G2G · Mantralaya Copilot
Mantralaya clears a 240-page inter-ministry file in one sitting.
The copilot summarises every annexure, surfaces precedent and flags policy conflicts — with full CAG-ready lineage.
G2C · Aarogya Agent
An ASHA worker in Gadchiroli gets a second opinion in 30 seconds.
Aarogya Agent screens symptoms against PMJAY protocols and books a teleconsult when the PHC is out of reach.
G2B · JNPT Flow
JNPT cuts container dwell time by a full shift.
Daisy plans berth, yard and gate moves together — and re-plans the moment a vessel slips schedule.
G2B · Udyog Suite
A Pune kirana files quarterly GST without an accountant.
GST Saathi reads invoices in WhatsApp, reconciles GSTR-1 and 3B, and pings the owner only when a human decision is needed.
Daisy Super AI
A multi-agent operating system for the state.
Daisy Super AI is the substrate underneath every MahaAI deployment — agentic, auditable, adaptive. Built on Daisy Nova, already in commercial production for Insure AI in North America.
Agentic Core
Autonomous AI agents that plan and execute multi-step workflows end-to-end — not chatbots, but task-completers.
Regulatory Audit
Built-in compliance, decision logging and reversibility. Every autonomous action is reviewable by a designated officer.
Adaptive Learning
Continuous improvement architecture. The system gets sharper with every transaction — without manual retraining cycles.
Architecture
Five layers. Everything above L0 is built and owned in Maharashtra.
Domain Agents
Pre-built agents for agriculture, MSME, industry, healthcare, education, public safety and citizen services. Configurable per department, swappable per ministry.
Orchestration Engine
Multi-agent planning, task delegation, decision arbitration and human-in-the-loop checkpoints. Every escalation routed to the correct officer.
Audit & Governance
Decision logs, explainability, bias monitors, regulatory rule packs and sovereign data-residency enforcement on every call.
Data & Model Layer
Vernacular language models — Marathi · Hindi · English · regional — with Indian-context fine-tuning and retrieval over departmental knowledge graphs.
Sovereign Compute
Maharashtra-resident GPU/TPU clusters with state-controlled key management. Onshore IP, onshore weights, onshore audit.
How it deploys
From in-principle approval to live agent in 90 days.
Scope a department workflow
Joint scoping with the nodal officer — pick a single high-value workflow (claims, grievances, scheme eligibility, compliance) for the first pilot.
Connect to the sovereign sandbox
Departmental data flows into the Maharashtra-resident sandbox under graded access. Daisy agents are fine-tuned in-environment.
Human-in-the-loop go-live
Agents shadow officers for 30 days, then move to assisted-autonomy with reversible recommendations and full audit logs.
Scale and export
A proven agent in one ministry becomes a re-usable module — packaged for other ministries, other states and South Asian governments.
For ministries & departments
Bring sovereign AI to your department.
Pilot a Daisy-powered agent inside your ministry in 90 days, with audit-grade decision logs and full Maharashtra data residency.