What changed
Businesses are increasingly asking AI tools questions about live operational data instead of relying only on dashboards, static reports and exported spreadsheets. SyncHub, an Auckland-based data integration provider, released a custom database connector in October so customers can query cloud software alongside on-premise systems, legacy databases and in-house applications.
The shift is being enabled by Model Context Protocol, or MCP, which connects tools such as SyncHub’s AI Assistant, Claude and ChatGPT more directly to business systems. The questions cover sales, stock, staffing, invoices and performance.
Why it matters
This changes the role of the dashboard. A manager may no longer need to hunt through several screens or ask a technical colleague to write a query before getting a basic answer. They can ask the question directly, provided the underlying data has been cleaned, structured and connected.
That condition is the important part. MCP can open a route to live information, but it cannot repair mismatched formats or unreliable records. Businesses often keep accounting, customer, human-resources, inventory and timesheet data in separate systems. Individual software vendors may connect their own data through MCP, but that still leaves a fragmented view of the wider business.
SyncHub’s connector reflects another pressure: AI-assisted coding has made it easier for companies to build internal applications, creating more databases outside standard SaaS platforms. Bringing those systems into the same queryable layer may make AI more useful, but it also makes data organisation a prerequisite rather than a tidy-up task for later.
The sensible near-term model is hybrid. Conventional reports and dashboards still suit formal reporting and ongoing monitoring. Natural-language querying is becoming the faster first step when someone needs a specific answer. The real test will be whether businesses can trust those answers across their full technology stack, not whether an AI assistant can produce a fluent response.
Comments
No comments yet.