Talk to your business data in natural language

Is it hard to explore your data? Do your colleagues spend hours extracting and consolidating information from your business tools and applications?

What if you could query your systems directly in natural language?

What was last month’s margin? Why did expenses rise? Which customer generates the most revenue?

Your data is spread across several tools. Find the information you need in a single conversation, without searching through multiple applications. 

With the Datavizin AI Assistant, you can ask questions about your business data in natural language from a chat interface. We connect your data sources and work with you to define which information users can access and which metrics to use.

Datavizin AI Assistant
What was our margin last month?
DATAVIZIN AI ASSISTANT

Within the agreed scope, the gross margin for July was 28 %. Here is how it changed over the past four months.

Your sales, margins and expenses, based on your business definitions.

Sovereign Swiss cloud

Data infrastructure and strategy

Free one-hour audit

Training and support

Does this sound familiar?

Revenue goes up.
Margin falls.
What now?

startup, whiteboard, room

You are preparing for a management meeting. You want to understand which products, customers or expense items are driving the change.

The information exists. But you have to open the ERP system, find an export and ask the person who knows where to look and how to find the answers.

Your question should be answerable without submitting another reporting request.

Your questions are straightforward. Why are the answers so hard to find?

One figure is in the ERP system, while the explanation is in another tool. In between, you have to ask the one person who knows where to look.

Your Datavizin assistant brings your data sources together in a single conversation. Ask your question, review the result and continue the analysis without hunting for the right screen or session in each application.

Track margins.

“What was our margin last month?”

Use your own definition of margin every time.

Review expenses.

“Compare July 2026 expenses with June by category.”

Identify which categories contributed to the differences.

Understand your customer portfolio.

“Which customer has generated the most revenue since the
start of the year?”

Specify the period, then explore the answer in more detail.

The business question

What drives your profitability?

“What was our margin by product family in July?”

Compare product families using the margin definition agreed with your company. Then ask for the details behind a result.

For your decision: identify the areas to review before adjusting your priorities.

Gross margin · July

28 %

One common metric, broken down by product family.

Family A 34 %
Family B 28 %
Family C 22 %

Example with sample data. Answers depend on the configured data sources and analyses.

How does our AI assistant work?

The user asks a question

Ask our AI assistant a question in natural language, as you would with another AI model or a colleague. For example: “Which data sources are available now?” or “What were last month’s sales?”

Our assistant interprets the question

The AI assistant analyses the question and checks it against the sources it can access. If it has the information needed to answer, it queries the relevant source. 

The assistant retrieves the available data

Using the tools and sources it can access, the AI assistant retrieves the data needed to answer the user's question. 

The user receives an answer

The assistant returns the data, with a table or chart when possible and according to your preferences.

See the tool in action

In this video clip, a user asks our AI model about past sales.

You can ask directly without setting up complex queries, submitting a request to IT or installing additional applications.

Ask the Datavizin AI Assistant a question and get an instant answer.

Connected AI with clearly defined access

Our AI agent is hosted in Switzerland, as are the chat interface, gateway and connectors to your data sources. The results needed to answer a question are sent to the model for processing. They are excluded from model training and are never shared with third parties.

SQL Server, PostgreSQL, MySQL, MariaDB and REST APIs: connections depend on the access available and the capabilities of each system. We can combine several sources at your request when the data and linking keys allow it.

Access tailored to each user

The administrator selects which sources users can access.
Accounts, groups and roles govern access. Row-level filtering is outside the scope of the
standard version.

Business definitions agreed with you

We work with you to define the metrics and data objects users can access. Answers are based on that scope and depend on the quality of the source data.

Personal conversation history

Each user can access their own history. The gateway logs calls to configured objects.

Frequently asked questions

Do we need to replace our ERP system?
No. We connect accessible sources in your existing systems. During scoping, we verify technical access and any vendor restrictions.

 

Can we combine data from multiple applications?
Yes, as long as the available data can be linked. A single interface can query multiple databases and APIs. How information can be matched depends on the data model.

 

Can the AI change our data?
The standard version operates in read-only mode. Actions or write access require a separate project.

 

Is data sent to the model?
The results needed to formulate an answer are sent to the model. With Datavizin’s hosted service, processing takes place in Switzerland and the data is excluded from training. Users can still access their conversation history.

 

Does it replace BI?
The Datavizin AI Assistant gives users conversational access to configured data. It can supplement existing reports and support specific needs without an in-house BI team. Data preparation and metric definition are still necessary.

 

Can it explain every change?
It can compare and break down the figures it can access. An observed increase or correlation alone cannot establish causation. The explanation depends on what information your systems hold.

Request a demo of our AI agent

Complete the form. We will contact you within 48 business hours to schedule a demo of our AI data exploration agent.

Contact team BI Starter Pack industriel