Ask any compensation platform vendor whether you can get your data into Tableau or Power BI, and the answer is always yes. Ask how, and the answer is almost always a pipeline: schedule an export, stand up a sync, land a copy in a warehouse, point the BI tool at the copy. It works, which is why the whole industry says yes. It also means the data your analysts trust most is a duplicate: aging from the moment it lands, maintained by an integration someone has to own, and governed by whatever rules the copy’s new home happens to enforce.
Optymyze offers a different answer, in two parts. Direct data access lets customers query their own Optymyze data live, from their own BI tools, with no pipeline and no copy. The Optymyze MCP server lets their own AI agents do the same in natural language. The principle behind both: your data should be usable where you work, and the governance should travel with the connection.
A live connection, not another copy
Direct data access is a certificate-based database connection from your BI environment straight into your Optymyze data. Analysts build dashboards, join compensation results with finance and HR data, and answer their own questions against live data, in the tool they already use every day.
What makes this different from the export-and-sync pattern is where the governance lives. The published data set is assembled in the application layer, which means anything outside it does not exist at the connection point. There is no broader schema an ambitious query can wander into, and no second copy drifting out of date or out of policy. Certificates are issued per connecting client and can be revoked individually, credentials are still required, and access is read-only. The security conversation with IT is short, because the answer to “what can this connection reach” is the published set, and nothing else.
Ask in natural language
The MCP server takes the same governed connection and puts an AI agent on the far end. MCP, the Model Context Protocol, is the standard the industry is converging on for giving AI systems structured access to tools and data. The protocol itself is not the story; most platforms will have an MCP endpoint eventually. The story is where the agent runs and what it is allowed to see.
The Optymyze MCP server registers as a tool inside the customer’s own agent environment, such as a Copilot Studio agent running in their own tenant under their own enterprise agreement. The AI sits on the customer’s side, inside their compliance boundary. Optymyze is not running an AI over customer data; the customer’s own agent is asking governed questions of the customer’s own published set. An analyst types “show me attainment by region for the last two quarters” instead of exporting three files and spending an afternoon in pivot tables. The agent sees exactly what the connection is allowed to see, which is exactly what the governance says it should.
The boundary that makes it trustworthy
Here is the part most AI announcements skip. Answers that come back through direct data access and the MCP server are for exploration and analysis. The platform remains the system of record for anything that determines a payment. A dashboard can show attainment; an agent can explain a trend; neither one moves a dollar. Compensation runs on calculations that are versioned, approved, and auditable inside the platform, and that does not change because the question arrived in natural language.
This boundary is what makes the capability safe to adopt. An agent exploring governed, read-only, scoped data can be rolled out without betting payroll accuracy on a language model, and the audit posture your governance team already approved still holds.
Where this goes
The questions customers raise the moment they see this point the same direction: can payees ask about their own results, can AI help design and change plans. The path there runs through certified data services, where an answer an agent returns carries the platform’s own calculation logic behind it rather than its own reading of raw tables. The metadata-first architecture Optymyze is built on is what makes that step possible, and the governed access available today is the foundation it stands on.
Your data has always been governed. Now it is governed where you work. To see direct data access or the MCP server against your own environment, talk to your Optymyze account team or book a demo.
