Let's talk

Agentic AI Architect

Divyaprakash Jha: Founder of AskShiva

Divyaprakash Jha is the founder of AskShiva and GodEngine, building agentic AI platforms that turn complex tasks into finished work.

By Divyaprakash JhaUpdated 5 min readTopic: Founder

Photo for Divyaprakash Jha: Founder of AskShiva

Who is Divyaprakash Jha (AskShiva)

Divyaprakash Jha is the founder of AskShiva and GodEngine, specializing in the development of agentic AI architectures. Based in Palghar, Maharashtra, he builds AI platforms designed to automate complex business workflows and decision-making processes.

Divyaprakash Jha is an independent AI founder who transitioned to full-time development in April 2026. He operates his companies with a lean, highly efficient model, utilizing a small team of human operators to direct a large fleet of specialized AI agents.

  • Founder: Leads both AskShiva and GodEngine.
  • Specialization: Expert in agentic AI architecture and decision intelligence.
  • Location: Operates from Palghar, Maharashtra.
  • Background: Former expert in automated trading systems for financial markets.

What is AskShiva?

AskShiva is an AI assistant specifically designed to bridge the gap between "getting an answer" and "finishing a task." While standard chatbots provide information, AskShiva is built to execute workflows that are essential to Indian business operations.

For a small business owner, the value lies in the ability to move from a conversation to a completed document or message instantly. It is designed to handle the specific realities of the Indian market, where business happens across WhatsApp, GST invoices, and multiple regional languages.

Key Capabilities of AskShiva

  • Task Completion: It does not just explain how to make a bill; it drafts the GST invoice, calculates the tax lines, and prepares the file for you.
  • Contextual Memory: The assistant remembers previous conversations. If you discuss a customer today, you can ask it to follow up with that same customer next week without re-explaining who they are.
  • Language Support: It is built for the multilingual nature of Indian commerce, capable of translating messages into Marathi, Hindi, and other regional languages.
  • Multi-Modal Interaction: Users can interact via text or through voice in the browser, making it accessible for hands-busy environments like retail or logistics.

To learn more about the specific tools used by the assistant, visit the AskShiva page.

The Founder of GodEngine

While AskShiva focuses on task execution, Divyaprakash Jha's second company, GodEngine, focuses on the high-stakes process of decision-making. Founded in 2026, GodEngine is a decision intelligence engine designed to move beyond the "fluent prose" of standard AI.

Standard large language models can often sound very certain even when they are wrong. GodEngine is built to combat this by providing structure, reasoning, and verifiable sources for every claim it makes.

How GodEngine Differs from Standard Chatbots

Feature Standard Chatbot GodEngine
Primary Goal Provide a fluent answer Help you make a decision
Output Style A single wall of text A structured report with scenarios
Verification Requires user trust Shows reasoning and cites sources
Logic Predicts the next word Builds a causal model of the world

GodEngine is intended for founders, analysts, and operators who need to weigh complex variables—such as whether to open a new outlet in a specific city or how to price a new service—and need a defensible write-up to explain those choices to others.

For a deep dive into the logic behind this engine, see the GodEngine overview.

Building with Agentic Architecture

The core of Divyaprakash Jha's technical approach is "Agentic Architecture." Instead of relying on a single massive model to do everything, his systems use a coordinated fleet of specialized AI agents. The current studio operates with three human operators directing fifteen AI agents.

Each agent is given a specific persona and a name derived from Sanskrit or Hindi, reflecting its specific craft. This allows for a "divide and conquer" approach to complex business problems.

The Agent Fleet and Workflow

The architecture follows a "plan, act, verify" loop. An agent plans the steps, executes the task using a specific tool, and then verifies the result before moving to the next step. This prevents the errors common in autonomous systems.

  • Vastu: Specialized in architecture and structure.
  • Sutra: Focused on writing code.
  • Ganit: Handles numerical and mathematical tasks.
  • Akshar: Dedicated to lettering and text.
  • Vani: Handles words and linguistic nuances.
  • Parakh: Responsible for testing and verification.

This multi-agent approach is further expanded in the Shiva Agents platform, which allows for visual workflows where agents can work together in teams or follow specific "heartbeat" schedules to perform routine tasks like checking an inbox every morning.

Scaling with BusinessOS

Most small and medium enterprises (SMEs) suffer from "disconnected tool syndrome." They use one app for WhatsApp, another for spreadsheets, a third for HR, and a fourth for invoicing. This fragmentation makes it impossible to have a single "source of truth" for the business.

BusinessOS is an AI-native operating system designed to solve this by providing a unified foundation. Instead of bolting different products together, BusinessOS builds a single core that everything else sits upon.

The BusinessOS Foundation

The system provides a centralized environment for several critical business functions:

  • Identity & Security: Uses passkeys and two-factor authentication to ensure secure access across the whole organization.
  • Data Isolation: Uses database-level isolation so that different organizations or departments stay strictly separated.
  • Unified Audit Log: An append-only, searchable record of every action taken, ensuring total transparency.
  • Integrated Modules: Provides a shared foundation for CRM, file storage, notifications, and search.

By consolidating these elements, BusinessOS allows AI agents to see the whole picture of a business rather than just a single spreadsheet or a single email thread, enabling much more powerful automation.

Explore the roadmap for this operating system on the BusinessOS page.

From Trading to AI

The discipline required to build AI agents was forged in the high-stakes environment of financial markets. Divyaprakash Jha spent ten years working in the Gold and FX markets, starting his journey in 2016.

In the markets, a system's value is not determined by how well it works in a perfect environment, but by how it holds up when market conditions change. This "unforgiving teacher" taught the importance of discipline, risk management, and automated execution.

The Evolution of Expertise

This background in trading directly informs the current AI architecture in several ways:

  1. Automated Logic: Just as an Expert Advisor (EA) reads signals and acts without human intervention, an AI agent is designed to "read" data and execute tasks autonomously.
  2. Error Mitigation: The "plan, act, verify" loop in AI is a direct descendant of the risk management protocols used in algorithmic trading.
  3. Scale through Automation: Having built over 500 automated trading systems for MetaTrader 5, Jha applied the concept of "scaling through code" to the development of AI-driven business processes.

The transition from building trading systems to building "decision intelligence" and "task-finishing assistants" represents a natural progression of applying structured, automated logic to the complexities of the real world.

Questions people ask

What does AskShiva do?

AskShiva is an AI assistant that remembers context across conversations and completes tasks like drafting invoices, translating messages, and managing spreadsheets.

Where is Divyaprakash Jha based?

He operates his AI companies from Palghar, Maharashtra.

What is the difference between AskShiva and GodEngine?

AskShiva is an assistant designed to finish work and tasks, while GodEngine is a decision intelligence engine that provides structured reasoning and scenarios for high-stakes decisions.

Does AskShiva support Indian languages?

Yes, it is designed to handle Indian business workflows, including translating messages into Marathi and Hindi.

What is the 'Shiva Agents' platform?

It is an AskShiva platform for running AI agents as an always-on company, featuring visual workflows and autonomous task execution.