The dashboard is moving into the chat

When it comes to how teams actually consume analysis, something has changed quietly over the last couple of years. The starting point used to be a dashboard, a person would open a BI tool, navigate to the right view, and read the chart that answered their question. Increasingly, however, that question now starts in a chat window, someone types it into Slack or Discord, an agent or bot picks it up, runs the analysis, and replies in the same thread. The dashboard has not disappeared, however it is no longer the first place people look.

##Why the conversation became the interface

There are a few practical reasons this shift happened. First, the question is usually conversational to begin with (e.g., 'how did the campaign do last week?' or 'which accounts look at risk?'), and a chat window is the most natural place to ask it. Second, the cost of answering has fallen, because an agent with access to the right tools can run a query, summarize the result, and respond in seconds, without a human assembling a view in advance. Third, most questions are temporary, they need an answer once, and building or maintaining a dashboard for each of them would be disproportionate to the value. This means that the chat thread, rather than the BI tool, has become the place where a large share of everyday reporting now happens.

We should be careful not to overstate this. Dashboards remain the right tool for metrics that a team watches continuously (e.g., revenue, active users, or pipeline health), where a persistent, always-on view is genuinely useful. The shift we are describing is mostly about the long tail of ad hoc questions, which has always been poorly served by dashboards, because nobody wants to build a permanent surface for a question that will be asked once.

##The new problem: the answer is stuck in the thread

However, answering in chat introduces a problem of its own. A good answer that lives inside a Slack thread is difficult to share, difficult to return to, and difficult to read once it grows beyond a couple of sentences. When the agent replies with a wall of text, a few numbers, and maybe a pasted table, the result is technically correct but hard to consume, and it is almost impossible to forward to a stakeholder who was not in the thread. This is the same formatting gap that has always existed, it has simply moved from the analyst's afternoon to the agent's reply.

##A link, not a wall of text

In this case the useful pattern is straightforward: instead of replying with the full analysis inline, the agent creates a structured report and returns its URL. The thread stays clean, because the conversation keeps a short summary and a link, while the detail (e.g., charts, metrics, tables, and notes) lives on a page built for reading. This way the team keeps the speed of asking in chat, however it also gets an artifact that survives the thread, which can be revisited later or sent to someone who needs the result without the conversation around it.

In summary, the interface for reporting is moving from the dashboard to the conversation, and that is mostly a good thing, because it matches how people actually ask questions. However, the conversation is a poor place to store an answer, and this is the gap worth closing. The analysis can start in chat and the watching can stay on dashboards, but the answer itself most likely belongs in a report that has its own link.

##Further reading

• Google Cloud: A closer look at Looker Conversational Analytics
• Gartner: 75% of new analytics content will use generative AI by 2027
• IBM: What is conversational analytics?