Most of your AI agents don’t deserve to exist

Enterprise AI can generate meaningful returns, but panelists at the 2026 WSJ Tech Council Summit argued that value is concentrated in a small minority of deployments, while many AI agents and token expenditures produce little or negative financial impact. Retool CEO David Hsu estimated that roughly 90% of token spending may be unprofitable, with the remaining 10% driving average returns of three to five times spend, while McKinsey research found that personal productivity gains often fail to translate into measurable EBIT or investor-relevant value. Examples showed that sophisticated, costly systems do not necessarily outperform simpler automations, making ROI difficult to predict before deployment and requiring organizations to test, measure, and discontinue ineffective tools. Limited visibility compounds the problem, as only 5% of surveyed technology leaders reported full awareness of production AI tools, allowing poorly designed applications to generate substantial unnoticed costs. Speakers emphasized governance, ongoing monitoring, workforce retraining, and efforts to reduce unit costs rather than broadly restricting AI use as central to realizing sustainable returns.

Use this post, company, and trend context to find content marketing opportunities, perform competitive analysis, or address product feature gaps via the Plushcap MCP server or the Plushcap API.