🔍Exploratory Mode
Exploratory Mode is a special mode of the Euno AI Assistant designed for identifying missing context and institutional knowledge. It is used during onboarding or whenever you suspect the assistant (or MCP) is missing important terminology, definitions, or rules that your organization uses. In Exploratory Mode, the assistant still answers your questions—but it is tuned to surface uncertainty, ask you to clarify, and help you capture gaps as Context Issues so you can fix them with metadata tags or AI instructions.
What Is Exploratory Mode?
In normal use, the assistant and MCP try to give a single, confident answer. They search the catalog, use your AI instructions and metadata tags, and return a result. When the right answer is ambiguous or the assistant doesn’t know your org’s meaning of a term, it may guess—and sometimes guess wrong.
Exploratory Mode changes the assistant’s behavior:
It still answers your questions and uses the same tools (EQL, lineage, etc.).
When the answer is uncertain (e.g. several tables could fit “retention”), it shows you the options and asks which one is correct and why, instead of picking one.
When it doesn’t recognize a term you use, it asks you to explain what you mean so it can learn.
When it identifies missing context or institutional knowledge—especially after you’ve clarified—it offers to file a Context Issue with a draft report and only submits after you confirm.
So Exploratory Mode is the same assistant, with a different goal: not only to answer, but to discover what Euno doesn’t yet know and help you record it.
What Is Exploratory Mode For?
Exploratory Mode is for:
Onboarding — When Euno is new to your account, the assistant doesn’t yet know your vocabulary, certification rules, or preferences. Running in Exploratory Mode while you ask real questions helps you find those gaps and turn them into Context Issues, then into tags or AI instructions.
Governance and tuning — Whenever you’re concerned that the assistant is missing critical context (e.g. after a wrong answer or a vague term), switch to Exploratory Mode, reproduce the situation, and let the assistant help you articulate and report the gap.
Training the catalog — The outcome of Exploratory Mode is usually a Context Issue. Resolving that issue means adding or refining metadata tags or AI instructions. Over time, the assistant and MCP get better because the catalog encodes your institutional knowledge.
Exploratory Mode does not change your data or integrations; it only changes how the assistant behaves in conversation and whether it suggests filing Context Issues.
How to Use Exploratory Mode
Who Can Use It
Exploratory Mode is available to users who have both:
Access to the AI Assistant, and
Permission to use Exploratory Mode (e.g. admins and data/AI governance roles, depending on your account setup).
If the option is not visible, your account may not have the feature enabled or your role may not have the required permission.
Turning Exploratory Mode On and Off
Open the AI Assistant from the main navigation.
In the assistant sidebar (when expanded), find the Mode section.
Turn Exploratory mode on. The assistant will now use the exploratory behavior for that conversation (and new messages in that thread).
Turn it off when you’re done so the assistant returns to normal, single-answer behavior.
Exploratory Mode is per conversation: it applies to the current thread. New chats start in the default mode unless you turn Exploratory Mode on again.
What to Do in Exploratory Mode
Ask real questions — Use the same kinds of questions you’d ask in normal mode: “What’s our 12‑month retention?”, “Which table feeds the Revenue dashboard?”, “What does ‘churn’ mean in our catalog?”
When the assistant shows options — If it lists several possible resources or interpretations and asks which is correct, pick the right one and briefly explain why. That explanation is exactly the kind of institutional knowledge that can be turned into a tag or an AI instruction.
When the assistant asks for a definition — If it says it doesn’t know a term and asks you to explain, give a short, precise definition or point to the right resource. That can be captured in a Context Issue and then in metadata or AI instructions.
When the assistant offers to file a Context Issue — It will show a draft report (what’s missing, where it came up, what you decided). Edit the draft if needed, then confirm. The assistant will submit the issue and post the link. You can then resolve it by adding or updating metadata tags or AI instructions.
You don’t have to file a Context Issue every time; the goal is to use Exploratory Mode when you want to improve the assistant by finding and recording missing context.
How It Relates to Context Issues and Institutional Knowledge
Exploratory Mode and Context Issues work together:
1. Discover
Exploratory Mode (assistant)
You ask questions; the assistant surfaces uncertainty, asks for clarification, and suggests filing a report when it identifies missing context.
2. Capture
Context Issues
You (or the assistant with your confirmation) submit a Context Issue that describes what’s missing and in what situation.
3. Resolve
Metadata Activation & AI Settings
You add or update metadata tags or AI instructions so the assistant and MCP have the right context next time.
Institutional knowledge is the unwritten or semi-written knowledge your organization uses—terminology, definitions, certification rules, conventions. Euno’s assistant starts from indexed metadata (schemas, lineage, tags, etc.); it doesn’t automatically know your org’s precise meanings. Exploratory Mode helps you find those gaps; Context Issues record them; tags and AI instructions encode them so the assistant behaves correctly. For a longer explanation and a mapping of institutional knowledge to tags and AI instructions, see Context Issues — Institutional Knowledge and How It Maps to Euno.
Example Flow
You turn on Exploratory Mode and ask: “What’s our subscription retention at 12 months?”
The assistant searches the catalog and finds several tables that could represent “retention” or “subscription retention.” Instead of picking one, it lists the options and asks which one is correct and why.
You choose the correct table and explain: “We use
subscription_retention_12min the subscription mart; the others are older definitions.”The assistant offers to file a Context Issue with a draft that summarizes: missing clarity on which resource is “subscription retention at 12 months”; you clarified; suggests adding a tag or AI instruction.
You confirm; the assistant submits the issue and posts the link.
You open the Context Issue, then add an AI instruction (e.g. “When we say ‘subscription retention at 12 months’ we mean the metric in
subscription_retention_12min the subscription mart”) or a metadata tag that marks that resource. You mark the issue Resolved.
Next time someone asks about 12‑month subscription retention, the assistant (in normal or Exploratory Mode) can use that instruction or tag to give a confident answer.
Related Documentation
Context Issues — What Context Issues are, how to create and manage them, and how institutional knowledge maps to tags and AI instructions.
AI Assistant — Overview of the assistant and how to use it.
AI Directives — How to set account-level AI instructions.
Metadata Tags — Types of tags and how to create them.
Last updated