OpenClaw Sales Demo Automation Assistant

Problem Addressed

In sales demos, the hard part is not describing automation; it is showing a runnable, verifiable path that covers data movement, repetitive entry, manual reporting, missed exceptions, and opaque workflow status. These pain points often span ERP, CRM, OA, databases, and Excel, and a single script rarely handles orchestration, RPA-like page actions, scheduled triggers, multi-channel alerting, and deliverable generation. This skill is aimed at demos and pre-sales validation, converging scattered pain points into reusable demo paths and producing workflow diagrams, execution logs, sample reports, and alert samples.

How It Works

• Requirement mapping: uses scenario questions to classify feedback into data movement, repetitive form entry, manual reporting, missed exceptions, workflow visibility, and template bootstrapping, then outputs a customer profile and recommended demo path.
• Workflow assembly: selects an industry template or minimal viable flow and configures data sources, field mapping, conditional branches, output actions, and success/failure notifications. Sources can include database queries, API calls, file reads, and Excel imports.
• Execution and display: triggers the demo with scripts/run_demo.py, showing progress, step-level input/output, timing, and data flow. Results can be emitted as HTML, PDF, Excel, structured JSON, or Markdown.
• Exception demos: simulates failed source connections, dirty data, threshold breaches, API 500 errors, and timeouts, then demonstrates WeCom, email, SMS, and Webhook alerting paths, including retry, degradation, and fallback strategies.
• Template recommendation: ranks the top 3 templates by fit, provides customization notes, a POC → pilot → rollout plan, and an ROI estimate based on saved labor hours and cost.

Boundaries and Caveats

The skill is demo-oriented and fits pre-sales validation, POCs, training, and solution walkthroughs; it is not a replacement for a production RPA platform. Connecting to ERP, CRM, OA, or internal databases still requires permission, credential, network, and data-masking controls. RPA-style page actions can break when layouts change, when sessions expire, or when verification challenges appear. Alerting and self-healing patterns are better treated as communication and recovery demos; production use should add audit logging, retry limits, idempotency, and escalation paths.