DEV Community: tajdin k

Noviz AI Multi-Agent Architecture: How Independent AI Agents Work Together in Enterprise ERP

tajdin k — Thu, 12 Nov 2026 06:43:44 +0000

Noviz AI Multi-Agent Architecture: How Independent AI Agents Work Together in Enterprise ERP

At Noviz AI, we follow one architectural principle: The ERP decides, and the AI reasons.
Our multi-agent architecture starts with a central orchestrator that receives each request, decomposes the task, delegates it to specialised agents and consolidates their results. Dependent tasks follow sequential handoffs, while independent tasks execute in parallel.
Each agent has a dedicated responsibility, covering ERP operations, analytics, reporting, utilities and deep analysis. They operate as independent services using the open A2A 1.0 standard, allowing the architecture to expand without tightly coupling individual agents.
All agents share common execution layers for MCP-based tool loading, ERP permission guards, authenticated execution handlers, conversation memory, server-side result references and auditable turn management.
The LLM handles reasoning and determines the next step, without holding business data or execution state.
For complex analysis, execution plans require user approval before datasets are retrieved in parallel, calculations are performed in a Python sandbox and results are independently verified.
Scheduled agents extend the same capabilities into background, read-only operations. Findings return as proposals, with human approval and ERP acceptance required before any transaction is committed.
The ERP retains complete authority over permissions, validations and approval workflows.
We are implementing this architecture across ERPNext, Microsoft Dynamics 365 Business Central, Odoo and QuickBooks Online, with specialist, deep-analysis and scheduled agents currently undergoing testing.
The interesting part of multi-agent AI is not the number of agents, but how independently operating agents collaborate through a common, controlled execution architecture.
www.noviz.in

MultiAgentArchitecture #AgenticAI #EnterpriseAI #ERP #A2A #MCP #ERPNext #BusinessCentral #Odoo #QuickBooks

Noviz AI | Agentic AI for Enterprise Workflow Orchestration

tajdin k — Sat, 07 Nov 2026 09:25:07 +0000

Noviz AI by Zeetech Soft Solution provides a unified AI orchestration architecture across Microsoft Dynamics 365 Business Central, ERPNext, Odoo and QuickBooks.
The architecture separates AI reasoning from business-system execution. LLMs handle intent interpretation, tool selection and execution planning, while Noviz Relay manages orchestration and connected platform extensions execute operations within the respective ERP environments.
This enables centralized workflow customization, automation and orchestration without duplicating native business logic. Existing permissions, validations, approval workflows and transaction controls remain enforced by the underlying systems.
The objective is to make enterprise workflows interoperable, extensible and AI-driven while preserving each platform as the authoritative system of record.
Systems of record remain. Intelligence becomes a service.
https://noviz.in/

NovizAI #AgenticAI #ERP #MicrosoftDynamics365 #BusinessCentral #ERPNext #Odoo #QuickBooks #WorkflowOrchestration #EnterpriseAI

Enterprise AI Workflow Automation AI Agents in Noviz AI

tajdin k — Tue, 03 Nov 2026 16:52:58 +0000

Enterprise AI Workflow Automation & Orchestration Experts Available on Contract | AI Agents, Autonomous Workflows & Intelligent Automation

From AI Agent Development to Enterprise-Grade Workflow Orchestration and Progressive Autonomy.

At Noviz AI, developed by Zeetech Soft Solution, we are working on bringing intelligence into enterprise software through AI agents, intelligent workflow orchestration, business process automation, and controlled autonomy. Our approach is to evolve enterprise AI progressively — starting with standard AI agents, introducing customized workflows and Autoflow, implementing human approval mechanisms, and eventually enabling more autonomous business operations through contextual learning and feedback.

Enterprise AI is not just about connecting an LLM to an ERP or CRM. The real engineering challenge is making AI understand business processes, coordinate multiple operations, manage dependencies, handle approvals, and execute workflows reliably.

Our Enterprise AI Implementation Approach

1. Standard AI Agent Implementation

Connect LLMs with enterprise APIs, business data, and tools to enable natural language interactions and business operations.

2. Customized Workflow & Autoflow Automation

Transform individual AI operations into coordinated business workflows with parallel execution, sequential dependencies, conditional branching, and automated task orchestration.

3. Human-in-the-Loop Approval Architecture

Implement multi-level approvals, role-based authorization, workflow pause and resume, exception handling, and audit trails.

