Analytics Model

What is an Analytics Model?

Analytics Model is a generative AI platform designed for enterprise data analysis and business intelligence. After connecting to a database or business system, users can pose questions in natural language; the system then generates queries, charts, dashboards, and actionable insights.

Its goal is to enable business professionals who are not familiar with SQL and traditional BI tools to explore data, while providing analysts with faster visualization and automation processes. The results generated by AI still require verification regarding metric definitions, query logic, and data integrity.

Main functions

1. Natural language query data

Users can directly describe the questions they want to know answers to, such as sales trends, customer churn, changes in profits, or inventory abnormalities. The platform converts these questions into data queries and returns the results, eliminating the need for users to write complex SQL statements manually.

• Use free text to pose analysis questions.
• Automatically identifies related tables, fields, and relationships.
• Returns values, trends, and explanations.
• Continue asking questions to narrow down the timeframe, region, or range of products.

Natural language is prone to ambiguity; terms such as revenue, active users, and conversion rate can have various interpretations. Before preparing a formal report, it is necessary to ensure that the correct fields, filtering criteria, and time zone are being used.

2. Automatic generation of charts and dashboards

The platform can generate a complete interactive dashboard from a single description, including charts, KPIs, filters, and explanations. Users can also make manual adjustments by dragging and dropping fields, selecting aggregation methods, and changing chart types.

• Bar charts, line charts, and other common visualizations.
• Filter, sort, and aggregate data by dimension.
• Combine multiple metrics into a business dashboard.
• Generate personalized views based on roles or industries.

The chart that is selected automatically is not necessarily the one that best represents the data. Analysts should examine the axes, the range of values displayed, the sample size, and the meaning of the colors, in order to avoid drawing misleading visual conclusions.

3. Over 500 data connections

The official website states that the platform can connect to more than 500 different data sources, listing common systems such as BigQuery, MySQL, Redshift, Snowflake, and Azure. The connection layer supports both real-time and batch data processing.

The actual number of searchable entries displayed on the public connector directory page may differ from the total capacity of \"over 500\"; the specific drivers, versions, and authentication methods should be confirmed within the account or during the pre-sales phase.

4. Real-time data and scheduled alerts

Users can add frequently used queries to the Feed and set the execution time and trigger conditions. When the metrics meet the specified conditions, the system sends a notification to the designated recipients.

1. Select and connect to the data source.
2. Express the issues that need monitoring in natural language.
3. Add the query to the Feed.
4. Set it for daily, weekly, or other intervals.
5. Configure thresholds or filtering criteria.
6. Specify the notification recipients and test the results.

Automatic warning systems must address issues such as data delays, repeated notifications, and abnormal values. For critical operational alerts, alternative monitoring channels should be established, as it is not advisable to rely solely on a single AI platform.

5. Prediction, trend, and anomaly analysis

The platform can generate trend analyses, predictions, and data-driven recommendations, and it helps to identify anomalies or opportunities. The official website illustrates how the platform can be used through examples related to product performance, advertising returns, financial aspects, logistics, and customer metrics.

Predictive results depend on historical data, assumptions, and models; they do not determine the future with certainty. Business leaders should consider the confidence intervals, external variables, and conditions under which the model fails.

6. Pivot perspective and manual control

When AI is unable to understand complex structures, users can use the Pivot function to manually place fields on the horizontal and vertical axes, and select the fields to be aggregated as well as the type of visualization to be used. This allows switching between automated generation and traditional analysis methods.

Pivot tables can still produce incorrect aggregates due to duplicate records, null values, and many-to-many relationships. Data quality checks should be carried out first before interpreting the charts.

7. Analytics Model API

The official API is used to generate charts and complete dashboards based on inputs; it can automatically select relevant data and return analysis results. It also supports the retrieval of metadata, such as schema, table structure, and data relationships.

• Integrate natural language analysis into internal systems.
• Interactive dashboards are created automatically.
• Connect to real-time or batch data sources.
• Customize data selection and visualization based on business insights.
• Data connections to other AI platforms are provided through a unified MCP layer.

The authentication methods, rate limits, and separate pricing for the API are not fully displayed on the public page. Before integrating it into their systems, companies should request the technical documentation, a sandbox environment, and service guarantees.

8. Multi-agent and MCP connection layer

The product description for Analytics Model in 2026 highlights generative AI, multi-agent intelligence, and MCP connectivity capabilities. Its goal is to enable different models or agents to access enterprise data and perform analyses through a unified layer.

An MCP connection expands the range of data that can be accessed; it is necessary to assign read-only permissions to the proxy, implement field masking, and ensure tenant isolation. Any write operations or actions that trigger business processes must undergo additional approval.

Which users is it suitable for?

• Business managers who need to quickly view KPIs and trends.
• Business teams that analyze sales, marketing, and customer data.
• Data analysts who create visualizations and reports.
• Financial professionals who monitor financial risks and unusual operational activities.
• Developers who wish to integrate natural language analysis into their products.
• Enterprise customers who need to control data residency and identity permissions.

How to get started?

