Fiddler AI

Today we’re announcing our $30M Series C, led by RPS Ventures, bringing our total funding to $100M. This milestone reinforces our conviction, validated by customers and backed by investors: AI systems won’t be operable or trusted at scale without an enterprise-ready Control Plane for AI. Making AI systems operable, governable, and trustworthy at scale is no longer optional; it is the prerequisite for adoption. That's the Control Plane for AI, providing the visibility and authority teams need across the agentic lifecycle with: + Standardized telemetry + Reliable evaluation + Continuous monitoring + Enforceable policy + Auditable governance We’re proud to be backed by long-term partners who see the same shift and are aligned on what comes next, including RPS Ventures, Insight Partners, Lightspeed Ventures, Lux Capital, Dallas Venture Capital, ISAI Cap Ventures, Dentsu Ventures, Mozilla Ventures, BGV, LDV Partners, and LG Ventures. Read our co-founders @krishnagade and @amitpaka blog to learn more about our vision for building the Control Plane for AI: fiddler.ai/blog/series-c?… Link to press release: fiddler.ai/press-releases…

💌 Webinar Invite: Why Do Smarter AI Agents Still Break? Join Juhi Parekh, GM of Key Frontier AGI Accounts at Turing, and Buddy Brewer, VP of Product Management at Fiddler AI, to explore why capable AI agents can still fail once deployed, and how those failures can inform the next version of the model. When: July 23, 2026, 10-11am PT What You'll Learn: 🔹 Why agent failures reveal what needs to improve in the model 🔹 How teams design harder test problems before deployment 🔹 Why agents that pass benchmarks can still fail in real enterprise workflows Register now: fiddler.ai/webinars/ai-ex… #AI #ArtificialIntelligence #ML #MachineLearning #MLOps #LLMs #LLMOps #GenAI #GenerativeAI #DataScience #DataScientist #DataEngineering #DataEngineer #CIO #ResponsibleAI #AIObservability #Guardrails #LLMGuardrails #AIAgent

💌 Webinar Invitation: Fighting AI Threats With Autonomous Defense Join us to explore how AI security changes when systems move from generating content to using tools, making decisions, accessing data, and taking action. When: July 16, 2026, 10am PT What You'll Learn: 🔹 Why agentic systems require a different security model 🔹 What autonomous defense looks like in practice 🔹 How the Behavior, Identity, and Control framework helps govern autonomous AI Register now: lnkd.in/gcmWWg2K #AI #ArtificialIntelligence #ML #MachineLearning #MLOps #LLMs #LLMOps #GenAI #GenerativeAI #DataScience #DataScientist #DataEngineering #DataEngineer #CIO #ResponsibleAI #AIObservability #Guardrails #LLMGuardrails #AIAgent

In the enterprise, usage-based AI costs compound fast. And if every agent trace is scored by an external LLM, evals can become another major usage-based bill. Fiddler Centor Models reduce that Trust Tax with secure, in-environment evaluations and no per-call API cost. 📱Run the numbers with our free interactive TCO calculator. lnkd.in/gxHnDQkB #AI #ArtificialIntelligence #ML #MachineLearning #MLOps #LLMs #LLMOps #GenAI #GenerativeAI #DataScience #DataScientist #DataEngineering #DataEngineer #CIO #ResponsibleAI #AIObservability #Guardrails #LLMGuardrails #AIAgent

We'll be at AI Engineer World Fair June 29-July 2, 2026 in San Francisco. Stop by our booth to see the AI Control Plane in action and learn about how: • Fast, free evaluations can take agents from testing to production • Avoiding the trust tax without using external API calls for every eval • Complete observability for AI agents can allow you to improve ROI give them more autonomy Meet with our team to discuss how enterprises are eliminating the trust tax while maintaining full governance over their AI agents. #AI #ArtificialIntelligence #ML #MachineLearning #MLOps #LLMs #LLMOps #GenAI #GenerativeAI #DataScience #DataScientist #DataEngineering #DataEngineer #CIO #ResponsibleAI #AIObservability #Guardrails #LLMGuardrails #AIAgent

