Rewiring Carta Fund Admin in the Era of AI

When we launched Carta Fund Administration in 2018, we set out to change a broken industry. The service model for venture and private equity funds relied on spreadsheets, disconnected tools, and human workarounds. By combining technology and service, we built a system of record for fund managers and their investors.

We introduced event-based accounting. We built workflows that allowed fund managers to execute “events” like capital calls and investments on Carta. The product automatically updates the general ledgers and all the outputs like Financials and PCAPs. It was a leap forward for the private markets: a unified, structured way to replace manual, error-prone processes.

In 2025 we’ve come so far.  We’ve expanded into building the ERP of Private Capital. Other than core fund accounting, the Carta platform has a suite of products from portfolio management, fund modeling, and valuations to closings, data room, and LP portal to name just a few. 

We can go even further. Technology has unlocked new possibilities for connectivity, and Carta is at the forefront of the transformation of private capital markets infrastructure. In our next chapter, this event-based accounting foundation allows us to 10X this platform through AI. 

Moving from event-based to agent-orchestrated

Events have been the anchor of our platform. Today, the workflows that help execute these events, like a valuation update, require human input. 

In this next phase, agents are the anchors.

Carta agents are trained in their specific task domain and combine Carta’s workflow automation with judgment. This takes even more work off of clients and our fund admin team. 

We’re building four types of Agentic/AI experiences (so far): 

1. Continuous monitoring agents 

2. Workflow agents 

3. Configuration agents 

4. Insight agents 

Our continuous monitoring agents look for gaps and discrepancies in the system to proactively address. For example, our SOI tagging agent will make sure all portfolio companies in a firm are tagged with the right geographies and industry to keep the SOI audit ready. Our cash reconciliation agent ensures that cash is always reconciled, and our health check agents spot discrepancies in the accounting to proactively resolve them. 

Our workflow agents can engage with clients and fund admins to make complex work quick and easy. Our expense allocation agent allows users to split expenses automatically through a preconfigured rule or on the fly with natural language input. Workflow agents power our closings tool that helps LPs easily complete a closing process in minutes (and thank goodness for that!). 

Our configuration agents allow the platform to tailor the experience for each firm, fund, and user. Today, we’re storing most LPA terms in our fund properties page as structured rules. Soon, agents will allow us to automatically translate the LPA to set up fund properties. Similarly, we’re building reporting agents that can help firms set up custom reports with calculations specific to their needs. 

The insight agents help firm users easily access data within Carta. This year we launched our Data Warehouse product, which aggregates firm data across captables, fund admin, LP administration, Fund Forecasting, and more into a firm-specific Snowflake instance. Our first version of the insights agent will allow you to write natural language prompts against this data warehouse. Eventually, insights agents will be embedded through the Carta product. Think AI powered tracebacks that would enable you to understand the lineage and calculations that led to a specific metric. Think proactive AI-powered alerts about what a user should pay attention to when they log into Carta. There’s so much possibility. 

Our agentic future is about more than just automation. It’s a shift from reactive fund administration to proactive orchestration—where the system anticipates what needs to happen next and executes without waiting for you to tell it.

Here’s what makes this model so powerful:

• Always On: Agents continuously monitor connected systems—Carta, banks, portfolio company data, LP records—to detect relevant signals instantly.

• Contextual Intelligence: Each signal is processed in context grounded in the LPA. For example, an LP contribution triggers different workflows depending on fund structure, investor terms, and entity relationships.

• Seamless Outputs: the signals directly inform transactions which makes it possible for financial reports, cap table updates, and LP communications to happen in near real-time.

With agents, we remove the need to log events manually. The fund admin platform does the listening, the reasoning, and the orchestration for you automatically.

How we use AI at Carta 

Naturally, the way companies use AI is top of everyone’s mind. At Carta, we use AI to improve our products, enhance the overall  experience, and help customers manage equity and their funds more efficiently.  We may use anonymized and aggregated data shared through the platform to support product development, analytics, and research – but never in a way that identifies individual users. While data can help power smarter features, we can confirm unequivocally that Carta does not train internal or third-party generative AI models on customer data. Trust that your information remains secure, and that we’re committed to sharing how we use AI to build better tools. For more details, visit /security/generative-ai-at-carta/.

Building an intelligent layer for fund administration

We’ve spent the last decade building the infrastructure for private markets. AI-powered agents are the next layer, a networked intelligence that ties these components together into a single, adaptive platform.

Our job is to build the platform that removes operational friction so GPs, CFOs, and LPs can focus on what they do best. Moving from event-driven to agent-orchestrated fund administration isn’t just a product evolution. It’s a step-change for the industry that brings transparency, speed, and intelligence to every layer of private capital operations.

–Vrushali