Datost Puts an AI Data Analyst in Slack So Your Team Stops Waiting for Dashboard Updates

The Macro: Every Company Has More Data Than It Knows How to Use

The data access problem in most companies follows a predictable pattern. The data team builds dashboards. Business stakeholders use the dashboards for a while. Then someone asks a question the dashboard does not answer. They submit a request to the data team. The data team adds it to their backlog. Days or weeks later, the answer arrives. By then, the business has either moved on or made a decision without the data.

This cycle repeats constantly. Data teams are perpetually bottlenecked. Business teams are perpetually frustrated. The data exists to answer most questions, but getting to it requires SQL skills, access to the right databases, and understanding of the data model. Most people in a company have none of those things.

BI tools like Looker, Tableau, and Metabase have tried to solve this with self-service analytics. They work well for pre-built dashboards and standard reports. But the moment someone needs an ad-hoc query or wants to explore data in a way the dashboard was not designed for, they are stuck. Self-service analytics is really only self-service for people who already understand data.

Datost, backed by Y Combinator, takes a different approach entirely. Instead of building another dashboard tool, they put an AI data analyst directly inside Slack.

The Micro: Ask Questions in Slack, Get Answers From Your Data

The product concept is straightforward. Datost connects to your databases, data lakes, docs, Slack history, and codebase. When someone has a data question, they ask it in Slack. Datost’s AI understands the question, translates it into the appropriate query, runs it against your data, and returns the answer in the Slack thread.

The key insight is meeting people where they already are. Nobody wants to learn another tool. Nobody wants to open a new tab, navigate a dashboard, and figure out which chart to look at. People want to ask a question and get an answer. Slack is where work conversations happen, so that is where data conversations should happen too.

What differentiates Datost from a generic AI chatbot connected to a database is that it has “its own computer.” It sees and understands your docs, Slack conversations, databases, data lakes, and codebase. This broader context means it can answer questions that require combining information from multiple sources, not just running a SQL query.

The founding team is lean. Maceo Cardinale Kwik is the 21-year-old CEO based in NYC. Jason Wang is the co-founder. Both are young builders moving fast.

The competitive space is active. Tools like ThoughtSpot offer natural language queries against data warehouses. Mode and Hex provide collaborative analytics notebooks. AI-native data tools like Equals and Supersimple are also targeting the gap between dashboards and ad-hoc analysis. But most of these are still separate applications. Datost’s Slack-native approach eliminates the friction of opening another tool.

The risk is accuracy. Data questions often have nuance. “What was our revenue last quarter” sounds simple, but the answer depends on how revenue is defined, which entities are included, and what time zone boundaries are used. If Datost gets these details wrong, users will lose trust quickly.

The Verdict

Data democratization has been a goal of the BI industry for two decades. Datost is taking a fresh approach by embedding the analyst directly in the communication tool everyone already uses.

At 30 days: how many data questions are being asked and answered through Datost per day? Usage frequency will show whether teams are adopting this as a habit or just trying it once.

At 60 days: what percentage of questions does Datost answer correctly without human data team involvement? The accuracy rate determines whether this reduces the data team’s burden or creates new cleanup work.

At 90 days: are data teams seeing measurable reductions in ad-hoc query requests? If the backlog is shrinking, Datost is genuinely democratizing data access.

I think the Slack-first approach is smart. The best tools are the ones you do not have to think about. If data answers appear in the same place where the question is asked, adoption will be natural. The challenge is building enough context and accuracy to earn trust with business users who do not have the data literacy to evaluate whether an answer is correct.