Uber used up its 2026 AI budget in four months. Before we scale an agent, I want to know what a finished task costs. That's why we start with one real process, at real volume, using the customer's own numbers. FT: hubs.ly/Q04BWYdX0 #AIAgents #EnterpriseAI #Meibel

Three spec sections referenced in the drawings. None in the spec book. On Nov 5, Brian Sampson, Aaron Aguillard and I will read two real bids live and show how Tetheree brings answers into Procore. 1 PM ET / 10 AM PT. Join us: hubs.ly/Q04BK95D0

AI that's 95% right still needs every page checked. Agents need data they can trust, controls on what they can do, and answers traced to the source. We built Meibel for that. I'm on stage at Venture Atlanta Oct 14. Come find me. #VentureAtlanta #AIAgents #EnterpriseAI #Meibel

Pilots prove an AI model can do the task. Production proves the system can keep doing it at volume. Meibel preserves document structure and sources, scores confidence, enforces access controls and retries failed work. #AIAgents #EnterpriseAI #Meibel

At Venture Atlanta, I get 3 minutes on stage. My point: AI automation breaks when errors compound. The fix isn’t a smarter agent. It’s better data—documents understood, structured, scored, and traced to source. Documents become data. Data becomes agents. #VentureAtlanta

AI that’s 95% right still needs someone checking the other 5%. That’s the problem with agents working from documents. We built Meibel so documents become reliable data, data powers agents, and every answer traces back to its source. I’m on stage at Venture Atlanta, Oct. 14.

Boston today for the AI & Data Transformation in Construction Summit. Construction already has the data AI needs in specs, drawings, RFIs, and addenda. We’re showing how Meibel turns it into connected data agents can use. #AIConstructionBoston #Meibel

We’re heading to Venture Atlanta. Meibel has been selected to present on stage. AI shouldn’t just be a second reader. It should run real work. That’s what we’re building: Documents become data. Data becomes agents. Oct 14–15 | Atlanta #VentureAtlanta #AIAgents #EnterpriseAI

Already using AI for documents? Most teams are. The challenge starts in production. Meibel unifies document understanding, agent execution, and governance in one platform, with no external vector DBs, glue code, or audit gaps. See how your stack holds up.

We're bringing our construction stickers to Boston! “Document AI Built for How You Build” captures what I'll discuss at my AMG roundtable: helping agents understand complex construction data. Find Aaron and me on Sept 29. Grab a sticker and say hi! #AIConstruction

Biotech AI needs more than plausible answers. It needs traceable evidence, confidence scores, and complete context. Aaron Aguillard will be at #BioTechX USA on Sept 30 at Boston’s Hynes Convention Center. Working on AI in life sciences? Connect with him. #Meibel

Tomorrow: Joe Small of York IE joins us to break down the document pipeline, from ingestion to agent action. We’ll cover structured context, provenance, confidence scoring, and execution policies. Sept 16 | 1 PM ET / 10 AM PT hubs.ly/Q04xxZwG0

Heading to Boston for the AI & Data Transformation in Construction Summit. I’ll be speaking on Track 4: Unlocking Your Enterprise Data to Build Agents, 2–3 PM. Sept 29 | Hyatt Regency Boston Come join the conversation. #AMGWorld #AIConstruction #EnterpriseAI #Meibel

Next Wednesday, I’m joining Aaron Aguillard and Joe Small of York IE to walk through the document layer beneath production AI agents. Register: hubs.ly/Q04wXB2r0

Six important updates shipped in Meibel in August. And anyone can start building today. Self-signup is live. A few clicks and you're in. Six updates. One direction. Structure preserved, steps removed, more people in the door.

Meibel is live on AWS Marketplace. For teams already on AWS, this matters. Available now on AWS Marketplace as SaaS, Bring-Your-Own-Cloud on AWS, or fully on-premises. More details here: hubs.ly/Q04vdYBS0

Batch processing gets ugly fast once you stop talking about a few hundred documents. That is why we built batch execution with a cost estimate and approval gate before the job starts, QA sampling while it runs, and checkpoint recovery if something breaks halfway through.

Meibel is officially live on the Microsoft Azure Marketplace. If you are running on Azure and need to move document AI from proof-of-concept to production, reach out and we'll set up a private offer for your team: hubs.ly/Q04tysPp0