Vellum

🚨🎁 Masterclass to becoming AI native from our team to yours Sing up here 👉 https://lnkd.in/gPnhN9-n

ATTENTION AI BUILDERS OF NYC 📣 Come have a drink with us in our new office 👉👈 AI industry titans and automation enthusiast will be helping us break the new space in! Don't miss out RSVP here: https://luma.com/wp6txvf5

In the AI race you either move fast or get left behind. Sitting on the sidelines is a compounding disadvantage costing time, opportunity, and a competitive edge. Yesterday’s “nice to have” is tomorrow’s “must have”. Here are some facts: - According to Stanford's HAI “AI Index 2025” report, 78% of organizations reported using AI in some function in 2024. - The global AI market: valued at ~$391 billion in 2025 and forecast to grow ~9× by 2033. - For generative AI specifically, expected annual growth (CAGR) of ~41.5% from 2025-2030. - Organizations with defined AI adoption strategies are twice as likely to see revenue growth compared to those without. If you’re not already asking “How can we build AI agents now that directly serve our users, automate meaningful but slow processes”, let’s change that. Here’s what you can do in the next few days: 1/ Identify one process in your team or org that’s repetitive, slows you down, or requires human follow-up. 2/ Sketch how an AI agent could take over, augment, or accelerate that process 3/ Login to vellum.ai and try to build it (prompt it!). AIAGENTS25 for 25% off! 4/ Let me know what you build - I’d love to promote it! 5/ Learn from this use case and start with step 1 again P.S: I had fun moving fast at Times Square today in front of our billboard. Thank you Omid Izadjou and Arc for getting us featured on Times Square!

Amazing to see how it only took Autobound a couple days to deploy an MCP server using Speakeasy to connect their data as a tool for their Vellum agents. This is the perfect example of what happens when great infra meets a great agent platform! 🤝

Vellum engineers shipping at ⚡️ speed!

The Vellum 8 balls we gave out at AI Agent World Tour NYC showed the future of agent building... Spoiler: it’s about prompt based building We brought that future to life in our hands-on workshop, where Anita Kirkovska and David Vargas Fuertes guided builders through creating agents just by prompting Vellum. No drag and dropping. No coding. Just natural language + Vellum. For the rest who missed out, our AI oracle Sidd Seethepalli forecasted why we (hopefully) may never need to build another agent again, breaking down: 1/ What we’ve learned from helping teams move beyond drag-and-drop frameworks with prompt based building 2/ Where better tool definitions, smarter testing, and full visibility make or break your agents By the time builders made it to our booth, the 8 Balls didn’t need to predict anything. Nico Finelli and Matt Mirick walked builders of all levels through the reality of prompt building agents on Vellum today! Massive thanks to the MLOps community team for putting together an incredible event, and to everyone who stopped by to see the magic of prompt building agents firsthand. See for yourself how easy prompt-based agent building can be on Vellum 👉 here: https://www.vellum.ai/

Turns out “No drag, no drop” is a bit of a crowd magnet. Most builders at AI Engineer Paris had built workflows the old way: connecting nodes, tweaking endless settings, and hoping everything works at run time. Few had ever built them from prompting. It didn’t take long for builders to start taking turns at the keyboard once Nico Finelli showed what Vellum could build off of a single prompt. Big thanks to everyone who stopped by the booth and built with us! This is just the beginning 🚀 Check out Vellum for yourself and see how fast it is to go from idea to agent in minutes! 👉 https://www.vellum.ai/ A special thanks to AI Engineer for putting together a conference built by builders, for builders. We can't wait to see what will be showcased at the next one!

Google Cloud and Vellum are back at it again! This time with a new tutorial on how to bridge your agents and data using Google’s MCP Toolbox for Databases and Vellum. 🤝 MCP is reshaping what AI agents can do. Google’s MCP Toolbox turns agents into data-aware systems that can query, interpret, and act on live databases. Combine Google’s toolbox with Vellum and you’ll get: - Tool discovery and integration (no custom endpoints) - Secure and consistent access to live, structured data - Reusable tool definitions that work across agents Explore the full walkthrough to learn how to: - Use an existing PostgreSQL database for your Agent - Define custom MCP tools for queries - Connect everything to an Agent in Vellum using the MCP toolset from Google’s Agent Developer Kit, with best practices for evaluations and monitoring built in Here 👉 https://lnkd.in/g_TE42Qe Big thanks to Tai Conley, Hasan Rahman, Drew Oros, and Kashaf Mazhar from the Google Cloud team for collaborating with David Vargas Fuertes, Vincent Hsieh, and Anita Kirkovska on this, and helping push the ecosystem forward!

No one on your team should have to dig through PDFs just to find a security policy. Enter Trust Center RAG Chatbot An agent that instantly answers compliance and security questions using your own trust documentation to: - Retrieve accurate answers directly from your Trust Center - Cite sources for full transparency - Handle follow-up questions with context - Work across your website, Slack, or internal tools Your team no longer has to search through endless files or ping security every time a question comes up. Turn this workflow into an AI App, so your team can use this chatbot to: 🟢 Answer security-related inquiries from employees or clients 🟢 Provide quick access to specific policies during audits 🟢 Assist compliance teams in responding to security questionnaires Clone the template, drop in your security docs, and spin up your own trust center chatbot in minutes here 👉 https://lnkd.in/gXpa2aU3 Which team in your company would use this chatbot the most?

Complex agentic workflows can take weeks of building, but in Vellum? It only takes a couple minutes. Most agent builder platforms can handle low stakes automations. Vellum unlocks complex workflows others can’t handle, with a single prompt. Akash Sharma built a healthcare claims review and error detection workflow that: - Extracts data from medical claim documents - Parses CPT + ICD codes with regex - Cross-checks codes against a vector database of billing guidelines - Flags anomalies through parallel GPT-4.1 nodes In minutes. When the workflow needed custom logic, Vellum instantly built a custom node with code. No drag and dropping. No need for engineers to add to their sprint plan. Just a single prompt. See for yourself here: 👉 app.vellum.ai What’s the first complex workflow you’d build?