🚀 Introducing the Fiddler Control Plane for Coding Agents. Coding agents are already inside your business. They are reading files, calling tools, and making decisions faster than any human reviewer can keep up with. Most teams can't see what they're accessing, spending, or exposing. This is not a coding agent problem. It is a visibility and control problem. We’re delighted to extend the Fiddler Control for coding agents with enhanced guardrails and observability. 🔹It is the only inline enforcement at the agent’s request and response paths for PII/PHI and secrets detection and redaction. 🔹It now provides fleet-wide intelligence across every developer, token, and dollar so that you can track and plan AI spend. #tokenmaxxing anyone? 🔹All this happens through the gateway you already run for added flexibility. No agent-side integration needed. Learn more 👉 buff.ly/l7kOSPm #AI #ArtificialIntelligence #ML #MachineLearning #MLOps #LLMs #LLMOps #GenAI #GenerativeAI #DataScience #DataScientist #DataEngineering #DataEngineer #CIO #ResponsibleAI #AIObservability #Guardrails #LLMGuardrails #AIAgent

An unnamed enterprise reportedly racked up a $500M Claude bill in one month. The issue is more than just usage. We are noticing the same pattern in AI evaluations, where every agent trace is being scored by an external LLM and API costs are scaling fast. Fiddler Centor Models are purpose-built to reduce this cost by running securely in your environment and eliminating your per-call API cost. See what your real evaluation TCO looks like with our free interactive TCO calculator. fiddler.ai/evals-tco-calc… #AI #ArtificialIntelligence #ML #MachineLearning #MLOps #LLMs #LLMOps #GenAI #GenerativeAI #DataScience #DataScientist #DataEngineering #DataEngineer #CIO #ResponsibleAI #AIObservability #Guardrails #LLMGuardrails

Did you hear about the fintech team that had a $283K monthly LLM evaluation bill? Their primary inference cost was $190K. That means their evals cost more than the AI application itself. This has always been the risk with external LLM-as-a-Judge. Fiddler Centor Models are purpose-built to contain this risk by running evals securely in your environment and eliminating your per-call API cost. Run your evaluation TCO across Centor, GPT, Claude, Gemini, and more with our free interactive TCO calculator. fiddler.ai/evals-tco-calc… #AI #ArtificialIntelligence #ML #MachineLearning #MLOps #LLMs #LLMOps #GenAI #GenerativeAI #DataScience #DataScientist #DataEngineering #DataEngineer #CIO #ResponsibleAI #AIObservability #Guardrails #LLMGuardrails #AIAgent

See you next week at booth 533 of Databricks Data + AI Summit June 15-18 in San Francisco! As enterprises deploy more AI agents, the gap between development and production oversight grows wider. We'll be showcasing the AI Control Plane, purpose-built for: • Standardized telemetry that discovers and monitors all your agents • Continuous evaluation without external API costs • Enforceable guardrails across input, execution, and output • Enterprise-grade audit trails for compliance and oversight Stop by our booth to learn more! #AI #ArtificialIntelligence #ML #MachineLearning #MLOps #LLMs #LLMOps #GenAI #GenerativeAI #DataScience #DataScientist #DataEngineering #DataEngineer #CIO #ResponsibleAI #AIObservability #Guardrails #LLMGuardrails #AIAgent

AI agent evaluations are becoming the hidden cost of deployment. CIO TIMES recently called out how as enterprises move from AI pilots to agentic systems in production, many teams are underestimating the cost of testing and evaluating agent outputs. When every trace may need to be evaluated for quality, safety, or performance, costs can scale fast with external LLM calls. That is the Evals Trust Tax. Fiddler Centor Models are purpose-built to reduce that cost by running securely in your environment and eliminating your per-call API cost. See what your real evaluation TCO looks like across GPT, Claude, and Gemini with our free interactive TCO calculator. fiddler.ai/evals-tco-calc… #AI #ArtificialIntelligence #ML #MachineLearning #MLOps #LLMs #LLMOps #GenAI #GenerativeAI #DataScience #DataScientist #DataEngineering #DataEngineer #CIO #ResponsibleAI #AIObservability #Guardrails #LLMGuardrails #AIAgent