1. Register for a trial or choose a purchase option from AWS Marketplace.
2. First, connect to the database or business system that contains only test data.
3. Define the definitions for key metrics such as revenue, customers, orders, and time.
4. Pose a simple question in natural language that can be verified manually.
5. Check the tables, fields, filters, and aggregation methods selected by the system.
6. Generate charts or dashboards, and manually adjust the display settings.
7. Save to Feed when needed and set alert conditions.
8. Expand the scope of data and users after verification using real business samples.

Price and deployment version

As of August 26, 2026, AWS Marketplace offers two monthly subscription plans for Professional Data Visualization and On Premise AWS EC2 Deployment, along with a free trial period. The page detailing the 12-month contract states that savings of up to 17% can be achieved, but the publicly available information does not show the total cost for an entire year.

The costs associated with AWS infrastructure may be charged separately; the implementation, maintenance of localized solutions, as well as the resources required for the models must also be specified in the contract. The prices listed above are valid as of August 26, 2026, and the final amounts will be determined based on AWS’s purchase options, the settlement page, and the contract terms.

Refunds and contract considerations

The AWS Marketplace page indicates that vendors generally do not provide refunds for services for which requests have already been made; refunds are only possible in a few cases where no request was submitted at all. The terms related to trial periods, monthly renewals, and annual contracts should be checked before making a purchase.

• Confirm how 500 messages are defined and whether failed requests are counted.
• Ask whether service will be discontinued or additional fees will apply after exceeding the limit.
• Verify the costs related to AWS computing power, storage, traffic, and models.
• Confirm the discounts, early termination options, and renewal procedures for annual contracts.
• Keep evidence of the trial period, cancellation, and request for a refund.

Data security and privacy

The terms of service state that the customer owns the customer data, and the platform accesses or uses this data only when providing services, as required by law, or upon the customer’s instructions. The data may be processed and stored in the United States or in other countries where the service provider’s facilities are located.

Customer administrators can access and manage end-user data, as well as obtain the necessary approvals. Health information protected by regulations such as HIPAA cannot be uploaded without official written consent and the corresponding agreements.

Local deployment

The On Premise option is deployed within the customer’s own AWS EC2 environment; according to the vendor, data and processing activities do not get sent to any external infrastructure. The customer has control over computing resources, where the data is stored, as well as over identity management and role permissions.

\"Localization\" does not automatically imply complete offline functionality or the provision of source code. Companies need to assess the network requirements related to software updates, license verification, telemetry, model invocation, and technical support.

Is the Analytics Model open source?

The Analytics Model is available as a commercial SaaS product as well as an option for enterprise deployment; however, no complete source code, models, or version that can be hosted independently on one’s own servers have been made publicly available by the developers. AWS EC2 offers an environment for deployment, but this does not equate to an open-source license.

The platform offers API and MCP connection capabilities, but this does not mean it can be classified as open source. If there is a need for source code hosting, further development, or disaster recovery options, these matters should be specified separately in the enterprise contract.

Usage restrictions and precautions

• AI queries may select the wrong table, field, association, or aggregation method.
• Unified definitions must first be established for business metrics in natural language.
• Predictions, recommendations, and anomaly detection are not guaranteed to be accurate.
• Connecting to multiple data sources increases risks related to permissions and privacy.
• The professional version has a limit of 500 messages per month.
• The terms of service set limits on the scope of liability, and high-risk decisions require independent review.

Product advantages

• Lower the barriers to data querying and visualization through natural language.
• It supports over 500 data sources, as well as real-time and batch connections.
• It can automatically generate complete dashboards and scheduled alerts.
• It also offers a web platform, APIs, MCP, and AWS deployment options.
• Preserve manual analysis controls such as Pivot and filtering.

Product limitations

• The message quota for the professional version is clear, but the separate price for the API is not transparent.
• The total number of public connectors and the criteria used for displaying them in the catalog need to be confirmed.
• AI-generated results still need to be reviewed by someone familiar with the business data.
• Localization solutions are costly, and AWS costs may apply additionally.
• No official open-source version was found.

Frequently Asked Questions

Is knowledge of SQL required for the Analytics Model?

Not necessarily. Users can use natural language to formulate queries and generate visualizations, but an understanding of data structures and the definitions of various metrics still helps in evaluating the results.

Which data sources can be connected?

The official website states that it supports more than 500 different data sources, listing systems such as BigQuery, MySQL, Redshift, Snowflake, and Azure; the specific connectors available need to be checked within the account.

How much is the Professional plan?

The current public price on AWS Marketplace is $99 per month; it includes up to 500 messages, as well as advanced visualization features, multiple connectors, and filtering options.

Can it be deployed on my own AWS?

Yes. The On Premise AWS EC2 option costs $5,000 per month; according to the provider, there are no limits on the number of users or on the usage of the platform, and customers have control over their data and processing environment.

Can analysis generated by AI be used directly for decision-making?

It is not recommended to use it directly. The data quality, SQL logic, metrics definition, charts, and prediction assumptions should be verified, and important decisions require manual review.

Is the Analytics Model open source?

No. It is a commercial SaaS product as well as an enterprise deployment solution; APIs, MCPs, and customer-based AWS deployments do not constitute open-source software.