The San Francisco startup, which says it works with more than a dozen large companies, wants AI agents to take over upkeep work now done by outsourced teams.
Dodge AI, a startup that uses artificial intelligence to maintain large enterprise software systems, said Sept. 29 that it had raised $2.65 million in funding led by Accel and Google’s venture arm.
New Build Venture Capital, Antler, Schema Ventures and angel investors from the SAP ecosystem also took part, the company said.
Dodge AI is building what it calls a control plane, software that sits across systems such as SAP, Salesforce, Microsoft Dynamics, Kinaxis and Oracle JD Edwards. The company said the platform resolves incidents and change requests and records the custom rules and exceptions that businesses have built into those systems over the years.
The problem it targets
Once enterprise software goes live, the business keeps changing, and companies add thousands of their own rules over time, Dodge AI said. When something breaks, the cause is rarely in one place. The company said that keeping such systems running costs enterprises more than $600 billion a year, a figure the announcement did not cite a source for.
Maintenance has traditionally gone to large systems integrators such as Accenture, TCS and IBM, which typically assign 20 to 50 people, often offshore, to handle incidents, change requests and background jobs, according to the company. Dodge AI argued that this model leaves fixes undocumented and customizations piling up, and deepens a company’s dependence on its contractor because the knowledge lives in tickets, consultants and memory.
“For a long time, the only way to maintain enterprise systems was to add more people,” said Rebhav Bharadwaj, co-founder and CEO. “We believe long-horizon agents can change that.”
Also Read: The New Boardroom Problem: How Do You Measure AI’s ROI?
Early customers and results
Dodge AI said it works with more than a dozen enterprises, half of them publicly traded, on incident management and process optimization across SAP, Kinaxis and Microsoft Dynamics. It said the platform handles hundreds of queries an hour. The company did not name its customers.
The company described two cases. In the first, a truck could not be loaded at a warehouse because a goods receipt note was printing incorrect information. Dodge AI said it traced the fault across SAP, Kinaxis and the customer’s warehouse software and delivered a fix within minutes.
In the second, a customer ran inventory planning overnight because SAP crashed during the morning run. Dodge AI said it modernized the process and made it 132 times faster, freed the team of 10 people who maintained it and cut order allocation time by eight hours.
Also Read: What Are VLA Models? How Vision, Language and Action Work
What comes next
Dodge AI sees maintenance as the entry point to broader modernization. Chief information officers are reluctant to risk breaking systems that work, it said, and tight budgets are consumed by daily incidents. The company also wants its record of exceptions and customizations to serve as a source of truth for AI agents operating in production.
“Frontier models are incredibly capable, but enterprise adoption still requires significant hill climbing,” said Aditya Thakur, co-founder and chief technology officer.
Prayank Swaroop, an investor at Accel, called application maintenance “one of the largest and least modernized categories in enterprise technology.”
