Hi,
Hope you have a great July 4th coming up. The last two weeks have been full of releases, a couple of milestones, and some fresh learnings from the field.
A few things we’re putting in front of you:
We crossed 100 organizations on the self-serve tier, just a month and a half after launch. Retail, ed tech, manufacturing, B2B SaaS, consultancies, and we’re glad to have every one of them.
A new customer story from life sciences, on how Genloop’s living context graph brought precision where a query-banking setup kept breaking.
The conversations with data leaders continue, and one thing stood out this time: how AI has added more noise, and made the signal harder to find.
Let’s get into it.
We opened Genloop’s self-serve tier six weeks ago, with no marketing budget behind it. As of this week, more than 100 organizations have connected, across retail, ed tech, manufacturing, B2B SaaS, and consultancies.
The part we like most is quieter. We run our own analysis on Genloop too, connected straight to our application and usage databases. This week we asked it for cumulative user growth since launch. No table names, no schema, just the question a business user would actually ask. It queried, charted, and narrated the answer back. That’s the person we built this for. The one who cares about the answer, not the query.
A store operator, two decades in, at one of the largest grocery chains in the US. He spotted a staffing problem that cost $200K to ignore. All he needed was his sales data laid against his schedule. It took a month and a half to get it.
The data existed the whole time. He just couldn’t reach it. That’s not a rare story. That’s the normal experience for most of the people actually running these businesses.
Next in our Agentic Analytics Bites series, we sat down with Derek Lilley. He spent a decade in data analytics across Mars and McDonald’s, on store-level performance, pricing, and forecasting, and recently founded Storied Analytics. We asked him how AI is changing analytics in CPG.
His answer wasn’t the one we expected. “In the absence of data science, you could find five plausible answers to explain an outcome. Marketing runs with one story. Sales runs with another.” Then the turn: “LLMs are actually creating more stories. People are running around with more versions of the truth.”
Most people frame AI as the fix for the noise. Derek thinks it’s making the noise worse. The hard part was never generating explanations; AI does that easily now. The hard part is knowing which one is true, and remembering it next time. He didn’t name the company Storied Analytics by accident.
Databricks just launched Genie One, an AI coworker that answers in plain English over governed data and takes action instead of stopping at a chart. It’s a strong product. It also assumes Databricks sits at the center of your stack.
Most teams don’t run on one clean platform. Data lives across Snowflake, BigQuery, Postgres, Redshift, and a dozen apps. So the question is which assumption matches your reality. Genie One reaches outside data through federation and stays on managed cloud. Genloop reads every source in place with zero copies, and runs anywhere: cloud, on-prem, VPC, or fully air-gapped with no external LLM calls.
Consolidated on Databricks? Genie One is a natural fit. Spread across clouds, or in a regulated environment? That’s the gap we fill. Full breakdown on the blog.
A leading life sciences and analytics company was building conversational intelligence into their product so commercial teams could talk to their data. The hard part was accuracy under complexity, with compliance rules and business logic that had to hold every single time.
Their first attempt, query-banking on a RAG setup, kept hitting walls. 90% accuracy on known questions, 76% on anything new. Every new client meant two-plus months of engineering and a freshly curated query bank. Roughly 300k tokens burned per query. Governance buried in instructions.
Then they put Genloop’s living context graph underneath their own agents and UI.
Accuracy up 22 points.
Onboarding cut from months to days.
10x fewer tokens per query.
Governance that’s deterministic, not hopeful.
One thread ran through everything this week: the tools keep getting better, but the real question hasn’t moved. Can the people actually running your business get a trusted answer from their data, the moment they need it? That’s what we’re building toward.
If you haven’t tried Genloop yet, connect your warehouse and start free. No demo, no sales call. We’d love to hear what you make of it.
Thanks for reading.
Ayush
CEO, Genloop
No posts
Comments
Nothing yet. Say the first thing.
Sign in to join the conversation.