Dodge AI Raises $2.65M for ERP Maintenance Agents

Dodge AI has raised a $2.65M Seed to build autonomous agents for enterprise application maintenance. Accel and Google's AI Futures Fund backed the financing, with Schema Ventures, New Build Ventures, Antler, and angels from the SAP ecosystem also participating. The round matters because enterprise software rarely operates according to the clean process diagram sold during implementation. SAP, Salesforce, Microsoft Dynamics, Kinaxis, and Oracle JDE environments collect custom code, approval paths, integrations, exceptions, and undocumented workarounds.

Dodge AI is trying to turn that accumulated operating context into a control layer that agents can use to diagnose incidents, deliver changes, and remember why a system behaves the way it does. Its wager is that the reasoning behind yesterday's repair can become infrastructure for tomorrow's incident rather than another rediscovery exercise.

What Happened

Tech Funding News reported the $2.65M Seed on September 29, 2026. Dodge AI was founded in 2025 by Rebhav Bharadwaj, co-founder and CEO, and Aditya Thakur, co-founder and CTO. The founders met at BITS Pilani before building the company across San Francisco and Bengaluru.

The financing follows Dodge AI's selection for the 2026 Accel Atoms and Google AI Futures Fund cohort. Accel's program announcement described a structure combining equity investment, Google compute credits, technical access, and startup-building support. Dodge AI has not disclosed a valuation, ownership terms, or a detailed allocation of the new proceeds.

Why Enterprise Maintenance Is the Entry Point

Enterprise application maintenance is an institutional-memory problem disguised as a ticket queue. A failed job, broken integration, or incorrect output may be caused by a local rule that was added years earlier and documented only in an old change request or a consultant's recollection. The technical fix can be small while the work required to reconstruct the surrounding business logic is expensive.

Dodge AI calls its product an AI control plane for enterprise application maintenance. The platform connects to business processes, customizations, IT service-management systems, and legacy configurations, then builds context around incidents and change requests. Its agents are intended to identify root causes, recommend or execute fixes, and preserve the reasoning behind those changes so the next problem does not begin from zero.

That starting point separates the company from a generic ERP copilot. A copilot can summarize a ticket or explain a screen. Maintenance software has to understand permissions, dependencies, exception logic, and the consequence of acting inside a production system that moves inventory, money, orders, and customer commitments.

What Dodge AI Says the Product Is Doing

Dodge AI says it works with more than a dozen enterprises, half of them publicly listed, and that its platform handles hundreds of queries per hour. Those figures are company-reported. The company has not named the customers behind most of its performance claims, so the numbers are useful evidence of the product thesis rather than independently audited proof of broad adoption.

One official case study describes an unnamed global consumer-products company whose SAP material-requirements-planning workflow kept generating unnecessary purchase requisitions. Dodge AI says its system helped identify the root cause, design an enhancement, and reduce MRP-related support tickets by 30%, save 11 planner hours per plant each week, and deliver the solution 4x faster. Tech Funding News also relayed a company claim that another inventory-planning process became 132x faster.

The strongest signal in those examples is not the multiplier. It is the shape of the work. The product is being aimed at incidents, configurations, and process exceptions where a result can be compared with the ticket backlog, resolution time, planner workload, or delivery estimate already sitting inside the enterprise.

Why the Investors Care

Accel's Dodge AI profile describes autonomous agents that manage and maintain enterprise SAP systems. Accel partner Prayank Swaroop framed application maintenance as one of enterprise technology's largest and least modernized categories, arguing that maintenance creates the context layer agents need before they can safely operate in production.

That logic gives the financing a practical capital story. Enterprises already spend heavily on application-management services, often through large systems integrators and offshore teams. If software can resolve routine L1 and L2 incidents, document changes, and shorten delivery cycles, some of that budget can move from repeated human reconstruction toward a product that retains the knowledge.

The shift will not happen because an agent can produce a plausible answer. It will happen only if enterprise buyers can trace what the agent saw, why it recommended a change, which permissions it used, and whether the outcome survived production. Google brings model and infrastructure access. Accel brings enterprise-software pattern recognition. Dodge AI still has to turn those advantages into customer trust.

The Market Test Ahead

The company frames the opportunity as a $600B enterprise-maintenance market, but that figure is Dodge AI's own estimate and sits well above several outside definitions of application management and outsourcing. The broader point does not require accepting the largest number: complex enterprises already pay substantial recurring costs to keep customized systems running, and the knowledge required to do that work is often fragmented.

Dodge AI also faces larger and better-funded competitors approaching the same software estates from migration, modernization, or code-understanding angles. Its wedge is the daily queue. Incidents and change requests create urgency, measurable outcomes, and access to the local context that a broader modernization platform eventually needs.

That wedge carries risk. A maintenance agent that works across multiple systems must handle data access, auditability, approval boundaries, and the possibility that a technically correct change creates an operational problem elsewhere. Dodge AI's website shows SOC 2 Type I and Type II as in review, so security maturity remains part of the evidence customers will watch.

What This Funding Changes

The $2.65M Seed gives Dodge AI room to deepen the product, expand enterprise deployments, and hire around the engineering and implementation work required for production systems. Current founder-led recruiting in Bengaluru supports that expansion signal, although the company has not published a detailed hiring target or proceeds allocation.

The more important asset will be the context accumulated through each deployment. Every resolved incident can add another decision, dependency, and exception to the system's record. If Dodge AI can make that record reliable, permissions-aware, and reusable, maintenance becomes more than the entry point. It becomes the place where the enterprise finally starts retaining the explanation for how its own software works.