Nama ERP's New UI and AI Integration Through MCP
Watch the full webinar on YouTube
This webinar presents a new generation of the Nama ERP user experience. It starts with the menu, list views, and data-entry screens; moves through chat, announcements, and dashboards; and concludes with connecting the system to AI tools through MCP. These improvements share one goal: a more flexible experience that helps users find and analyze data and complete tasks in fewer steps, while preserving access controls.
Webinar Chapters
- The new menu and its display modes — 00:18
- List view: columns, filters, and sorting — 02:09
- Pinning and resizing columns — 05:58
- User, interface, and theme settings — 10:03
- Mobile support — 16:16
- Chat and mentions of system records — 17:32
- General announcements and audience targeting — 20:29
- Data-entry grid improvements and shortcuts — 22:46
- New dashboards and interactive filtering — 27:13
- Creating a dashboard with AI — 33:28
- The generated payroll dashboard — 42:04
- Two ways to connect AI to Nama — 47:15
- MCP tools and secure queries — 51:03
- Why use MCP with AI agents? — 55:59
- Permissions and access controls — 59:02
- Internal AI and specialized use cases — 01:04:11
The New Menu and Its Display Modes
The menu can now be pinned so that it remains open while users move between screens. It also offers multiple ways to display modules and commands. Users can choose a progressive view that shows only the contents of the current level, or a tree view that displays nested levels together. Menu search covers all available commands and records, including items that may not appear in global search.
Icons and colors make system areas easier to distinguish visually. Users who prefer a quieter interface can disable the icon colors.
List View: Direct Control of Data
List views received a substantial set of improvements. Users can show or hide columns, filter directly from column headers, clear all filters, and sort records by clicking a column. To sort by multiple columns, hold Shift while selecting them; the grid displays each column's sorting priority.
Search criteria now support operators such as “one of” and “not one of,” allowing users to select several references—such as customers or branches—without creating a separate criterion for every value. Numeric fields can be filtered using equals, greater than, or less than, even when those fields are not included in the traditional search-criteria screen.
Pinning Columns, Fitting the Grid, and Copying Data
Important columns can be pinned to the right or left of the grid so that they remain visible during horizontal scrolling. This is especially useful for document numbers and codes. Users can move columns, resize them manually, or ask the system to fit them to the available screen width or the data they contain.
The grid's side panel provides another way to manage columns and filters. Copy behavior can also be configured to return the displayed value, a reference code, or a reference name. When users select several numeric cells, the system shows a quick summary—including the total, minimum, and maximum—without requiring an Excel export.
Settings for Each User's Working Style
The user-data screen provides access to contact details and password changes, along with several interface preferences:
- Normal, compact, or extra-compact spacing.
- Dark mode to reduce eye strain.
- Showing or hiding menu icon colors.
- Horizontal or vertical field layouts on edit screens.
- Two or four fields per row, depending on the available space.
- Hiding action-button labels to free additional screen space.
Themes allow administrators to define a shared appearance, including font sizes, spacing, and colors for light and dark modes. Individual users can also create personal themes that do not affect anyone else.
A Better Mobile Experience
Parts of the interface, including approvals, have been redesigned for smaller screens. The system and its grids are now more usable on mobile devices, while display-density options remain available so each user can select a layout appropriate for their screen.
Chat Inside Nama
Chat supports real-time communication between system users and is also available in the mobile app. It includes one-to-one conversations, groups, images, voice messages, and emoji.
More importantly, a message can mention a Nama record—such as an employee or an invoice—and turn that mention into a direct link to the record. The same infrastructure can be used in customized customer apps to connect customers or sales representatives with the internal team.
General Announcements
Nama distinguishes notifications tied to events for a particular user from general announcements intended for a wider audience, such as a change to a holiday date or an upcoming company event.
An announcement can target all users, a specific user, a user group, or a security profile. Administrators can define its publication period, deactivate it manually, and display it on the home page and in the mobile app.
Data-Entry Grids and Keyboard Shortcuts
The new grids continue to support keyboard shortcuts for opening referenced records in a new tab or pop-up, or for creating a new referenced record. Search results also provide more information so users can inspect a record before selecting it.
One of the most practical improvements is multi-cell editing: select multiple cells, edit one, and press Ctrl+Enter to apply the value across the selection. Users can also copy complete lines from one invoice to another or exchange data with Excel, saving significant time in invoices and production documents with many detail lines.