4. Enterprise Prompt & Business Context Customization

Customize AI instructions, business policies, prompts, tool selection, and contextual intelligence according to organizational requirements.

5. Progressive AI Autonomy & Learning

Use operational feedback, execution history, and workflow performance to improve automation, recommend workflow changes, and gradually introduce controlled autonomy.

Enterprise AI Workflow Orchestration Expertise

We work across:

Enterprise AI agent development.
AI workflow automation and orchestration.
Multi-agent systems and task coordination.
ERP and CRM AI integration.
Parallel and sequential execution.
Human-in-the-loop approval workflows.
LLM and prompt engineering.
Autonomous business process automation.
Workflow state management and recovery.
Enterprise AI architecture and customization.

Looking for AI Workflow Automation Experts on Contract?

Zeetech Soft Solution is open to contract-based engineering engagements with ERP companies, SaaS platforms, enterprise software vendors, AI startups, and technology consulting organizations.

Whether you need to build an AI agent, customize an existing workflow engine, implement complex business orchestration, or develop an autonomous enterprise AI architecture, we can collaborate on your engineering requirements.

Need Enterprise AI Workflow Automation or Orchestration Experts? Let's discuss your project.

Explore Noviz AI: https://noviz.in/

Zeetech Soft Solution

Building intelligent enterprise systems where AI doesn't just answer questions — it coordinates, executes, and improves business operations.

🚀 Noviz AI: Expanding into the Microsoft Dynamics Ecosystem

tajdin k — Fri, 30 Oct 2026 10:25:38 +0000

🚀 Another Milestone for Noviz AI: Expanding into the Microsoft Dynamics Ecosystem

Zeetech Soft Solution is now part of the Microsoft Dynamics ISV Partner Ecosystem, creating new opportunities to bring Noviz AI's enterprise agentic capabilities to the Microsoft Dynamics 365 ecosystem.

Our journey is driven by a broader ambition: to establish strategic partnerships across leading enterprise technology platforms and build a common intelligence and orchestration infrastructure that works across business applications.
We began with ERPNext, expanded into Odoo, and are now advancing into the Microsoft Dynamics ecosystem. QuickBooks is also part of our expansion roadmap.

Building Beyond Traditional Enterprise Integrations
We believe the future of enterprise AI goes far beyond connecting language models to business applications.Enterprises need intelligent systems capable of understanding business context, designing customised workflows, coordinating multi-step operations and executing tasks within secure, governed environments.Achieving this requires much more than an AI interface. It demands carefully engineered orchestration, standardised integration mechanisms, execution controls and a deep understanding of enterprise workflows.

This is the architectural foundation we are developing at Noviz AI.An Infrastructure That Evolves With Every Domain

Every enterprise platform introduces its own business logic, operational patterns and integration challenges. As we expand across technology ecosystems, our experience in workflow engineering and agentic orchestration continues to grow.
We are evolving our shared infrastructure to make complex enterprise workflows increasingly adaptable, reusable and easier to implement across different industries and business applications.Rather than developing isolated integrations for every platform, our objective is to establish a scalable foundation on which domain-specific intelligence and customised workflows can be built.

We recognise that enterprise agentic AI is a complex engineering discipline. Designing systems that can reason, orchestrate and execute reliably while maintaining enterprise governance requires experienced engineering and architectural thinking.
That is the challenge we are committed to solving.
Our journey continues. Our architecture evolves. Our vision grows.
Noviz AI | By Zeetech Soft Solution
🌐 https://noviz.in/

NovizAI #MicrosoftDynamics365 #MicrosoftISV #EnterpriseAI #AgenticAI #EnterpriseArchitecture #AIOrchestration #ERPNext #Odoo #DigitalTransformation

