DataSays

Problem

AskAxon proved that people wanted to talk to their data. But it also exposed the next problem.

A plain-English answer is helpful. A decision often needs more.

Teams need the context behind the number, the chart that explains it, the dashboard that tracks it, and the slide that communicates it to someone else. In most companies, that work is scattered across analysts, spreadsheets, BI tools, presentation decks, and Slack threads.

The result is slow decision-making. Not because the data is missing, but because turning data into a useful business narrative still takes too much manual work.

Approach

I built DataSays as a standalone AI data workspace.

Unlike AskAxon, which lived inside Slack, DataSays has its own environment. It connects to data warehouses and business data sources, then lets users interact with their data in natural language.

The product is designed to move from question to output. A user can ask what is happening in the business, explore the underlying data, generate charts, create dashboards, extract insights, and turn the answer into a presentation-ready narrative.

The product direction is simple: make the AI useful beyond the first answer. DataSays is not just a chatbot for data. It is a workspace for analysis, explanation, and communication.

Outcome

DataSays is live with two customers: raven.inc and BPD Healthcare.

With Raven, I am also exploring a revenue-sharing model to distribute DataSays to their client base. That makes the product both a direct SaaS opportunity and a potential channel-led distribution business.

The early traction has helped sharpen the product thesis: the winning AI data product will not only answer questions. It will help teams turn answers into decisions, artifacts, and operating rhythms.

Stack

• OpenAI - reasoning, summarization, and insight generation
• Python / FastAPI - backend application layer
• React / Vite - frontend application
• LangChain - orchestration and agent workflows
• SQLGlot - SQL parsing and query generation
• BigQuery / data warehouse connectors - data access layer
• Pandas - data transformation and analysis
• Charting / dashboard layer - visual outputs and reporting