Keep your data below decks

For a small firm, the data is the business, and SaaS tools make it easy to forget who holds it. Before you add AI, make sure your most valuable data sits in something you own.

I worked with a 35-person executive search firm whose value is its relationships: which candidates it knows, who it has placed where, and which clients come back. All of that lived in SaaS tools, chiefly an applicant tracking system and a CRM. They are good tools, and the firm kept them. The problem was that the history of the business existed only inside them, in the vendors' formats and behind the vendors' exports.

What does below decks mean for a small business?

On a ship, what matters most is stowed below, secured, and not at the mercy of the weather on deck. For a firm, below decks is a database the firm owns, in an account the firm controls, holding the records that make up the business: candidates, clients and placements.

The SaaS tools stay on deck, where people work every day. The core does not replace them. It makes sure they are not the only place the business exists.

SaaS at the edges

The rule is simple. People keep working in the tools that suit them. Data flows from those tools into the core, where it is kept in one shape the firm defines rather than in each vendor's. Anything that needs the whole picture, from reporting to AI, reads from the core and never directly from the vendors.

That rule has two useful consequences. Changing a vendor becomes a matter of connecting a new tool, not of migrating the firm's memory. And questions that cross tools, such as which clients have hired candidates the firm first met years earlier, can be asked in one place.

Why it matters for AI

AI makes a data core more valuable, not less. Retrieval over a firm's own history is only as good as the history it can reach. Pointed at one well-kept core, an AI assistant can draw on the firm's candidates, clients and placements together. Pointed at several vendors, each with its own export and permissions, it sees fragments.

The core also keeps the firm in charge of what it shares. Every AI vendor will offer to connect to the firm's tools. With a core in place, the firm decides what an AI service may read, and it can change AI services as the market moves without handing its history to each one.

For a firm of this size, the core does not need to be a large platform. A well-modelled database, dependable sync from the tools that matter and clear ownership of the account are the whole of it. The discipline is in the model and the rule, not the size of the stack.

Where to start

Start with a list, not a platform. Write down the records that would hurt most to lose, or to be unable to move: for a search firm, candidates, clients and placements. Put the core in an account the firm owns, not one held by a supplier or a contractor.

Then sync one tool, check the records against the source, and point one report at the core. Once the firm trusts that report, it is time to point AI at the same data.

What carries over

• Know which data is the business, and keep it in something you own.
• Let SaaS tools be where people work, not where the business is stored.
• Sync into one shape you define, and point reporting and AI at the core, never at the vendors.
• Choose AI services that read from your core, so you can change them without moving your history.

Cloud tools make an excellent deck. Keep the cargo below.