Source: https://internal-agents.com/agents/embat-mr-batt # Embat — Mr. Batt Embat's internal agent platform began as a support tool for the Product Experience team and now serves the whole company through Slack. A router sends each message to one of more than ten specialist agents, each mapped to a module, task, or tool set. - Company: [Embat](https://internal-agents.com/organizations/embat) - Collection: Agents - Approach type: Agent family - Deployment stage: Deployed - Autonomy: Unknown - Evidence strength: Limited primary - Status: Internal - First reported year: 2026 - Work: Support - Interfaces: Slack - Invocation: Interactive - Entry reviewed: 2026-10-08 ## Purpose ### Summary Embat's internal agent platform began as a support tool for the Product Experience team and now serves the whole company through Slack. A router sends each message to one of more than ten specialist agents, each mapped to a module, task, or tool set. Fact · Reported · Not assessed confidence · `embat-mr-batt--summary` Confidence reason: No explicit confidence assessment is recorded for this claim. Evidence: - Supports · [1] [How We Built an Internal AI Agent Platform at Embat](https://www.embat.io/tech/internal-ai-agent-platform) · [Preserved copy](https://github.com/steel-experiments/internal-agents-map/blob/main/archive/sources/embat-mr-batt-source-1/content.md) · Preserved content.md, lines 20-24 - Supports · [1] [How We Built an Internal AI Agent Platform at Embat](https://www.embat.io/tech/internal-ai-agent-platform) · [Preserved copy](https://github.com/steel-experiments/internal-agents-map/blob/main/archive/sources/embat-mr-batt-source-1/content.md) · Preserved content.md, lines 44-48 Representative workflow: Support question in Slack → routed specialist answer. ## How it works ### Route to a specialist A lightweight step reads the conversation and picks the agent that best fits before any specialist runs Fact · Reported · Not assessed confidence · `embat-mr-batt--primitives-route-to-specialist` Confidence reason: No explicit confidence assessment is recorded for this claim. Evidence: - Supports · [1] [How We Built an Internal AI Agent Platform at Embat](https://www.embat.io/tech/internal-ai-agent-platform) · [Preserved copy](https://github.com/steel-experiments/internal-agents-map/blob/main/archive/sources/embat-mr-batt-source-1/content.md) · Preserved content.md, lines 52-54 ### Ask the user when routing is unsure When the router is not confident, it asks the person to choose; messages that are not support requests get a separate place Fact · Reported · Not assessed confidence · `embat-mr-batt--primitives-ask-user-when-router-unsure` Confidence reason: No explicit confidence assessment is recorded for this claim. Evidence: - Supports · [1] [How We Built an Internal AI Agent Platform at Embat](https://www.embat.io/tech/internal-ai-agent-platform) · [Preserved copy](https://github.com/steel-experiments/internal-agents-map/blob/main/archive/sources/embat-mr-batt-source-1/content.md) · Preserved content.md, lines 54 ## Where people stay involved - **Support question in Slack → routed specialist answer** — Unknown · Level unknown **Reported:** The review boundary for the specialists' answers is not documented; the reported human involvement is the router asking the person to choose when unsure. ### Supervision evidence #### Operating model assessment Unclassified for Support question in Slack → routed specialist answer; human attention boundary: unknown. Inference · Catalog judgment · Unverified confidence · `embat-mr-batt--operating-models-0` Confidence reason: The article documents human involvement only at routing, where an unsure router asks the person to choose; it does not state whether anyone reviews the specialists' answers before use. Qualifications: - Observation date: 2026-07 Evidence: - Supports · [1] [How We Built an Internal AI Agent Platform at Embat](https://www.embat.io/tech/internal-ai-agent-platform) · [Preserved copy](https://github.com/steel-experiments/internal-agents-map/blob/main/archive/sources/embat-mr-batt-source-1/content.md) · Preserved content.md, lines 50-54 ## Implementation details ### Harness Built on the Claude Managed Agents