Ashley Cooper on Lovable, Local-First AI Apps, and Safe Agent Workflows for MSPs

The host interviews Ashley Cooper (COO at Cyber Drain, VP of Community at Rewst) about her use of Lovable and other AI coding tools to build small, local-first, deterministic apps and learn through prompting. Ashley describes receiving a Lovable contributor gift after ranking in the top 0.01% of users, her early failed attempt to rebuild community software, and her shift to rapid, browser-based prototypes like JSON-to-CSV converters, receipt/expense and food trackers, and a webhook-driven PSA time-entry timer. She explains her workflow moving from Lovable scaffolding to GitHub/VS Code with Copilot, plus experimenting with Bolt and tools like OpenClaw, emphasizing trust boundaries, least privilege, and treating agents like employees. They discuss prompt engineering using philosophical mental models, risks of unvetted "vibe-coded" SaaS, and advising MSPs to start with data readiness and education before deploying AI for clients.

This episode is brought to you by Opsleader Pro. A place for MSP owners and managers to get the systems and tools they need to build a stable and growing MSP. Part group coaching, part peer group, everything you need to run a successful MSP.

00:00 Meet Ashley Cooper

01:17 Lovable Gift Surprise

03:01 Early Builds and Lessons

05:25 Hackathon One Shot Apps

09:34 Finding Ideas to Build

11:00 Viral AI Side Projects

14:09 Keeping or Killing Projects

18:22 Local First Automation Wins

19:51 Beyond Lovable Toolchain

24:03 OpenClaw Risks and Power

30:49 AI As Employee

31:18 Stress And Bias

32:02 Context Anxiety

34:52 Prompt Writing Workflow

37:23 Mental Model Prompts

41:44 Deep Instruction Sources

47:20 SaaS Apocalypse Debate

49:53 MSP AI Readiness

53:47 Education And Governance

56:30 Closing Thoughts