𝗕𝗲𝘆𝗼𝗻𝗱 𝗠𝗖𝗣 𝗖𝗼𝗻𝗻𝗲𝗰𝘁𝗶𝘃𝗶𝘁𝘆: 𝗕𝘂𝗶𝗹𝗱𝗶𝗻𝗴 𝗜𝗻𝘁𝗲𝗹𝗹𝗶𝗴𝗲𝗻𝘁 𝗘𝗻𝘁𝗲𝗿𝗽𝗿𝗶𝘀𝗲 𝗔𝗴𝗲𝗻𝘁𝘀

tajdin k — Thu, 29 Oct 2026 07:40:05 +0000

𝗕𝗲𝘆𝗼𝗻𝗱 𝗠𝗖𝗣 𝗖𝗼𝗻𝗻𝗲𝗰𝘁𝗶𝘃𝗶𝘁𝘆: 𝗕𝘂𝗶𝗹𝗱𝗶𝗻𝗴 𝗜𝗻𝘁𝗲𝗹𝗹𝗶𝗴𝗲𝗻𝘁 𝗘𝗻𝘁𝗲𝗿𝗽𝗿𝗶𝘀𝗲 𝗔𝗴𝗲𝗻𝘁𝘀
The Model Context Protocol (MCP) is an important step towards a more interoperable AI ecosystem. By standardising how AI applications connect with tools and external systems, MCP creates new opportunities for enterprise automation.
However, 𝗰𝗼𝗻𝗻𝗲𝗰𝘁𝗶𝘃𝗶𝘁𝘆 𝗮𝗹𝗼𝗻𝗲 𝗱𝗼𝗲𝘀 𝗻𝗼𝘁 𝗰𝗿𝗲𝗮𝘁𝗲 𝗶𝗻𝘁𝗲𝗹𝗹𝗶𝗴𝗲𝗻𝘁 𝗲𝗻𝘁𝗲𝗿𝗽𝗿𝗶𝘀𝗲 𝗮𝗴𝗲𝗻𝘁𝘀.
Enterprise agents must understand business context, interpret domain-specific rules, coordinate operations across systems, handle exceptions, and respect organisational policies. These capabilities require domain intelligence, workflow design, and reliable orchestration beyond tool connectivity.
𝗙𝗿𝗼𝗺 𝗧𝗼𝗼𝗹 𝗜𝗻𝘁𝗲𝗴𝗿𝗮𝘁𝗶𝗼𝗻 𝘁𝗼 𝗕𝘂𝘀𝗶𝗻𝗲𝘀𝘀 𝗜𝗻𝘁𝗲𝗹𝗹𝗶𝗴𝗲𝗻𝗰𝗲
Instead of repeatedly building similar integrations, we should encourage an ecosystem where platform providers expose reusable capabilities and developers focus on specialised agents, domain expertise, and composable business workflows.
The opportunity is to move beyond making tools available to making them work together intelligently to achieve measurable business outcomes.
𝗪𝗵𝗲𝗿𝗲 𝗡𝗼𝘃𝗶𝘇 𝗔𝗜 𝗙𝗶𝘁𝘀 (www.noviz.in.)
At Noviz AI, our guiding principle is: 𝗦𝘆𝘀𝘁𝗲𝗺𝘀 𝗼𝗳 𝗿𝗲𝗰𝗼𝗿𝗱 𝗿𝗲𝗺𝗮𝗶𝗻. 𝗜𝗻𝘁𝗲𝗹𝗹𝗶𝗴𝗲𝗻𝗰𝗲 𝗯𝗲𝗰𝗼𝗺𝗲𝘀 𝗮 𝘀𝗲𝗿𝘃𝗶𝗰𝗲.
We focus on connecting AI reasoning with enterprise context and operational workflows, enabling intelligent agents to work with existing business systems while respecting permissions and governance.
Imagine developers building specialised agents across platforms such as ERPNext, Odoo, QuickBooks, and Microsoft Dynamics 365. By reusing available tools and MCP servers, developers can focus on domain-specific intelligence and workflow design rather than repeatedly rebuilding integrations.
MCP can provide the connectivity foundation. Domain intelligence and workflow orchestration can turn that foundation into practical enterprise value.
𝗧𝗵𝗲 𝗳𝘂𝘁𝘂𝗿𝗲 𝗼𝗳 𝗲𝗻𝘁𝗲𝗿𝗽𝗿𝗶𝘀𝗲 𝗔𝗜 𝗶𝘀 𝗻𝗼𝘁 𝗷𝘂𝘀𝘁 𝗮𝗯𝗼𝘂𝘁 𝗰𝗼𝗻𝗻𝗲𝗰𝘁𝗶𝗻𝗴 𝗺𝗼𝗿𝗲 𝘁𝗼𝗼𝗹𝘀. 𝗜𝘁 𝗶𝘀 𝗮𝗯𝗼𝘂𝘁 𝗼𝗿𝗰𝗵𝗲𝘀𝘁𝗿𝗮𝘁𝗶𝗻𝗴 𝘁𝗵𝗲𝗺 𝗶𝗻𝘁𝗲𝗹𝗹𝗶𝗴𝗲𝗻𝘁𝗹𝘆 𝘁𝗼 𝘀𝗼𝗹𝘃𝗲 𝗿𝗲𝗮𝗹 𝗯𝘂𝘀𝗶𝗻𝗲𝘀𝘀 𝗽𝗿𝗼𝗯𝗹𝗲𝗺𝘀.
Learn more about our vision for enterprise AI at www.noviz.in.

