This episode of the Jukebox Podcast (WP Tavern) features Matt Schwartz, founder of Inspry, an Atlanta WordPress and WooCommerce agency, and maker of the WordPress testing product CheckView. Matt has run an agency since 2011 and has recently focused much of his attention on how AI can change the way agencies operate.
Big picture
Rather than the common conversation about AI generating sites or content, Matt argues the biggest opportunity for WordPress agencies is using AI behind the scenes: to improve processes, SOPs, client communications, and internal operations. Execution — building a basic brochure site — is becoming a commodity. The real value agencies must prove is in their process, strategy, and how they help clients solve business problems. AI is useful because it can scale and speed up many of those non‑visible tasks.
Transparency and ethics
Matt’s ethical line: if AI is changing the deliverable a client receives, tell them. For internal workflows that speed up the agency, disclosure isn’t typically necessary unless you’re helping a client redesign their internal processes with AI. He stresses guardrails and human oversight — especially on client-facing work — to avoid mistakes that can damage reputation.
Why now?
AI adoption is accelerating. Clients know the possibilities and will have higher expectations for speed, touchpoints, and intelligence. Agencies that ignore AI risk falling behind in client experience and efficiency. But Matt is clear: you should be deliberate about adding AI, not haphazard.
An AI vision document
Matt recommends creating an AI vision document — a written plan that maps your agency’s processes and flags where AI makes sense. Use time-tracking, ticket history, and other data to find repetitive, high-effort tasks where AI can add the most value. The document should also list places where AI is off‑limits or where human judgment is required. Writing this down gives the team clarity and protects the agency from inconsistent or risky implementations.
AI as a core service
If your agency learns to apply AI well, that competency itself can become a sellable service. Many clients won’t want to hear the label “AI”; they want results (cost savings, faster workflows, better data). Agencies can productize automation and internal tools — lightweight apps or integrations that used to be too expensive — and sell them as services or recurring maintenance. Tools like automation platforms (e.g., n8n/Make) let you stitch together processes and deliver measurable operational value.
Data and marketing insights
AI’s strength in pattern recognition allows agencies to offer analysis that was previously too costly. Agencies can ingest client support tickets, analytics, and CRM data to produce audience segments, persona insights, or prioritized product/market recommendations. That enables service offerings that were out of reach for many small clients before.
Practical uses inside agency operations
– Meeting summaries and proposals: AI can transcribe meetings, summarise key points, and draft more tailored proposals and SOWs by combining call notes, the client’s website, and previous proposals. This saves time and improves proposal quality while still requiring human review.
– Support workflows: Rather than replacing human support agents, AI can provide richer context for every ticket by pulling logs, prior tickets, server info, and suggested fixes. This helps humans resolve issues faster, raising throughput without degrading quality. Matt warns agencies to tread carefully: poor automation at a customer’s moment of pain can cause frustration, and some users will always prefer human contact.
– Debugging and site management: AI agents connected to a site’s APIs and logs can speed up troubleshooting (finding the likely culprit, suggesting fixes). For one‑off, niche problems, AI can save many developer hours. Many page builders and platforms are also adding agent-friendly syntaxes so AI can edit content blocks — but Matt insists on human review of edits and changes.
Human in the loop
Across use cases the rule of thumb is: use AI to augment speed and pattern recognition, but keep a human in the final review loop depending on risk. Agencies need policies for when AI output is acceptable and when a human must intervene. Relying blindly on AI is risky; verification and accountability remain essential.
WordPress as the foundation
WordPress itself is becoming an effective scaffold for AI-driven workflows. Instead of locking functionality inside WordPress, the core approach increasingly allows tools and agents to communicate with WordPress as the platform. That means many users may never need to visit the admin UI again — they’ll instruct AI agents to make changes while WordPress remains the reliable infrastructure layer beneath.
Wrap up
This episode covers eight of Matt’s sixteen notes and focuses on practical, process-driven AI adoption for agencies: be intentional, write an AI vision, protect customer experience, and keep humans responsible for critical judgments. The conversation continues in Part 2, where Matt will explore additional examples, risks, and tactics for integrating AI across sales, delivery, QA, and more.