Datacoves

Big news from the Data & Analytics Summit Chicago. Our co-founder Noel Gomez is taking the stage to dig into a question every CPG data team is wrestling with right now: what has to be true before you put AI in front of business users, without losing control of the numbers. No hype. Just the real groundwork: governance, trust, and guardrails that let AI actually earn its seat at the table. Catch the session if you're in Chicago.

⛵ Come join the Datacoves Yacht Club. We'll be waiting at booth 208. Mike Doll, Guitar Center's VP of Data is speaking today at 4:15, Reliable by Design in Room Gracia 8, Level 3 Swing by before or after; can't wait to see you! #dbtsummit

Today at 4:15 pm Mike Doll, VP of Data at The Guitar Center Company, presents Reliable by Design: Lower Cost, Better Delivery, Trusted AI. Location: Level 3, Gracia 8 We are at booth 208; stop by if you want to learn how you can achieve similar results. dbt Labs #dbtSummit

Today we are launching Atlas. Atlas is the agreement layer. It helps teams and their AI agents settle what data means before anyone answers a question, and turns that agreement into a governed Definition of Record that is versioned, attributed, and owned by you. It works with the warehouse, dbt, and tools you already have, over the open Model Context Protocol. No migration. No lock-in. Run it as a managed service, or entirely inside your own network. https://datacoves.ai

Three things before you land in Vegas. 1. Mike Doll's session, Reliable by Design, Wednesday at 4:15 pm. Their journey to reliability. 2. Booth 208, near the main entrance. Find us there all week. 3. There's a scratch-off in every swag bag. Bring yours to our booth to claim a prize. Book focus time: https://lnkd.in/gMN2B2wg dbt Labs #dbtSummit

We are doing two big things this week. 1️⃣ We are a Silver Sponsor at dbt Summit 2️⃣ We are launching a new product I talk a lot about our philosophy. Years ago, we took a small investment from TinySeed for mentorship, not the growth-at-all-costs VC path. We do what we tell our customers to do. Build a solid foundation, go slow to move fast, and focus on the long term. Along the way, we gained the trust of customers like Johnson & Johnson, The Guitar Center Company, Orrum Clinical Analytics, and Sevita. Different industries, different problems, same reason they stay: we did not cut corners, and they don't worry about infrastructure. We also accelerated their data stack maturity and set the foundation needed for AI. It's been a rewarding journey. See you at booth 208 this coming week. #dbtsummit

Anthropic ran a self-service analytics experiment. Without a structured way for Claude to reliably find and use the right definitions, its accuracy on internal analytics questions did not exceed 21%. Add a layer that routes it to the correct governed source before it ever writes a query, and that number goes above 95%, regularly near 99% in some domains. Same model both times. They also tried an automated shortcut: have an LLM draft the underlying business definitions, pulled from raw tables and old queries. It backfired. The AI-written definitions just encoded the same ambiguity everyone already disagreed on, dressed up to look decided. Their fix was not more automation. It was a person owning each definition, with AI helping document it, not decide it. Everyone is focusing on how technology will "fix" the deficiencies in the org. Almost nobody is fixing the thing that determines whether the answer is right. Come to Datacoves booth no 208 and see how the agreement layer fixes this. Hint: it's not more work for the tech team. dbt Labs #dbtSummit

When you are early in building your data platform, you do not know what you do not know. You talk to a warehouse vendor, an ingestion vendor, an orchestration vendor, a BI vendor, a migration vendor. Each one is honest about their part. None of them can describe the whole elephant, because none of them is responsible for the whole elephant, and none of them lives with the consequences if your elephant turns out to be a whale. This is not a knock on any of them. No one selling you the trunk is lying about the trunk. They just cannot tell you what the tail needs, because that was never their job. We do not show up and deploy dbt and Airflow and call it a day. We tell you there is real work: understanding your pain points and goals, figuring out what is holding you back, architecting the foundation that holds up for the next ten years, not just through the migration. We know that moving fast means moving a little slow at the start. Your org owns the agility, user experience, and reliability of your platform. You need to drive the migration because tools alone won't truly transform your org. The job to be done is not to "stand up new tools." It's to see the whole elephant at once so the platform holds together instead of having stakeholders optimize only their slice. We have done this enough times that we can show a team the whole elephant early. Our customers are loyal because we don't sugarcoat the truth; we help them chart the map to get to a mature process as quickly as possible. Because we know what true transformation involves.

One week out. Here is where to find us at dbt Summit. Booth 208, near the main entrance. We are a passport program stop, so you can get your card marked while you are there. Also worth knowing: there is a scratch-off card in every swag bag, and winners claim at our booth. If you would rather book dedicated time: https://lnkd.in/gMN2B2wg dbt Labs #dbtSummit