MCP #EnterpriseAI #AgenticAI #AIArchitecture #ERPNext #Odoo #QuickBooks #MicrosoftDynamics365 #NovizAI

AI Integration and Data Privacy: In Enterprise AI, Trust Is the Real Question

tajdin k — Thu, 15 Oct 2026 07:52:18 +0000

Recently, we have received many questions about Noviz AI’s data privacy and security. These questions deserve an honest discussion, because the concern goes far beyond any single AI product.

Look at the enterprise software industry today. SAP, Microsoft Dynamics 365 and Oracle NetSuite have established themselves as trusted platforms for managing critical business operations. Their customers entrust them with financial records, customer databases, employee information, commercial strategies and years of accumulated business intelligence.

Now, these same enterprise ecosystems are increasingly partnering with major AI providers and embedding LLMs into their products.

But who ultimately controls the intelligence built around your business data, and how much of that trust can a customer independently verify?

Your business data is more valuable than you think
The AI industry operates in a highly competitive market. Companies invest billions in infrastructure, research, model training and product development. Data, domain knowledge and access to useful information all have commercial significance.

We have already seen serious public controversies over how AI companies collect and use information.

In 2026, Canada’s federal and provincial privacy regulators found that OpenAI had collected and used personal information from ChatGPT interactions to train models without obtaining the express consent they considered necessary. The findings concern personal information and consumer interactions, not proof that enterprise ERP records were secretly used for training. Nevertheless, they demonstrate why questions about consent and data use cannot be dismissed as hypothetical.

This brings us to a much bigger issue.

When businesses entrust their most valuable information to global enterprise giants such as SAP, Oracle and Microsoft, they place enormous trust in these companies not only to secure their data, but also to govern how it is accessed and used as AI becomes deeply integrated into their cloud ecosystems.

These companies are increasingly forming major commercial partnerships with LLM providers, bringing together enterprise infrastructure, sensitive business information and powerful AI technologies.

In an industry where enterprise data carries immense commercial value and AI capabilities are a competitive advantage, how much trust should businesses place in these ecosystems, and how can they independently verify that their data will never be used beyond its intended purpose?

The question is not whether these companies offer privacy commitments. It is whether businesses should rely on those commitments alone when the same ecosystem controls their infrastructure, AI integrations and access to commercially valuable information.

The uncomfortable reality is that a privacy promise and an architecture that minimises data exposure are not the same thing.

Intelligence without exploiting customer records
We believe AI should become more useful through collaboration, experience, agreed business rules and reviewed operational patterns not through the accumulation of confidential customer records.

Learn through collaboration and experience, never from your records.

This is not a claim that every major AI provider misuses customer data. It is a challenge to an industry in which businesses are increasingly expected to trust complex AI ecosystems with some of their most commercially valuable information.

We should be asking harder questions about who controls that information, what the AI actually needs, and whether the architecture itself limits unnecessary exposure.

At Noviz, that is the principle behind our approach to enterprise AI.

Trust should not be just something a vendor asks you to believe. It should be something the architecture is designed to earn.

Read more about our approach to data privacy: Data Privacy | Noviz AI | Zeetech Soft Solutions

How to build Build an enterprise AI agent

tajdin k — Thu, 08 Oct 2026 18:45:25 +0000

A Thought on the Next Generation of Enterprise AI

tajdin k — Sun, 04 Oct 2026 09:32:29 +0000

As enterprise platforms continue to evolve, an interesting question is emerging: if an ERP already has a mature framework for documents, workflows, permissions, business rules and security, should an AI agent create a separate system around it?

Perhaps the more sustainable approach is for AI to work within the systems businesses already trust.

This is particularly relevant as we extend our work from ERPNext to Microsoft Dynamics 365 Business Central. While these platforms have different architectures and ecosystems, the principle remains the same: the existing ERP should continue to govern its data, processes, permissions and business logic.

The AI layer should complement that foundation rather than compete with it.

