First-party coverage protects your organisation from losses you suffer directly when an AI agent causes or contributes to an incident. Rather than covering claims made against you by others, this coverage pays for the real operational and financial damage that hits your own balance sheet — whether that’s rebuilding a corrupted model, absorbing lost revenue, or covering unexpected infrastructure bills.Documentation Index
Fetch the complete documentation index at: https://redberrylabs.com/docs/llms.txt
Use this file to discover all available pages before exploring further.
Data and model restoration & retraining
When an AI incident corrupts training data or forces you to roll back a model, you face significant costs to restore datasets, run retraining pipelines, and validate outputs before returning to production. This coverage reimburses those direct restoration and retraining expenses. Example: A supply-chain disruption corrupts the historical demand data used to train your inventory forecasting agent. You need to re-source clean data, retrain the model from scratch, and run weeks of validation before redeployment. First-party coverage picks up those costs.Business interruption
Revenue losses caused by an AI-driven operational failure — not just the cost to fix it — are covered here. If a misbehaving agent triggers a cascade of downstream penalties, missed SLAs, or halted workflows, this coverage addresses the lost income during the disruption period. Example: Your AI logistics agent misroutes a large inventory shipment. Fulfilment delays trigger cascading contractual penalties and cancelled orders. Business interruption coverage compensates for the revenue lost while operations are restored.Incident response & regulatory costs
Responding to an AI incident often requires forensic investigators, legal counsel, and formal notifications to regulators or affected parties. These costs accumulate quickly and sit entirely outside normal operating budgets. This coverage pays for the investigation, legal fees, and mandatory notification obligations. Example: A model update introduces unexpected behaviour in a regulated workflow. You engage external forensic consultants to scope the incident, retain legal counsel to assess your disclosure obligations, and notify the relevant regulator. Incident response coverage handles those costs.Runaway usage & infrastructure overage
Security events — particularly denial-of-service patterns targeting AI endpoints — can generate infrastructure and API costs that far exceed any planned budget. This coverage addresses unexpected spikes in compute, token, or API spend that result from such events. Example: A botnet floods your public AI endpoint with long-context queries, driving your LLM API bill up sharply before your team can respond and block the traffic. Runaway usage coverage reimburses the overage costs caused by the attack.Incorrect funds transfer
Financial losses that arise directly from an AI agent making an erroneous payment decision — approving a transaction it should have flagged, misrouting funds, or processing an order at the wrong value — are covered here. Example: Your AI procurement agent autonomously approves a bulk GPU purchase order totalling $4.45 million, far outside intended approval thresholds. Incorrect funds transfer coverage addresses the financial loss from that erroneous transaction.To file a first-party claim, log in to app.redberrylabs.com and navigate to Claims. Reference the specific agent involved — coverage is issued per agent, so selecting the correct agent is required to open a claim against the right policy.