Explore the Key Agents and Prerequisites

Learning Objectives

After completing this unit, you'll be able to:

• Summarize the purpose of the four specialized agents in CLAIRE GPT.
  
• List the prerequisites required to use CLAIRE GPT in an organization.
  
• Explain the benefits of using CLAIRE GPT for data governance and productivity.
  

Prepare to Work with the Agents

CLAIRE GPT gets its usefulness from a set of specialized agents working behind the scenes. Each agent handles a different kind of task. Your organization also needs some setup before anyone can start asking questions. Knowing what each agent actually does, rather than treating CLAIRE GPT as one undifferentiated assistant, helps you know what kind of question to ask and what kind of answer to expect back.

Ana, the Cloud Kicks data analyst, gets her organization ready to use CLAIRE GPT and meets the agents that power her daily analysis. Before Cloud Kicks rolled out CLAIRE GPT company-wide, its administrators ran a short pilot with a handful of analysts, Ana included. The pilot demonstrated which agents mattered most for day-to-day work and confirmed the right data sources were connected.

The Four Specialized Agents

Ana relies most on the discovery agent to locate the right sales datasets. She uses the data quality agent to confirm those datasets are complete before she builds a report on them. Early on, she also uses the product help agent to navigate IDMC, though she uses it less often now that she's familiar with the platform.

Knowing which agent improves efficiency and helps you frame your questions more precisely. If you ask CLAIRE GPT to “Find me a dataset on shoe returns,” that's a discovery agent task. If you then ask “Is this dataset trustworthy enough to report on?” that's a task for the data quality agent. CLAIRE GPT routes your question to the right agent automatically, but understanding the distinction helps you interpret the answer and ask the right follow-up.

CLAIRE GPT Prerequisites

Before Cloud Kicks analysts can ask CLAIRE GPT questions, the IDMC administrator needs to put a few things in place. An individual analyst can’t configure any of these prerequisites alone. It’s still worth understanding them, even if you aren’t the one doing the setup. Knowing what’s required also helps you troubleshoot when CLAIRE GPT seems unable to find data you expect.

Here’s the CLAIRE GPT prerequisites.

• IDMC access with the appropriate license that includes CLAIRE GPT. Not every IDMC license tier includes CLAIRE GPT by default, so this is usually the first thing an administrator checks.
  
• Administrator setup, including enabling CLAIRE GPT for the organization and configuring which data sources and metadata it can access. This step determines the boundaries of what CLAIRE GPT is allowed to access on behalf of any user in the organization.
  
• Connected metadata sources. CLAIRE GPT can only discover and reason about assets that are already cataloged in IDMC. If a dataset has never been connected or cataloged, CLAIRE GPT has no way to know it exists, no matter how specifically you ask about it.
  
• Appropriate user permissions to view the underlying data and metadata that CLAIRE GPT surfaces in its answers. It respects existing access controls rather than granting new ones, which means that CLAIRE GPT doesn’t present data that you aren’t able to access.
  

For example, Ana used CLAIRE GPT to answer sales-related questions just fine. But before Ana could ask CLAIRE GPT about defect rates on the factory floor, she needed the right permissions. Until that setup was complete, CLAIRE GPT couldn’t answer questions about manufacturing since that data simply wasn’t accessible to it yet. The Cloud Kicks administrator connected the manufacturing data source, and then confirmed Ana's permissions covered that data. After getting the right permissions, Ana can ask cross-domain questions and get answers from sales and manufacturing systems in the same conversation.

Why Governance and Productivity Matter

CLAIRE GPT reinforces good data governance while it saves analysts time, and it’s worth understanding both sides of that benefit.

• Governance: CLAIRE GPT respects existing permissions and pulls from cataloged, governed metadata. It can't surface data a user shouldn't access. Its answers stay traceable to a real, governed source instead of a black box, which means that when Ana shares a finding with her manager, she can point to exactly where that finding came from rather than asking her manager to simply trust an AI-generated number.
  
• Productivity: Tasks that used to mean navigating several tools, such as checking lineage in one place and quality scores in another, now happen in a single conversation. This saves clicks and reduces the chance that an analyst skips a step, such as checking data quality, simply because it was inconvenient to open yet another tool.
  

With CLAIRE GPT in place, Ana cut the time she spent locating and vetting data for a typical report from a full afternoon to a few minutes, all while staying within the access and governance rules Cloud Kicks already had in place.

That time savings compounds across a team. What used to be a bottleneck for one analyst on one report becomes a small, repeatable step across every analyst and every report. Cloud Kicks expanded CLAIRE GPT access beyond its initial pilot group.

For data teams that need to demonstrate audit readiness, the governance side is equally important. Because CLAIRE GPT answers are grounded in cataloged, permission-controlled metadata, a team can point to the source of any AI-assisted finding rather than treating the answer as opaque. That traceability is what makes CLAIRE GPT a tool that works with your data governance strategy rather than around it.

Wrap Up

You now have the CLAIRE GPT agentic workflow and specialized agents in your toolkit. You're ready to ask better questions of your data, trust the answers you get, and spend less time searching for data and more time analyzing it.