The problem

Some questions are simple: "When does the O&M contract expire at Sunny Solar?" Squirrel answers in seconds.

But others are big: "Review all preventive maintenance reports from the past 6 months across my French assets and summarize what happened." That kind of task involves pulling dozens of documents, cross-referencing entities, and synthesizing findings. In a single chat thread, it bogs down, and you're stuck waiting.

How sub-agents work

When Squirrel recognizes a task that's too large or multi-step for a single conversation turn, it can now delegate parts of that task to sub-agents. These are specialized workers that run in the background.

You keep talking. While a sub-agent processes your bulk request, you can continue asking Squirrel other questions. The sub-agent works independently and reports back when it's done.

Scoped and focused. Each sub-agent gets a clear, bounded task. "Summarize defects for assets A, B, and C" might become three parallel sub-agents, each handling one asset. Results are aggregated and delivered together.

Use cases

• Cross-portfolio comparisons. "Compare O&M availability guarantees across all my wind assets." Sub-agents handle each asset in parallel.
• Bulk document review. "Review all statutory inspection reports from Q1 and flag any new defects." Each report gets its own sub-agent.
• Compliance setup. "Based on the contracts for this new asset, suggest all compliance items I should create." The sub-agent reads through multiple documents and proposes a structured list.
• Data export prep. "Prepare a summary of all contract key terms for the board meeting." The sub-agent compiles, you review.

Benefits

• No more waiting. Complex tasks don't block your workflow. Keep working while the heavy lifting happens.
• Better results on big asks. Breaking a large task into focused sub-tasks produces more accurate and thorough output than forcing everything through one thread.
• Transparency. You see what sub-agents are working on and get notified when results are ready.