This also opens an important possibility for organisations with customised ERP environments. AI integration should be capable of adapting to the existing application and its business context, rather than requiring organisations to redesign the way they operate.

With this approach, AI becomes an additional intelligence layer across established enterprise systems, while the ERP remains the source of truth and the organisation remains in control.

This is an important direction for enterprise AI, and one we are continuing to explore across different platforms and business ecosystems. You can find Noviz AI in our marketplace https://cloud.frappe.io/marketplace/apps/noviz_ai
www.noviz.in

Intelligence that evolves with every business

tajdin k — Sun, 27 Sep 2026 09:55:24 +0000

Intelligence that evolves with every business
Every organisation runs on business software. Very soon, every one of them will also run on AI agents. I founded Noviz because I believe the way those agents are built will decide which businesses truly benefit from this shift, and which are left maintaining yet another system of their own.

Look at the history of business technology and one pattern repeats. Companies once wrote their own applications, until packaged software arrived. They ran their own data centres, until the cloud arrived. They installed and maintained their own software, until Software as a Service arrived. Each time, what was built alone became shared, and the shared model prevailed, because it improved for everyone. Intelligence is the next layer to make that transition.

Today, AI agents are where business software was thirty years ago. Every enterprise builds its own, against its own systems, with its own team. These agents are isolated and static; they break with every upgrade, and the same engineering is paid for again and again. An agent built in isolation is at its best on the day it launches. From then on, the world moves and the agent does not. The measure of an agent is not what it can do on day one. It is how fast it improves.

That is why Noviz is built on a simple thesis: decouple the intelligence from the system of record. Platforms such as SAP, Microsoft Dynamics, Oracle and ERPNext should continue to do what they do best, holding the data and the transactions. Noviz supplies the intelligence that acts upon them, as a service: a business-domain AI agent that works within existing workflows under each user’s own permissions, evolves with the domain it serves, and requires no in-house AI team to maintain.

Because the agent is shared, every deployment strengthens it for every customer. Reviewed operating patterns become new capabilities for all, while one principle remains non-negotiable: patterns are shared, customer data never is. In business software, trust is not a feature; it is the price of entry. So governance is built into the architecture, not left to a prompt. Every action executes under the signed-in user’s own session, rules are validated in code before anything is committed, and every action is on the record.

This is not a concept. Our first deployment is in production today, inside ERPNext, and available on the Frappe Cloud Marketplace. It answers questions, generates reports and executes transactions on live business data, always under each user’s own permissions. You can watch it at work.

Where this leads is clear to me. Today, the agent implements what it is asked. Next, it will be authorised to handle routine operations within the limits a business defines. Over time, industry-trained agents will run entire processes, while people retain ownership of strategy and exceptions. Autonomy is earned, never switched on, and because the agent is shared, the trust earned in one industry accelerates the next.

We do not compete with the platforms our customers rely on; we make them more valuable. No business should have to replace the systems it trusts in order to become intelligent.

Systems of record remain. Intelligence becomes a service. That is what we are building at Noviz.

Taj Din

Founder & CEO, Noviz AI

Zeetech Soft Solutions

Noviz AI Agent for ERPNext — A Reasoning Layer for ERPNext

tajdin k — Wed, 02 Sep 2026 11:46:58 +0000

# Noviz AI Agent for ERPNext — A Reasoning Layer for ERPNext

Noviz AI is an AI Agent built specifically for ERPNext.

Instead of simply answering questions, Noviz is designed to understand user intent, reason about the ERPNext environment, and work with existing ERPNext capabilities to accomplish tasks.

How It Works

User → Noviz AI → ERPNext

The user provides the intent, Noviz handles the reasoning, and ERPNext remains the underlying source of business data, permissions, DocTypes, workflows, reports, and actions.

What Developers Can Expect

Work with ERPNext DocTypes and records
Analyze ERPNext data using natural-language requests
Generate reports, tables, and charts
Execute actions on ERPNext records
Work with workflows and business processes
Handle notification-driven processes
Coordinate multi-step ERPNext operations
Work with relevant ERP and email context

Built Around ERPNext

Noviz does not attempt to replace ERPNext's existing architecture.

It is designed to work with capabilities ERPNext already provides, including:

DocTypes
Metadata
APIs
User permissions
Reports
Workflows
Notifications
Business data

Permission-Aware

Noviz operates within the permissions of the currently logged-in ERPNext user.

