The news
Dodge AI, a San Francisco startup whose AI agents maintain corporate software, announced $2.65 million in seed funding on September 29, 2026, led by Accel and Google's venture arm. It is going after work that enterprises typically hand to large consulting teams.
New Build Venture Capital, Antler, Schema Ventures and angel investors from the SAP ecosystem also joined the round, according to the company's announcement. SiliconANGLE, which reported the deal, identified the Google investor as its AI Futures Fund. Neither source disclosed a valuation.
Dodge AI describes its product as a control plane, a central layer that sits across a company's business applications, for maintaining systems such as SAP, Salesforce, Oracle and Microsoft Dynamics. Its agents take incidents and change requests, trace the root cause and fix or recommend fixes. The company says the system also records undocumented customizations and exceptions, building a store of operational knowledge as it works.
The startup says it is already working with more than a dozen enterprises, half of them publicly listed, and handling hundreds of queries per hour. It did not name customers. In one example, the company says its agents traced a warehouse truck-loading failure across several systems and fixed it within minutes; SiliconANGLE reported the cause was an incorrect goods receipt note recorded in multiple systems.
In another case, Dodge AI says it rebuilt a customer's overnight inventory planning process, making it 132 times faster, freeing the 10-person team that maintained it and improving order allocation time by eight hours. These are the company's own figures from unnamed customers.
The company was founded by chief executive Rebhav Bharadwaj and chief technology officer Aditya Thakur; Aditya Patil is chief operating officer. In the announcement, Bharadwaj said that for a long time "the only way to maintain enterprise systems was to add more people."
The numbers
- Seed round
- $2.65 million
- Annual cost of keeping enterprise systems running (company estimate)
- More than $600 billion
- Typical integrator team per deployment (SiliconANGLE)
- 20 to 50 people
- Enterprise customers (company claim)
- More than a dozen, half publicly listed
- Speed-up on one customer's inventory planning process (company claim)
- 132 times faster
Why CEOs should care
For CIOs who run SAP or Salesforce through an application-maintenance contract, Dodge AI is aiming straight at that line item. Dodge AI says keeping these systems running costs enterprises more than $600 billion a year, and SiliconANGLE noted that integrators such as Accenture (ACN) and IBM (IBM) often assign teams of 20 to 50 people to a single deployment. If agents can close even routine tickets, the size of those teams becomes negotiable.
CFOs should treat the headline claims with care. The 132-times speed-up and the freed 10-person team come from a single unnamed customer and are not audited benchmarks. A better use of them is as a prompt at renewal time: ask your integrator how much of its ticket volume is repeat, rules-based work, what share it already automates, and how its price would change if an agent handled the rest. Ask any agent vendor for a pilot on your own incident backlog, with the before-and-after numbers measured by your team.
CISOs and audit committees have a different question. An agent that can change ERP configuration, the settings that drive orders, inventory and finance, needs the same controls as a human administrator: least-privilege access, segregation of duties, a change ticket for every fix and a human approval step for anything that touches production. Dodge AI says it documents custom logic as it works; buyers should check that this record is complete enough for auditors and stays in their own environment.
The bigger picture
The pitch targets the business model behind much of enterprise IT services. Bharadwaj's point is that maintenance has scaled with people; agent startups argue software can absorb a growing share of that work. The backing of Accel, Google's venture arm and angels from the SAP ecosystem suggests investors think this category is worth testing, though a $2.65 million seed round is small next to the market Dodge AI says it is chasing. Large integrators have the customer relationships and the system knowledge; the open question is whether they partner with agent startups, build their own tools or wait to see whether results like Dodge AI's hold up at scale.
What’s next
The announcement did not name customers, so the next test is public references from the publicly listed companies Dodge AI says it works with, and independent measures of ticket volume closed without human help. Watch also for how large integrators respond, since any shift in how application maintenance is priced would show up first in their contract renewals.
What “Fact-checked” means
Fact-checking means testing a story’s facts against the evidence before it is published. This story went through at least two separate checks before this version was published.
- What we checked
- Its names, figures, dates, job titles, quotes and who said what were checked against the story’s sources, including its main source where it could be opened. The headline was checked for accuracy and overstatement.
- How
- A first check reviewed the whole story. If it passed, a second, skeptical check went back to the sources to look for mistakes in the most important facts. If a check flagged the story, it was edited to fix the problems found, and a separate re-check then reviewed the whole story again.
- Who
- The checks are made by our newsroom, as steps kept separate from the writing, under rules set by our editor, Hussein Mukhtar. A story the checks still flag is held for the editor, who decides whether it is fixed, published or dropped.
- If something is wrong
- “Fact-checked” does not mean error-free. If a material error is found after publication, we correct the story and add a note saying what changed. Report an error
Companies in this story
