> ## Documentation Index > Fetch the complete documentation index at: https://help.daylit.com/llms.txt > Use this file to discover all available pages before exploring further. # Customer AI signals and collection insights explained > Daylit's AI generates risk signals, conversation summaries, and collection strategy plans for each customer to guide your AR team's actions. ## Purpose **Jobin (AR manager)** at **Acme Corporation** uses AI signals to prioritize outreach without reading months of email history. Signals flag worsening payment behavior early and attach **Conversation Summary** and **AI recommended strategy** so **Jobin (AR manager)** knows exactly what to write in the next follow-up. ## Acme Corporation — priority risky accounts These five customers appear throughout the docs as **Jobin (AR manager)**'s typical Monday triage list: | Customer | Signal | What it means | | ------------------------ | --------------------- | ----------------------------------------------------------------------------------------------------------------------------- | | Apex Solutions | Credit risk | Liquidity crisis on the customer side — **Jobin (AR manager)** escalates to **Maria (VP Finance)** before aggressive dunning. | | Mesa Valley Construction | Promise to pay active | Broken payment promise — **Jobin (AR manager)** sends a firm follow-up via **Next Step**. | | Redwood Chemical Supply | Promise to pay active | Broken payment promise — **Jobin (AR manager)** may **Rewrite** the draft or **Save draft** until after a call. | | Orion Electric Supply | Disputed pricing | Active dispute driving short payments — **Jobin (AR manager)** pauses program touchpoints and applies the **Disputed** label. | | Sterling Logistics | Contract dispute | Adversarial tone, legal involved — **Jobin (AR manager)** routes internally to **Maria (VP Finance)**; no automated outreach. | AI signals transform a reactive collections process into a proactive one. Instead of waiting for an invoice to hit 60 days past due, the AI flags worsening payment behavior immediately, providing drafted emails and call scripts that **Jobin (AR manager)** can execute in one click. ## What AI signals are An AI signal is a pattern detected on a customer account that warrants attention. Signals are generated by analyzing a combination of payment behavior data and communication history. Common signal types include: * **Payment behavior worsening** — The customer is paying materially slower than their own historical baseline. * **Severe or chronic late pattern** — The customer has a persistent history of paying late. * **Accelerating deterioration** — Lateness has worsened month over month. * **Balance growing** — Open balance is increasing while payment velocity is flat or declining. * **Ghosting** — The customer is unresponsive to outreach despite having an overdue balance. * **Churn risk** — Open balance has dropped sharply, the customer has gone quiet, and remaining balance is deeply aged. * **Frequent short payer** — The customer routinely pays less than the invoiced amount. ## Signal severity levels Every signal is assigned a severity to help your team prioritize: | Severity | Meaning | | ----------- | ----------------------------------------------------------------------------- | | **Risky** | Immediate attention required. The AI has detected a significant risk pattern. | | **Watch** | Action recommended soon. The pattern is concerning but not yet critical. | | **Healthy** | No collection action needed. The account is in good standing. | ## Conversation summary The AI reads every email and logged communication in a customer's thread and generates a concise summary of where things stand. The summary answers: * What is the current status of the collection conversation? * Who needs to take the next action? * Is there a commitment date or a predicted payment date in the conversation? ## Collection strategy plan For each customer, the AI generates a recommended next step — either send an email, make a call, or escalate internally. The strategy plan includes: * **Recommended action** — The specific action type (email, call, or internal follow-up). * **Reasoning** — A brief explanation of why this action is recommended. * **Urgency** — A score and bucket (Critical, Today, This Week, Upcoming) indicating how time-sensitive the action is. * **Email draft or call script** — A pre-written draft with the appropriate tone and template. You can act directly on the strategy plan from the customer detail page or from **Inbox**.