The data and actions available through the agent are governed by the user's existing ERPNext access.

Your ERPNext login. Your access. Your permissions.

Data Efficiency

Noviz is designed so that the complete ERP database does not need to be sent to a language model for every request.

Processing can take place within the ERPNext environment, while AI reasoning can use metadata and only the relevant context required for the task.

From AI Assistant to AI Agent

The goal is not simply to build another chatbot for ERPNext.

Noviz is being developed around a different approach:

Understand the intent → Reason about the ERP → Use ERPNext capabilities → Get the task done.

This makes natural-language interaction a way to work with ERPNext, rather than simply ask questions about it.

Now Available

Noviz AI is now available on the Frappe Marketplace:

https://cloud.frappe.io/marketplace

We would love to hear from Frappe developers, ERPNext implementers, consultants, and businesses about real-world use cases, feedback, and ideas for where an AI Agent can be useful inside ERPNext.

Noviz AI — Your AI Agent for ERPNext.

Building an AI Layer for ERPNext: Using Frappe as the Execution Engine

tajdin k — Tue, 25 Aug 2026 07:47:14 +0000

I have been working on Noviz AI, an AI agent integration for ERPNext, and one architectural principle has become particularly important during development:

The LLM should reason about the user’s request, but it should not become the database, reporting engine, or visualization engine.

ERPNext and Frappe already provide most of the infrastructure required to execute business operations reliably. The AI layer can sit above that infrastructure and use it through a controlled tool interface.

Architecture
The integration is intentionally divided into separate responsibilities.

The LLM / AI layer understands the user’s intent, reasons about the required operations, selects the appropriate tools, determines filters, grouping and metrics, and interprets the returned results.

The Frappe execution layer performs the actual DocType operations, queries, aggregations, pagination and analytics while operating within Frappe’s existing permission model.

ERPNext remains the source of truth for the underlying business data, DocTypes, relationships, permissions, reports and document-generation capabilities.

This separation is important because the LLM is not being turned into an ERP database engine.

simplechart
simplechart
871×784 35 KB
Generating reports without sending the entire dataset to the LLM
One of the areas we have focused on is analytics and reporting.

Suppose a user requests a customer-wise sales analysis.

The reasoning layer can translate the request into an analytical operation containing the relevant entity, filters, grouping field, metric, aggregation and pagination requirements.

The execution layer performs that operation directly against ERPNext.

The complete result does not necessarily need to become part of the LLM’s context. The model can receive a compact representation of the analytical result, while the complete dataset remains available to the application for presentation.

This creates a useful separation:

LLM context → information required for reasoning

UI / report layer → complete information required for presentation

That becomes increasingly important when dealing with hundreds or thousands of records.

Tables, charts and reports as reusable capabilities
Tables, charts and reports do not have to become separate hard-coded AI features.

An analytics operation can produce structured data, and that same result can be consumed by different presentation capabilities.

A table can display the complete dataset. A chart can visualize the aggregated values. A report can use the same underlying query definition, and the report can subsequently be rendered as a PDF.

The LLM’s role is primarily to determine what needs to be calculated and how the result should be presented.

The deterministic execution and presentation layers perform the actual work.

This makes the system much more composable. New reporting or visualization capabilities can be added without requiring the reasoning layer to understand every implementation detail.

Reducing unnecessary token usage
This architecture also has an important effect on LLM costs.

A naive ERP AI implementation could retrieve a large number of records and repeatedly place those records into the model’s context during multi-step reasoning.

That is expensive and often unnecessary.

Instead, Noviz separates:

data retrieval → analytical processing → presentation → reasoning

For example, an aggregated result may contain hundreds of groups, but the reasoning model may only need summary information and pagination metadata to decide what to do next.

The complete dataset can remain outside the model context and be rendered directly to the user.

The LLM can therefore concentrate on the part where it provides the most value: reasoning and decision-making.

Working with Frappe’s existing permission model
Another important aspect is security.

The public Noviz Frappe application does not maintain a separate ERP credential for performing operations on behalf of the user.

The execution request is handled by the Frappe application under the authenticated user’s session. The existing ERPNext permission system therefore remains an important enforcement boundary.

The AI can request an operation, while Frappe determines whether that operation is permitted for the current user.

This keeps authorization close to the system that owns the data rather than attempting to recreate ERPNext’s permission model inside the AI layer.