More Interactive Dashboards
Dashboards have changed substantially in both appearance and capability. A dashboard can now contain tabs and present KPIs, charts, and tables in multiple formats. The webinar demonstrates a group-of-companies dashboard with an overview of revenue, costs, and expenses, followed by analyses for sales, inventory, payroll, projects, and other activities.
Dashboard data responds immediately to filters. Selecting a company or branch refreshes the figures across all tabs. Dashboards also support cross-filtering: clicking a value or chart element filters the other components using the same selection. Users can combine several selections and later remove them from the filter bar.
Creating a Dashboard with AI
Watch the demonstration from 33:28
The presenter demonstrates the creation of a payroll dashboard using an AI coding agent. The agent first receives access to the Nama ERP Support Knowledge Base repository, which contains documentation for the business-intelligence features and their technical reference, as well as the data model describing system entities, tables, and relationships.
After reading these sources, the agent receives a natural-language request to create a JSON dashboard for the current month's payroll. It proposes filters for branch, department, and month; KPIs for net payroll, additions, deductions, and employee count; charts that distribute payroll by branch and department; and a detailed table.
The benefit goes beyond generating a visual design. The agent combines the user's business request with technical documentation and the data model to produce a file that Nama can import. If an error occurs, the user can return the system error or a screenshot to the agent so it can correct the file and try again.
WARNING
Sensitive or complex dashboards should be implemented with technical-support assistance. Always review queries and results before adoption, particularly when they involve financial or payroll data.
Connecting AI to Nama: Internal Chat or MCP?
The webinar describes two connection methods:
- AI chat inside Nama: Users can ask a model questions from within the system. This method requires an API key and file-store configuration, and its capabilities remain limited to the model and tools available inside the application.
- MCP (Model Context Protocol): An external agent, such as Codex or Claude Code, receives a defined set of Nama tools through credentials belonging to a particular user.
The presentation emphasizes MCP because an external agent can read files, run tools, and work with Excel sheets or invoice images before calling Nama tools to complete the requested workflow. The experience is therefore not limited to text-based questions and answers.
What Can an AI Agent Do Through MCP?
Administrators decide which tools the agent can use. Examples include:
- Importing and exporting data.
- Running an entity flow or an approved system action.
- Running a report and returning its result.
- Executing a secure read-only query and returning the data to the agent.
This allows users to ask an ad hoc analytical question—such as which items sold the most or which customers have a low profit margin—and let the agent query the data, summarize it, or visualize it. Information that managers need every day is still better placed on a permanent dashboard.
During the demonstration, the presenter asks the agent to import the payroll dashboard it created earlier. When the first attempt fails, he sends it a screenshot of the error; the agent revises the file and eventually creates the dashboard in Nama.
Permissions Are Part of the Design
MCP does not automatically give an agent unrestricted access. The agent operates with the credentials of a specific Nama user and cannot perform actions that user is not allowed to perform. For example, if the connection uses an HR user who cannot create dashboards, the agent cannot create one either.
Each tool can be restricted to particular users. A powerful tool may be reserved for the system administrator, while narrower tools are made available to specific job roles. The dashboard itself is also protected by view permissions, so a payroll dashboard can be hidden from anyone without payroll access.
DANGER
A technically successful connection is not necessarily appropriate for every user. Design user permissions, tool permissions, and dashboard permissions together, and use a limited account instead of a full administrator whenever possible.
External Agents and Internal AI
Watch the discussion from 01:04:11
The final discussion compares external coding agents with an internal retrieval-augmented generation (RAG) system in Nama. External agents are more capable because they can read files, write and run helper programs, and process their results. For intensive work, subscriptions to some agent tools may also cost less than direct API usage.
An internal solution remains useful for narrow, repeatable scenarios configured for a particular customer. One example is reading a batch of invoice images, extracting the supplier, invoice number, total, and item details, and preparing the records for an accountant to review before submission. The webinar confirms that these scenarios can be delivered through configuration tailored to the customer's workflow.
Conclusion
The webinar brings together two complementary tracks: improving everyday work in the user interface and opening the system to controlled, intelligent automation. The new UI reduces the effort required for searching, filtering, and editing; interactive dashboards improve analysis; and MCP lets AI agents use Nama data and perform defined tasks. The real value depends on three foundations: comprehensive documentation, precisely defined tools, and permissions designed before sensitive data or operations are exposed.