QueryStory: Making AI Data Analysis More Trustworthy
A new startup wants enterprises to stop guessing whether their AI-generated data insights are correct, and actually be able to prove it.
Key points
- QueryStory has raised $6 million from Brightmind Ventures and New York Life Ventures.
- Co-founders include CEO Shapor Naghibzadeh, CTO Stanley Yang and CPO David Glusic.
- The platform surfaces confidence indicators so users can see why an AI analysis reached its conclusion.
- It targets large enterprises whose staff lack dedicated data science support.
Shapor Naghibzadeh spent years at Google tracing cyberattacks through tangled networks, learning that verified knowledge is worth the effort. That experience shaped his view of what AI-powered data analysis should look like: not a fast guess, but a grounded answer you can defend. He co-founded QueryStory, where he is CEO, with CTO Stanley Yang, a former Google engineer and ex-lead engineer at EvolutionIQ, and CPO David Glusic, who previously worked at Accenture.
QueryStory emerged from stealth this week carrying a $6 million seed round raised in late 2025, valuing the company at $60 million. It targets large enterprises sitting on proprietary databases, giving sales teams and operations managers a single place to query data and review the results.
Tim Del Bello, a partner at New York Life Ventures and an investor in the round, is also a user. He told TechCrunch the platform "was built for people like me: decision-makers seeking the ground truth who need to work with complex, disparate data sources but don't have a data science or BI team at their disposal." What once required several people to produce a quarterly business review now runs as a real-time dashboard.
How does QueryStory work?
The platform turns a database query into a narrative: users ask questions, the system assembles the answers into a structured story tied back to the underlying data. It generates visualisations and dashboards, and crucially, it surfaces a confidence indicator explaining why each conclusion holds. SQL queries appear automatically, and users can flag any analysis for a colleague to review, with that review logged in the platform.
A practical illustration: a database of satellite activity that took a developer several weeks to visualise manually produced sophisticated dashboards inside a few hours during a test run.
Should users be worried about AI reliability?
Fairly, yes. Tayler Sipperly, a partner at Brightmind Ventures, told TechCrunch that "AI is more brittle than people realize when it comes to building things that have to be durable and have large scale businesses relying upon them." Naghibzadeh agrees, and points to a specific failure mode: when thousands of employees each query a company database separately, they generate competing versions of the truth with no shared record tying any of it back to the data.
QueryStory is built to be model-agnostic, meaning it is not locked to a single AI provider, though it currently leans on the latest frontier models. The pitch to finance chiefs is predictable cost: rather than charging per token or per compute unit, QueryStory sells the reliability of the answer. That is a meaningful distinction worth watching as enterprises move from experimenting with AI to depending on it.
Common questions
How does QueryStory ensure AI accuracy?
QueryStory includes a confidence indicator that explains why its AI analyses are considered accurate. It also allows users to review and flag data for human oversight.
Who benefits most from QueryStory?
Large enterprises with sizeable proprietary databases and decision-makers who lack a dedicated data team benefit the most. The platform gives them tools to analyse complex data without needing specialised staff.