                 User Request
                      │
                      ▼
            ┌──────────────────┐
            │   LLM / AI Layer │
            │                  │
            │ Intent + Reasoning
            │ Tool Selection   │
            │ Planning         │
            └────────┬─────────┘
                     │
               Tool / Call Spec
                     │
                     ▼
          ┌──────────────────────┐
          │ Frappe Execution     │
          │                      │
          │ DocTypes / ORM       │
          │ Filters / Queries    │
          │ Permissions          │
          │ Analytics            │
          └──────────┬───────────┘
                     │
                 ERPNext Data
                     │
          ┌──────────┼──────────┐
          ▼          ▼          ▼
        Table      Chart      Report
                                 │
                                 ▼
                                PDF

Why this architecture fits Frappe particularly well
Frappe already provides the fundamental building blocks required for this type of AI integration: DocTypes, ORM, permissions, reports, print formats, APIs and the Desk environment.

The AI layer therefore does not need to rebuild an ERP abstraction from scratch.

Instead, it can act as an intelligent orchestration layer over the existing Frappe ecosystem.

A generic execution protocol is particularly useful here. Rather than exposing every internal implementation detail to the model, the AI platform can translate its reasoning into controlled operations such as retrieving documents, retrieving lists, creating or updating documents, and executing analytical operations.

The Frappe application then performs those operations using its native mechanisms.

A useful direction for AI + Frappe
This architecture opens an interesting direction for ERPNext.

Instead of building an AI feature for every individual report, dashboard or analytical question, the AI can be given a well-designed set of composable capabilities.

The intelligence determines the required operation.

Frappe executes it.

The analytics layer processes the data.

The UI renders the result.

The goal is not to replace ERPNext’s reporting or business logic. It is to make those existing capabilities accessible through a reasoning interface.

This makes it possible to build increasingly sophisticated natural-language workflows while keeping the underlying execution deterministic, permission-aware and efficient in terms of LLM context.

For AI integrations with Frappe, the separation between reasoning and execution provides a practical foundation for natural-language reporting, analytics, charts, tables, PDFs and broader ERP workflows

𝗔𝗜 𝗔𝗴𝗲𝗻𝘁𝘀 𝗶𝗻 𝘁𝗵𝗲 𝗘𝗻𝘁𝗲𝗿𝗽𝗿𝗶𝘀𝗲: 𝗕𝘂𝗶𝗹𝗱𝗶𝗻𝗴 𝗦𝗺𝗮𝗿𝘁𝗲𝗿 𝗘𝗥𝗣 𝗗𝗮𝘀𝗵𝗯𝗼𝗮𝗿𝗱𝘀 𝘄𝗶𝘁𝗵 𝗣𝘂𝗿𝗽𝗼𝘀𝗲 𝗕𝘂𝗶𝗹𝘁 𝗧𝗼𝗼𝗹𝘀

