Art Galleries Need a Unified Database Before AI Can Deliver Real Value

Context: Many Nigerian galleries operate across dozens of spreadsheets for exhibitions, stock, shipping, and pricing. This fragmented web functions only because staff spend hours navigating it daily. AI cannot impose order on chaos; it will merely speed up the confusion if the underlying data remains disjointed. The first step toward useful automation is creating one source of truth for every artwork, artist, contract, and image.

Key Facts: The computing adage “garbage in, garbage out” reminds us that poor input yields poor output. Luca and Will from Sydney University were chosen for their disciplined, practical approach rather than grandiose AI visions. Their project followed the migration principle to retain all data during transition, flagging ambiguities for review but discarding nothing of value.

What It Means: By building a structured database and a colleague‑facing web app first, the gallery created a stable foundation. The team kept the scope fixed, retained ownership of all code, and excluded advanced AI agents until the core system proved stable. This guardrail approach reduces overreach and ensures the technology adapts to workflows, not the reverse. Once the database is live, AI can support tasks like pricing suggestions, provenance tracking, and demand forecasting with reliable inputs.

Watch next: Pilot AI agents for automated pricing and condition reporting, measuring adoption rates and accuracy improvements over the next six months.