Day 2 Ops
AI Copilot for Day 2 Ops sits on top of Exemplar's service catalog, integrations, and Context Lake. Ask questions, draft automations, and invoke allowed actions with guardrails—whether you are in the console, your editor via MCP, or a script calling APIs.
IDE and MCP clients use the same tools as the dashboard
Monitoring, catalog, and webhooks surface as tools—one policy surface for chat and MCP.
Ask in plain language; the copilot maps intent to catalog queries, runbooks, webhooks, and platform APIs—so you are not memorizing twenty URLs.
Answers tie back to services, owners, dependencies, and live signals from integrations—not a generic model guessing about your stack.
Expose the same actions to Cursor, Claude, and other MCP clients: one governance model whether you click in the console or type in the editor.
Sensitive actions stay behind approvals, policies, and audit trails. The copilot suggests and executes within the boundaries your platform team defines.
Illustrative scenarios—actual behavior depends on your integrations and policies
Find workloads, who owns them, and what depends on what—without opening five tools.
Q: Who owns checkout in prod and what did we deploy last week?
A: Illustrative: the copilot resolves checkout to the owning squad, linked services, and recent deploy events from your connected CI and catalog.
Shorten time-to-context when something breaks: symptoms, recent changes, and owners in one thread.
Q: Errors spiked on billing-api—what changed?
A: Illustrative: summarize recent deploys, dependency shifts, and linked incidents using catalog and observability data you have connected.
Describe the outcome; wire webhooks, notifications, or self-service actions with review where needed.
Q: Notify the platform channel when a prod deploy fails.
A: Illustrative: draft a failure hook to your collaboration tool with retries and signing—subject to your org’s automation policies.
Pick the surface that matches the workflow
Built-in assistant in the Exemplar dashboard for operators and developers.
Model Context Protocol for AI clients that support tools—Cursor, Claude Desktop, and compatible IDEs.
Programmatic access for scripts, CI, and advanced integrations where chat is not enough.
Capabilities
One short example per area—rollout and tooling vary by tenant.
Turn questions into catalog lookups, summaries, and next steps—without switching context constantly.
You: Which services still run Node 18?
Copilot: Illustrative: returns candidates from catalog metadata and ownership so upgrade campaigns have an assignee per row.
Suggestions for cost, risk, and hygiene—scoped to what Exemplar can see from integrations and policies.
You: Where should we tighten IAM first?
Copilot: Illustrative: prioritizes findings linked to production paths and teams—your security tools and rules still govern enforcement.
Describe automations in prose; get structured steps, configs, or handoffs to operators for approval.
You: Back up stateful data before the database migration window.
Copilot: Illustrative: outlines checks, scheduling, and notifications aligned with webhooks and automation you have enabled.
Uses the service or incident you are in—so answers are specific, not generic platform advice.
You: I am on payment-service—what should I know before this release?
Copilot: Illustrative: summarizes owners, dependencies, recent changes, and policy notes from the graph around that entity.
At a glance
Tools
Catalog, SRE, automation, webhooks—via MCP and chat
Context
Grounded in your graph and integrations
Guardrails
Policies, approvals, and auditability
Open the console to explore the platform—service catalog, integrations, and AI Copilot capabilities.