tajdin k — Sun, 16 Aug 2026 06:28:53 +0000

🚀

One of the new capabilities we’ve been working on in our ERP AI agent is 𝗔𝗜 𝗴𝗲𝗻𝗲𝗿𝗮𝘁𝗲𝗱 𝗱𝗮𝘀𝗵𝗯𝗼𝗮𝗿𝗱 𝘃𝗶𝗲𝘄𝘀, allowing users to describe the view they need and letting the agent construct it from the underlying ERP data.
The key challenge wasn’t simply getting an LLM to generate a dashboard.
It was making the generation 𝗮𝗰𝗰𝘂𝗿𝗮𝘁𝗲, 𝗰𝗼𝗻𝘀𝗶𝘀𝘁𝗲𝗻𝘁 𝗮𝗻𝗱 𝗲𝗳𝗳𝗶𝗰𝗶𝗲𝗻𝘁.
We addressed this by defining 𝗽𝘂𝗿𝗽𝗼𝘀𝗲 𝗯𝘂𝗶𝗹𝘁, 𝘁𝘆𝗽𝗲𝗱 𝘁𝗼𝗼𝗹 𝗰𝗮𝗹𝗹𝘀 around the actual operations required to build a dashboard.
The LLM orchestrates the workflow, while the tools handle the deterministic work.
🔹 𝗙𝗲𝘁𝗰𝗵
Retrieves and aggregates the required ERP data through operations such as 𝗴𝗿𝗼𝘂𝗽𝗕𝘆, counts, sums, correlations and other supported aggregations.
🔹 𝗦𝗵𝗮𝗽𝗲
Converts those results into a validated chart and view specification. Server side calculations, category limits and structural constraints are handled outside the model.
🔹 𝗥𝗲𝗻𝗱𝗲𝗿
Converts the validated specification into the actual dashboard components and desktop view.
This gives the LLM a more precise responsibility.
Rather than asking the model to retrieve data, calculate metrics, design charts and implement the UI in a single generation, it determines 𝘄𝗵𝗶𝗰𝗵 𝗼𝗽𝗲𝗿𝗮𝘁𝗶𝗼𝗻𝘀 𝗮𝗿𝗲 𝗿𝗲𝗾𝘂𝗶𝗿𝗲𝗱 and orchestrates the corresponding tool calls.
For example:
𝗖𝗿𝗲𝗮𝘁𝗲 𝗮 𝘀𝗮𝗹𝗲𝘀 𝗱𝗮𝘀𝗵𝗯𝗼𝗮𝗿𝗱 𝘀𝗵𝗼𝘄𝗶𝗻𝗴 𝗼𝗿𝗱𝗲𝗿 𝘀𝘁𝗮𝘁𝘂𝘀, 𝘁𝗼𝗽 𝗽𝗿𝗼𝗱𝘂𝗰𝘁𝘀 𝗮𝗻𝗱 𝗺𝗼𝗻𝘁𝗵𝗹𝘆 𝘀𝗮𝗹𝗲𝘀 𝘁𝗿𝗲𝗻𝗱𝘀.
The agent can fetch the required datasets, shape each visualization and compose the resulting components into a single desktop dashboard.
We tested the approach against real dashboard generation, focusing on 𝗮𝗰𝗰𝘂𝗿𝗮𝗰𝘆, 𝗲𝘅𝗲𝗰𝘂𝘁𝗶𝗼𝗻 𝗲𝗳𝗳𝗶𝗰𝗶𝗲𝗻𝗰𝘆 and 𝗰𝗼𝗻𝘀𝗶𝘀𝘁𝗲𝗻𝗰𝘆.
One important finding was that reducing the implementation burden on the LLM can improve the result.
The model focuses on 𝗶𝗻𝘁𝗲𝗻𝘁 𝗮𝗻𝗱 𝗼𝗿𝗰𝗵𝗲𝘀𝘁𝗿𝗮𝘁𝗶𝗼𝗻, while deterministic operations are handled by tools that can be validated and tested independently.
Another important aspect is 𝗧𝗼𝗼𝗹 𝗖𝗼𝘃𝗲𝗿𝗮𝗴𝗲.
In several cases, the model correctly identified the required operation, but the corresponding ERP entity or capability was not exposed through the tool layer.
This is the direction we’re taking with the new dashboard capabilities:
𝗡𝗮𝘁𝘂𝗿𝗮𝗹 𝗹𝗮𝗻𝗴𝘂𝗮𝗴𝗲 𝗶𝗻𝘁𝗲𝗻𝘁 → 𝗽𝗿𝗲𝗰𝗶𝘀𝗲 𝘁𝗼𝗼𝗹 𝗰𝗮𝗹𝗹𝘀 → 𝘃𝗲𝗿𝗶𝗳𝗶𝗲𝗱 𝗱𝗮𝘁𝗮 𝗼𝗽𝗲𝗿𝗮𝘁𝗶𝗼𝗻𝘀 → 𝗰𝗼𝗺𝗽𝗼𝘀𝗲𝗱 𝗲𝗻𝘁𝗲𝗿𝗽𝗿𝗶𝘀𝗲 𝗱𝗮𝘀𝗵𝗯𝗼𝗮𝗿𝗱.
𝗜𝘁 𝗶𝘀 𝗮𝗯𝗼𝘂𝘁 𝗴𝗶𝘃𝗶𝗻𝗴 𝘁𝗵𝗲 𝗟𝗟𝗠 𝘁𝗵𝗲 𝗿𝗶𝗴𝗵𝘁 𝗮𝗯𝘀𝘁𝗿𝗮𝗰𝘁𝗶𝗼𝗻𝘀 𝘁𝗼 𝗱𝗼 𝗶𝘁𝘀 𝗽𝗮𝗿𝘁 𝗽𝗿𝗲𝗰𝗶𝘀𝗲𝗹𝘆.

AI #AIAgents #LLM #EnterpriseAI #ERP #SoftwareEngineering #DataViz #OpenSource