Fiserv's AgentOS, Agentic Red Teams Go GA, and Big Tech's Agent-Driven Workforce Shift
Tools
Sweet Security's Sweet Attack Ships Continuous Agentic Red Teaming for Live Production Environments
Sweet Attack GA's an AI red-team agent that runs continuous attack-chain discovery against your real runtime topology -- not a synthetic model of it -- including shadow IT and undocumented MCP servers. At a moment when AI-assisted attacks are compressing exploit timelines to under 44 days, periodic human red teams are no longer fast enough; this is the automated counter.
Analysis
136,000 Tech Layoffs in 2026 as Amazon, Meta, and Oracle Explicitly Cite Agentic AI Workflows
2026 is tracking as the worst year for tech employment on record, with companies now explicitly naming agentic workflow integration as the driver -- not just post-pandemic corrections or cost cuts. Amazon confirmed its latest round of layoffs targets middle management and solution architects in AWS, the exact roles agentic automation is absorbing first.
GitLab Restructures for the Agentic Era: 30% Country Footprint Cut, Savings Reinvested in AI Agents
GitLab is flattening management and cutting its global footprint as agentic coding shifts the unit of platform value from seats to tasks. The deeper signal here is that GitHub already froze new Copilot signups because agent-driven coding sessions broke the economics of unlimited-use pricing -- the whole dev tools pricing model is getting repriced around agents.
A detailed breakdown in The Hacker News shows that time-to-exploit has collapsed from over 700 days in 2020 to just 44 days in 2025, with 28% of CVEs now exploited before patches even ship. The direct line from agentic coding tools to offensive capability acceleration is now backed by hard incident data -- builders need to factor this into their threat model.
News
Fiserv and OpenAI Form Strategic Collaboration to Build Banking Agents at Frontier Scale
Alongside the agentOS launch, Fiserv and OpenAI formalized a broader collaboration covering agent co-development, core modernization, and cybersecurity -- with joint teams already working across all four focus areas. This is the clearest signal yet that core banking infrastructure vendors are racing to become agent deployment platforms, not just software vendors.
Broadridge Ships Production-Ready Agentic Post-Trade Automation with Built-In Audit Trails
Broadridge launched production-ready agentic capabilities (May 13) that chain ontology-backed data normalization with supervised agent workflows to automate exception resolution across post-trade and client-services ops. For anyone building in financial infrastructure, this is the clearest example this week of how to package agentic automation for a regulated environment without skipping compliance controls.
OpenAI formally spun up a Deployment Company with 19 global partners (including Bain, which invested) to take enterprise agent rollouts from pilots to production at scale. Bain's accompanying research pegs the agentic AI SaaS opportunity at $100B in the US alone -- and enterprise now makes up over 40% of OpenAI's revenue, on track for parity with consumer by end of year.
Stay Ahead
Delivered each morning.