AGS AI Analytics

Last updated on July 23, 2026
note

AGS AI Analytics is in early access and connects to AI assistants that support the Model Context Protocol (MCP). Rendering charts and dashboards inside the conversation also requires support for the MCP Apps extension. Availability may vary by plan.

Overview​

AGS AI Analytics lets your whole team ask questions about your player data in plain language and get answers back in seconds. It connects an MCP-compatible AI assistant to AccelByte Gaming Services (AGS), so you can ask about the same game telemetry and analytics events that power your analytics dashboards, without writing a query by hand or filing a ticket to your data team.

Out of the box, a question returns a general answer. When you add a short context file that describes how your game works, the same question comes back segmented, ranked, and specific to your game. You can then pin the charts you care about into a dashboard your team can refresh on demand.

Who it's for​

AGS AI Analytics changes who gets to ask. Instead of routing every question through whoever owns the query language, anyone on your team can ask directly, in plain language, and get an answer back in the same conversation. It is built for everyone who needs player data but does not want to write SQL or wait on someone who can:

  • Live Ops teams measuring how an event or campaign performed.
  • Product and design teams validating whether players engage with new content as intended.
  • Player support and community teams answering player questions quickly.
  • Analysts and engineers who want a faster path to a first answer before a deeper dive.

How it works​

AGS AI Analytics turns your plain-language question into a query, runs it against your analytics data, and returns the result as an answer or a chart. You bring your own AI assistant, and AGS provides the data and the game context.

  1. Connect your AI assistant to AGS. Add the AGS AI Analytics MCP server to an AI assistant that supports MCP, then log in to your environment. To get charts and dashboards rather than text-only answers, use an assistant that also supports the MCP Apps extension.
  2. Ask a question in plain language, for example, "Show me the active players for the last 30 days." Out of the box, you get a general answer.
  3. Add a context file that describes what counts as a real player, which accounts are test or quality assurance (QA), what your recent events were, and what your custom events mean.
  4. Ask again. The same question now returns a segmented, ranked answer that reflects how your game actually works.
  5. Pin charts to a dashboard. Keep the charts you care about, such as registrations, login volume, or Matchmaking health, in one view your team can refresh.

For the full walkthrough, see Getting started with AGS AI Analytics.

What you can ask​

The questions are your own, in plain language. Some examples of what teams explore:

  • Which players returned after an event ended, and which dropped off.
  • How a newly shipped feature is affecting engagement and revenue.
  • Daily and monthly active players, new versus returning registrations, and retention trends.
  • Which bundles sold best last month, kept split by currency and login platform.
  • Where players drop off in a reward campaign, from eligible to rewarded to purchasing.
  • Whether sign-ins are failing more on a specific path, such as direct, Steam, PSN, or Epic.
  • Matchmaking wait times (median and 95th percentile) and ticket failure rates by region.
  • Whether a release caused a drop in logins, sessions, or completed matches.

Because you ask follow-up questions in the same conversation, one question can grow into a full picture, for example turning "how are players doing?" into a ranked list of who to re-engage first.

Key concepts​

  • Plain-language queries: You describe what you want in natural language. AGS AI Analytics generates and runs the underlying query for you, so no one on your team needs to know SQL.
  • Context files: A context file is a short document that teaches the AI assistant how your game works, covering player eligibility, event windows, and the meaning of your custom events. Context is what turns a generic answer into a game-specific one. See Writing a context file.
  • Custom events: Beyond AGS standard telemetry, you can query your own game telemetry events. Sending each event type under its own event name keeps queries fast and specific.
  • MCP Apps: The interactive charts and dashboards render directly in your conversation through the MCP Apps extension to MCP. Your AI assistant must support MCP Apps for these visualizations to appear; without it, the same answers come back as text only.
  • Dashboard: Pin AI-generated charts into a shared view and refresh them on demand.

Next steps​

Last updated on
On this page