasks. The piece is largely qualitative and unlikely to have an immediate stock impact, though it underscores changing operating models for software and services firms.

Analysis

The market is still pricing AI as a software feature upgrade, but the article points to a deeper enterprise re-architecture cycle: if managers start budgeting for digital labor, the spend shifts from discretionary copilots to durable operating expense. That is constructive for platform vendors that can sit in the workflow and identity/control layer, because “agent sprawl” creates a new need for governance, permissions, auditability, and policy enforcement. In that framing, OKTA is better positioned than pure model vendors to capture the second-order monetization of AI adoption.

The more important signal is that productivity gains are being blocked less by model capability than by organizational design. That creates a lag between AI enthusiasm and margin realization, which is why near-term earnings upside across the software complex may disappoint even if usage metrics look strong. Over the next 2-6 quarters, the winners are likely to be companies selling the plumbing for agent management; the losers are point solutions whose value proposition is easily abstracted into a broader workflow layer.

W retains relevance only if AI meaningfully improves merchant and customer-service productivity faster than peers, but the article suggests that enterprise value capture is still bottlenecked by human process change, not just tooling. That means benefits to retailers may arrive unevenly and with a longer lag, while labor-intensive firms with weak change-management culture risk paying more for AI pilots without offsetting headcount flexibility. The contrarian read is that the current consensus underestimates how much of the AI spend will migrate into governance, security, and orchestration budgets rather than pure compute or model exposure.

Key Decisions for Investors

• Long OKTA on a 3-6 month horizon: use any post-news consolidation to add, targeting a rerating as agent governance becomes a budget line item; risk/reward favors a move toward prior software multiple bands if AI agent adoption broadens.
• Pair trade: long OKTA / short W over the next 1-2 quarters; the thesis is that identity and control monetize AI workflow redesign more directly than retail labor savings, while W faces slower/less visible margin conversion.
• Buy medium-dated OKTA call spreads 4-6 months out to express upside from AI governance demand with defined premium risk; ideal if management commentary shifts from experimentation to budgeted deployment.
• Avoid chasing broad AI application names that rely on rapid productivity realization; use them as sources of short alpha if near-term margins fail to inflect despite rising AI spend.
• Set a watchlist on enterprise software vendors with workflow, IAM, and security exposure; if agent budgeting shows up in corporate capex/opex plans over the next earnings season, rotate toward that basket before the market fully prices the theme.