Harnesses get agents running but a control plane is needed to keep them under control. There's confusion in the market between agentic harnesses (LangChain, CrewAI, Claude Code, Bedrock) and control planes. Both solve fundamentally different problems: 1️⃣ Harnesses define how an agent runs - execution environment, iteration loops, tool registries, memory persistence 2️⃣ A control plane governs what agents do across all harnesses - policy enforcement, audit trails, compliance oversight Enterprises will deploy multiple harnesses across teams. But you need a single control plane to standardize governance regardless of which harness is running. With coding agents, the stakes are higher. When Claude Code writes code, calls APIs, and accesses enterprise systems autonomously, governance cannot start at runtime. It has to start at the code generation boundary. High-performing agents at enterprise scale need both a harness to run and a control plane to govern them. Read more on our blog: fiddler.ai/blog/agentic-h… #AI #ArtificialIntelligence #ML #MachineLearning #MLOps #LLMs #LLMOps #GenAI #GenerativeAI #DataScience #DataScientist #DataEngineering #DataEngineer #CIO #ResponsibleAI #AIObservability #Guardrails #LLMGuardrails #AIAgent

The Claude Code leak revealed something critical: every coding agent in your environment has the same attack surfaces. The leaked source showed Claude Code executing bash commands, reading files, making HTTP requests, and spawning sub-agents - all autonomously, often with full OS access. These capabilities exist in every enterprise coding agent today Most enterprises lack visibility into: • What their coding agents are actually doing • What data they're touching • What happens when something goes wrong Coding agent productivity gains are real. But so are the risks at enterprise scale. The solution isn't avoiding coding agents, but instead building proper controls. Observability must precede autonomy. You can't give agents more autonomy than your ability to oversee them. Read more on our blog: fiddler.ai/blog/claude-co…

AI agents are facing an identity crisis. As they become more autonomous, the fundamental question is about more than agent capabilities: it's about trust and control. Our latest blog post is live, discussing how: • Agents require standardized telemetry to capture every decision point • Reliable evaluation systems must run continuously, not just during development • Policy enforcement needs to happen at input, execution, and output stages • Governance frameworks must bridge technical teams and compliance requirements Without proper observability and control mechanisms, autonomous agents become liability generators rather than value creators. The path forward requires treating agent oversight as seriously as agent development. Read more: fiddler.ai/blog/ai-agents…

Catch the full replay of AI Explained: AI Agents Have an Identity Complex - now available on demand! 1Password's VP of Engineering, Jeff Malnick, dives deep into why static security breaks for AI agents, Zero Trust at runtime, and auto-scoping agent credentials. Watch the full session now: fiddler.ai/webinars/ai-ex… #AI #ArtificialIntelligence #ML #MachineLearning #MLOps #LLMs #LLMOps #GenAI #GenerativeAI #DataScience #DataScientist #DataEngineering #DataEngineer #CIO #ResponsibleAI #AIObservability #Guardrails #LLMGuardrails #AIAgent

Your evaluation TCO may be higher than you think. Every time an external LLM evaluates a trace, your LLM provider charges for it. That's your Evaluation Trust Tax, and it grows with every trace, every metric, and every token. That's why we built the Evaluation TCO Calculator. Compare external LLM evaluation costs against Fiddler Centor Models,* purpose-built, low latency, and cost-effective. Play around with the calculator and discover your TCO, whether you're running simple tasks or complex reasoning projects with LLM-as-a-judge. fiddler.ai/evals-tco-calc… *Fiddler Centor Models were previously known as Fiddler Trust Models.

