AI Agents & Workflows

1270 summarised stories about AI Agents & Workflows, each linking back to the original source. Browse all topics →

+ Follow this topic

Key players

OpenAI Anthropic The New Stack SiliconANGLE Meta Claude Code

Wednesday, 18 February 2026

IBM and UC Berkeley Diagnose Why Enterprise Agents Fail Using IT-Bench and MAST

Hugging Face 8 months ago 46

IBM and UC Berkeley created MAST, a diagnostic framework that categorizes why AI agents fail in IT automation tasks rather than just reporting success rates, and applied it to analyze 310 execution traces across three language models. Gemini-3-Flash averaged 2.6 failure modes per failed trace while GPT-OSS-120B averaged 5.3, showing that smaller open-source models suffer from cascading failures that compound over time. The analysis revealed that incorrect verification (agents declaring success without checking results) is the strongest failure predictor, enabling developers to deploy targeted fixes like external verification gates instead of blind prompt engineering.

Read the full story

AI Agents & Workflows Agent Evaluation Agent Benchmarking

Partnering with Firetiger

Sequoia 8 months ago 39

Firetiger launched an observability product that uses always-on agents to monitor customer-facing behavior, detect anomalies, verify impact, and fix issues before customers notice. The product is available fully self-service for anyone to use. This shifts incident and dashboard management from engineering teams to automated agents running 24/7.

Read the full story

AI Infrastructure Developer Tools AI Agents & Workflows Agent Orchestration

The daily briefing

Every AI story that matters,
in your inbox by 8am.

TLDRocket reads all relevant sources, removes duplicate coverage, and summarises the day in two minutes. Follow companies and topics for alerts, or get the briefing in Slack. Free, no spam, unsubscribe anytime.

TLDRocket 27 October 2026

OpenAI’s push from “you prompt, it responds” to “it keeps working” took center stage. dots, an always-on personal assistant for ChatGPT Pro subscribers and higher plans, reframes OpenAI’s coding/agent capabilities as a single proactive agent that can connect to email and calendar, research upcoming work items, and then ask for timing and privacy preferences—so the assistant behaves more like a scheduler with an inbox, less like a chat box. That same agent focus shows up in the API price drop for GPT-6.1 Terra, where OpenAI cut standard pricing by about 50% versus GPT-5.6 Terra and expanded availability to ChatGPT Work and Codex users. The catch is already visible elsewhere: Workiva’s Amplify event leaned hard on enterprise governance, noting only 11% of executives say their data quality is sufficient for AI, even as 84% feel comfortable with AI outputs for annual reports without human review. The practical takeaway is that agent automation demands traceability, approval rails, and workflow redesign—because “more proactive” also means “more accountable.”

Read the full briefing →