About this episode

The episode discusses the challenges and innovations in building agentic AI products at enterprise scale with insights from industry leaders.

If you're betting on agentic AI, hear directly from the builders navigating the challenges, innovating pricing models, and creating the coming reality where every employee manages thousands of agents. Topics Include: Four panelists represent the full AI stack: build, run, secure, monitor. Agentic AI moved faster than anyone predicted just 18 months ago. Writer AI's no-code agent builder missed both its target personas entirely. Non-technical users now just prompt agents instead of building workflows. Fireworks AI processes over ten trillion tokens daily across open models. DeepSeek's Christmas release tripled Fireworks' capacity needs almost overnight. Okta identified agent identity as a security problem from day one. 91% of organizations are already using AI agents in some capacity. Datadog evolved naturally from dashboards to autonomous investigative agents. Bits.ai agents now diagnose production incidents before engineers wake up. Trust requires explainability — black-box agents stall enterprise adoption cold. Human-in-the-loop remains essential; risk tolerance varies wildly by organization. Writer AI compressed a four-month retail workflow down to one week. Multi-provider…

More episodes of AWS for Software Companies Podcast

• Ep209: Starburst Data's Blueprint for the AI Era with AWS · June 2, 2026 · 17 min
• Ep208: Built to Survive: CockroachDB's Role in the Agentic AI Era · May 26, 2026 · 17 min
• Ep207: The AI Arms Race: How Vectra AI Uses Agentic AI to Outpace Cyber Attackers · May 19, 2026 · 14 min

Explore listener stats, chart rankings, contacts and more on the AWS for Software Companies Podcast podcast page.