Taryn Plumb

New VentureBeat Pulse Research finds enterprises run three AI orchestration platforms at once, yet 21% still can't stop runaway agent spending in real time.

Target SVP Siobhán Mc Feeney says AI models aren't the company's real edge — the governance layer it built so every agent has to earn its autonomy is.

One startup said it cut cyberattack containment time from seven hours to twelve minutes. See four other approaches to running AI agents safely at scale.

At VB Transform 2026, Instacart CTO Anirban Kundu said AI-generated code gets rebuilt so often that tech debt has stopped being a concern for his team.

Liberty Mutual built an AI backbone with fully swappable models underneath — so when Anthropic's Fable 5 went offline, the insurer simply routed around it.

Shopify built an LLM proxy and distillation pipeline so its engineers keep working when any model goes away — and often get better results from smaller tuned models.

Stanford's DeLM lets AI agents coordinate without a central controller, cutting multi-agent inference costs 50% and beating SWE-bench baselines by 10.5%.

MassMutual caps vendor contracts at 12 months and runs a multi-model architecture — cutting contact center resolution times from 10 minutes to one.

Pinterest CTO Matt Madrigal explains how gutting Qwen3-VL's vision layer and replacing it with proprietary embeddings cut AI costs 90% at 620M users.

Merck cut a drug discovery cycle by 33% and ships compliant marketing 80% faster. Mastercard is rethinking fraud disputes. Both say infrastructure came first.

RAG retrieves documents but not decision logic, causing agents to act on expired rules. Decision context graphs encode applicability and time-scoped memory.

LlamaIndex CEO Jerry Liu argues the framework era is over: agent loops are now capable enough that context quality is the real competitive edge.

Stanford's 2026 AI Index: frontier models fail one in three attempts, lab transparency is declining, and benchmarks are saturating faster than they're replaced.

Outshift by Cisco's Vijoy Pandey argues AI agents can connect but can't yet think together — and is building the protocols to close that gap.

AI agents are replacing traditional search for serious work — and LLM-referred traffic converts at 30-40%, far above SEO and paid social. Here's what enterprises need to know about answer engine optimization (AEO) and how to make their content citable before competitors do.

Technology leaders from MassMutual and Mass General Brigham share how they shut down ungoverned AI pilots and got to production — with results including 30% developer productivity gains and help desk resolution times cut from 11 minutes to one.

Intuit shipped AI agents to 3 million customers and hit 85% repeat usage — but the key to adoption wasn't better models. It was keeping human accounting, tax and payroll experts in the loop at critical moments. Here's how the company built it and what comes next.

AI agents fail in production for predictable reasons: fragmented data, undefined workflows, and runaway escalation. Burley Kawasaki of Creatio outlines three disciplines — data virtualization, bounded use-case loops, and agent monitoring with real KPIs — that enterprise teams are using to reach 80–90% agent autonomy without multi-year data overhauls.

LangChain co-founder and CEO Harrison Chase explains why harness engineering — not just smarter models — is what gets AI agents from prototype to production.

Intuit lost 42% of its market cap as AI agents threaten to replace QuickBooks and TurboTax. Here's what the company says agents can't replicate.