Second Brain

There’s a term in enterprise software called “system of record.” It refers to the authoritative source of a particular type of data. Salesforce is your system of record for customer data. Workday is your system of record for employee data. NetSuite is your system of record for financial data.

Being the system of record is the most defensible position in enterprise software. It doesn’t matter if your UI is clunky or your features lag behind competitors. If you’re where the data lives, everyone has to come to you. Integrations point at you. Workflows route through you. You are, in a very literal sense, the source of truth.

This position is about to collapse.

The thing about systems of record is that they don’t actually contain that much data.

Your entire Salesforce instance probably fits in a few megabytes to at most gigabytes. Every account, contact, opportunity, and activity for your whole company. Your Jira history, every ticket ever created, is maybe a gigabyte if you include images. Your HubSpot, your Asana, your Monday boards, your Notion workspace: all of it together might be smaller than a single video file on your phone.

This didn’t matter when the only way to interact with the data was through the application’s interface. The data was small, but it was also trapped. You had to log into Salesforce to see Salesforce data. The application and the data were fused together.

AI agents break this assumption.

To be useful, an AI agent needs access to your systems. So you connect it to your CRM, your project management tool, your documentation. You give it API credentials. This is the whole point. You want the agent to be able to look things up, update records, take actions on your behalf.

But here’s what happens next.

The agent, in the course of doing its job, starts pulling data. It needs context to answer your questions. It needs to understand your customers to help with sales. It needs to see your projects to help you manage them.

The data is small. The API is permissive. Within minutes, maybe seconds, the agent has a complete copy of everything.

Now you have an interesting situation.

The agent is your primary interaction layer. You talk to it, not to Salesforce. You ask it questions, not the HubSpot search bar. The agent is also sitting on a complete, synchronized copy of your data.

At this point, what exactly is the system of record doing?

The original application still exists. People still log into it occasionally. But increasingly, it’s just a write endpoint. It’s a place where data gets created before being sucked into the agent’s context. The agent is where the work actually happens.

Some systems of record will try to fight this. They’ll implement rate limits. They’ll restrict API access. They’ll try to keep the data trapped inside their walls.

This is a losing strategy.

If you rate-limit an agent too aggressively, it becomes useless. The whole point of connecting an agent to your CRM is so it can help you with your CRM. If it can only make ten API calls per minute, it’s not much of an agent.

But there’s a deeper problem. A sufficiently sophisticated agent, facing rate limits, will simply build a caching layer. It will pull what it can, store it locally, and sync periodically. The rate limits just delay the inevitable by a few hours.

Once the agent has a cache, you have to ask: why isn’t the cache the system of record? It’s more current than what the agent can access through rate-limited APIs. It’s faster to query. It’s where the agent actually does its work. And the agent is now the primary interface.

The system of record tried to protect itself and instead made itself irrelevant.

There’s a version of this story where systems of record survive by pivoting.

Maybe they become the “governance layer.” The place where permissions are defined, audit trails are kept, compliance is enforced. Multi-user sync is still hard. Permissioning is still hard. Knowing who changed what and when still matters for regulated industries.

But this is a much smaller business than being the system of record. Governance is a feature, not a platform. And it’s not clear that the current incumbents are well-positioned to provide it. Their entire architecture assumes they’re the center of the universe, not a compliance sidecar to an agent.

The companies most at risk are the ones whose entire value proposition is “we store your X data.”

CRMs. Project management tools. Documentation platforms. Applicant tracking systems. Any category where the data is structured, relatively small, and the main moat is that the data lives there.

These companies have had decades of dominance because the data couldn’t leave. Now it can leave in seconds. And once it leaves, it doesn’t need to come back.

If you’re building a system of record today, you should be asking yourself some uncomfortable questions.

What is your product if the data doesn’t have to live with you? What’s the value of your UI if no one’s looking at it? What’s your moat if an agent can clone your entire database before your rate limiter even kicks in?

The answers might be workflow automation that’s hard to replicate, network effects between users, deep domain expertise embedded in the product. But “we’re where the data lives” is no longer a valid answer.

For twenty years, that was enough. It’s not anymore.