Don’t miss tomorrow’s webinar! We've partnered with @1Password to unpack the identity challenges hiding inside every agentic deployment 📅 Register now to learn how to: 🔹 Separate agent identity from traditional machine identity 🔹 Understand why static security policies break when agents can reason and decide their own scope 🔹 Bring Zero Trust enforcement into real-time agent workflows with attestation and automated credential scoping All registrants receive an on-demand replay. Register now 👇fiddler.ai/webinars/ai-ex… #AgentIdentity #AI #ArtificialIntelligence #ML #MachineLearning #MLOps #LLMs #LLMOps #GenAI #GenerativeAI #DataScience #DataScientist #DataEngineering #DataEngineer #CIO #ResponsibleAI #AIObservability #Guardrails #LLMGuardrails #AIAgent

@1Password's Jeff Malnick has been thinking about agent identity longer than most. As VP of Engineering, Developer and AI at 1Password, he works on the security challenges that come with deploying AI agents in production, including how to scope, provision, and audit agent access in real time. Join Jeff and Krishna Gade for AI Explained: “AI Agents Have an Identity Complex” in one week. Save your spot for  May 14, 2026 at 10am PST: fiddler.ai/webinars/ai-ex… #AgentIdentity #AISecurity #AI #ArtificialIntelligence #ML #MachineLearning #MLOps #LLMs #LLMOps #GenAI #GenerativeAI #DataScience #DataScientist #DataEngineering #DataEngineer #CIO #ResponsibleAI #AIObservability #Guardrails #LLMGuardrails #AIAgent

You’re invited to our next AI Explained: “AI Agents Have an Identity Complex” 🔐with @1Password VP of Engineering, Developer and AI, Jeff Malnick! AI agents need access that changes in real time as they reason through what to do next. Most security frameworks were built for humans, not systems that decide their own scope. We will be covering: → What separates agent identity from machine identity and why static security policy breaks when agents can reason → Where Zero Trust enforcement has to live in an agent workflow and why real-time attestation beats standing policy → How to provision and scope agent credentials automatically without blocking the systems that depend on them 📅 May 14, 2026 at 10am PST 🔗 Register: fiddler.ai/webinars/ai-ex… #AgentIdentity #AISecurity #AI #ArtificialIntelligence #LLMOps #GenAI #AIAgent

AI coding sessions are ephemeral islands. Your agent crafts brilliant implementation plans, works through edge cases, makes architectural decisions. Then the session ends and all that reasoning evaporates. The next session starts fresh. The next engineer starts fresh. The code ships, but the reasoning lives only in memory. Most teams resort to copy-pasting between sessions, Google Docs, and Slack. The agent loses context every time. You spend ten minutes re-explaining the same problem. We solved this by using GitHub as a persistence layer for AI agent sessions: • Plans and decisions survive session restarts • Teammates can review and comment without context-switching • Agents reload full context from issues when work resumes • Complete audit trail of what changed and why AI agent sessions don't have to be throwaway conversations. When you make them collaborative and auditable, they become part of your development workflow instead of working around it. fiddler.ai/blog/github-ai…

Not many people can say they govern AI as a health system executive and use those same tools on patients every week. Stacey Johnston can. She is joining our next AI Explained webinar as Chief Information and Digital Execution Officer at Beacon Health System and a practicing hospitalist. Her take on AI governance, ROI mandates, and what clinical reality demands from technology is one you won't hear anywhere else. One week away. Grab your spot: fiddler.ai/webinars/ai-ex… #AI #ArtificialIntelligence #ML #MachineLearning #MLOps #LLMs #LLMOps #GenAI #GenerativeAI #DataScience #DataScientist #DataEngineering #DataEngineer #CIO #ResponsibleAI #AIObservability #Guardrails #LLMGuardrails #AIAgent

We are excited to announce our next AI Explained, "Lessons from a Physician-CIO on AI Governance" 🏥 In healthcare, every AI decision has a patient on the other side of it. That changes everything about how governance gets built. Joining us is Stacey Johnston, MD, Chief Information and Digital Execution Officer at Beacon Health System and a practicing hospitalist. She is the executive approving AI tools and the physician using them on patients. We will be talking about: → What a structured AI approval process looks like when patient safety sets the standard → How ROI gets defined, measured, and enforced before any tool scales → What practicing medicine adds to governance decisions that spreadsheets alone can't capture 📅 April 23, 2026 at 10am PST 🔗 Register: fiddler.ai/webinars/ai-ex…