platform, with a lightweight router in front of the specialists; one tool spins up a Google ADK agent instead Fact · Reported · Not assessed confidence · `embat-mr-batt--architecture-harness` Confidence reason: No explicit confidence assessment is recorded for this claim. Evidence: - Supports · [1] [How We Built an Internal AI Agent Platform at Embat](https://www.embat.io/tech/internal-ai-agent-platform) · [Preserved copy](https://github.com/steel-experiments/internal-agents-map/blob/main/archive/sources/embat-mr-batt-source-1/content.md) · Preserved content.md, lines 26-30 - Supports · [1] [How We Built an Internal AI Agent Platform at Embat](https://www.embat.io/tech/internal-ai-agent-platform) · [Preserved copy](https://github.com/steel-experiments/internal-agents-map/blob/main/archive/sources/embat-mr-batt-source-1/content.md) · Preserved content.md, lines 60-64 ### Tool access Embat databases, logs, bank providers, the company's own code repositories, and the Semantic Layer Fact · Reported · Not assessed confidence · `embat-mr-batt--architecture-tool-access` Confidence reason: No explicit confidence assessment is recorded for this claim. Evidence: - Supports · [1] [How We Built an Internal AI Agent Platform at Embat](https://www.embat.io/tech/internal-ai-agent-platform) · [Preserved copy](https://github.com/steel-experiments/internal-agents-map/blob/main/archive/sources/embat-mr-batt-source-1/content.md) · Preserved content.md, lines 22 - Supports · [1] [How We Built an Internal AI Agent Platform at Embat](https://www.embat.io/tech/internal-ai-agent-platform) · [Preserved copy](https://github.com/steel-experiments/internal-agents-map/blob/main/archive/sources/embat-mr-batt-source-1/content.md) · Preserved content.md, lines 48 ### Knowledge Each agent is defined by a system prompt, its allowed tools, and the repositories it may read; routing, Slack plumbing, memory, and tool execution are shared Fact · Reported · Not assessed confidence · `embat-mr-batt--architecture-knowledge` Confidence reason: No explicit confidence assessment is recorded for this claim. Evidence: - Supports · [1] [How We Built an Internal AI Agent Platform at Embat](https://www.embat.io/tech/internal-ai-agent-platform) · [Preserved copy](https://github.com/steel-experiments/internal-agents-map/blob/main/archive/sources/embat-mr-batt-source-1/content.md) · Preserved content.md, lines 46 ### Interfaces slack Fact · Reported · Not assessed confidence · `embat-mr-batt--architecture-interfaces` Confidence reason: No explicit confidence assessment is recorded for this claim. Evidence: - Supports · [1] [How We Built an Internal AI Agent Platform at Embat](https://www.embat.io/tech/internal-ai-agent-platform) · [Preserved copy](https://github.com/steel-experiments/internal-agents-map/blob/main/archive/sources/embat-mr-batt-source-1/content.md) · Preserved content.md, lines 66-68 ### One specialist per domain Each agent maps to a module, task, or tool set, and the team that owns the domain owns the agent behavior Fact · Reported · Not assessed confidence · `embat-mr-batt--primitives-per-domain-specialist` Confidence reason: No explicit confidence assessment is recorded for this claim. Evidence: - Supports · [1] [How We Built an Internal AI Agent Platform at Embat](https://www.embat.io/tech/internal-ai-agent-platform) · [Preserved copy](https://github.com/steel-experiments/internal-agents-map/blob/main/archive/sources/embat-mr-batt-source-1/content.md) · Preserved content.md, lines 40-42 ### Declare an agent in one place An agent is a system prompt, a tool allowlist, and a list of readable repositories; people who cannot code can configure one Fact · Reported · Not assessed confidence · `embat-mr-batt--primitives-configure-agent-declaratively` Confidence reason: No explicit confidence assessment is recorded for this claim. Evidence: - Supports · [1] [How We Built an Internal AI Agent Platform at Embat](https://www.embat.io/tech/internal-ai-agent-platform) · [Preserved copy](https://github.com/steel-experiments/internal-agents-map/blob/main/archive/sources/embat-mr-batt-source-1/content.md) · Preserved content.md, lines 46 ### Implementation coverage - **Model:** unreported — The post relies on the Claude Managed Agents platform and one Google ADK tool, and names no specific model. - **Harness:** reported - **Sandbox:** unreported - **Tool access:** reported - **Knowledge:** reported - **Context management:** unreported - **Credentials:** unreported - **Interfaces:** reported ## Validation and failure handling **Not reported:** The available sources do not describe how the team checks the output. ## Reported observations Observation: Implementation scale · Reported measurement · Specialist agents in Embat's internal platform ### Headline claim Grew from two agents to more than ten specialist agents Metric · Reported · High confidence · `embat-mr-batt--headline-metric` Confidence reason: The article states directly that the platform started with two agents and is now past ten. Qualifications: - The source does not report the denominator of this figure. - Reported by: Embat - Scope: Count of specialist agents on the internal platform since the first two - Observation date: 2026-07-21 Evidence: - Supports · [1] [How We Built an Internal AI Agent Platform at Embat](https://www.embat.io/tech/internal-ai-agent-platform) · [Preserved copy](https://github.com/steel-experiments/internal-agents-map/blob/main/archive/sources/embat-mr-batt-source-1/content.md) · Preserved content.md, lines 48 ## Lessons ### Lesson Embat's single generalist agent needed one large system prompt that no team fully owned, and prompt changes in one domain degraded performance in another; splitting into per-domain specialists removed the interference. Fact · Reported · High confidence · `embat-mr-batt--lessons-learned-generalist-prompt-interference` Confidence reason: The article names the shared-prompt ownership problem and the cross-domain degradation as the reported reason for the split. Evidence: - Supports · [1] [How We Built an Internal AI Agent Platform at Embat](https://www.embat.io/tech/internal-ai-agent-platform) · [Preserved copy](https://github.com/steel-experiments/internal-agents-map/blob/main/archive/sources/embat-mr-batt-source-1/content.md) · Preserved content.md, lines 40-42 ### Lesson With more than ten agents, Embat kept a human in the loop at routing, where an unsure router asks the person to choose instead of picking wrong. Fact · Reported · High confidence · `embat-mr-batt--lessons-learned-router-asks-rather-than-guesses` Confidence reason: The article states the router asks the person to choose when it is not confident. Evidence: - Supports · [1] [How We Built an Internal AI Agent Platform at Embat](https://www.embat.io/tech/internal-ai-agent-platform) · [Preserved copy](https://github.com/steel-experiments/internal-agents-map/blob/main/archive/sources/embat-mr-batt-source-1/content.md) · Preserved content.md, lines 54 ### Lesson Embat chose to borrow the Claude Managed Agents platform rather than build an agent loop, and spent its effort on tools, data access, and answer speed instead. Opinion · Reported · Medium confidence · `embat-mr-batt--lessons-learned-borrow-harness-spend-on-tools` Confidence reason: The team reports its preference to borrow the platform and spend effort on tooling; the post does not compare the choice against a from-scratch build. Evidence: - Supports · [1] [How We Built an Internal AI Agent Platform at Embat](https://www.embat.io/tech/internal-ai-agent-platform) · [Preserved copy](https://github.com/steel-experiments/internal-agents-map/blob/main/archive/sources/embat-mr-batt-source-1/content.md) · Preserved content.md, lines 26-30 ## Question coverage and scope - **purpose:** Reported - **workflow:** Reported - **human involvement:** Reported — The review boundary for the specialists' answers is not documented; the reported human involvement is the router asking the person to choose when unsure. - **implementation:** Reported - **validation:** Not reported - **observations:** Reported - **lessons:** Reported ## Sources 1. [How We Built an Internal AI Agent Platform at Embat](https://www.embat.io/tech/internal-ai-agent-platform) - Engineering blog · First party · Evidence - Original URL: - Published: 2026-07-21 · Accessed: 2026-10-08 · Last verified: 2026-10-08 - Preserved copy in the repository: