Nathan Wrigley welcomes listeners to the Jukebox Podcast from WP Tavern for a conversation with Matt Schwartz, founder of Inspry, an Atlanta-based WordPress and WooCommerce agency (founded 2011). Matt also builds CheckView, a WordPress testing product. The episode focuses less on AI generating sites or content and more on how AI can be embedded in agency workflows, processes, and services to deliver more value behind the scenes.
From accidental beginnings to agency life
Matt explains he stumbled into web design as a kid and then chose to stay in it. Like many freelancers and small agency owners, he learned processes by doing—crafting SOPs, pricing, emails and workflows without formal guidance. That iterative, DIY approach is common, and it’s the area where AI can produce the most meaningful improvements.
The bigger shift: execution as commodity
Both hosts agree: simple execution (e.g., brochure sites) is increasingly commodified. Tools and AI reduce the time and cost of building basic websites, so the differentiator becomes the agency’s process, client touchpoints, strategy and the value it brings. Because clients are aware of what AI can do, expectations are rising. Agencies that use AI thoughtfully can improve processes and compete better, but indiscriminate adoption brings risks.
Transparency: when to show AI to clients
Matt’s stance: disclose AI when it changes the client’s product. For internal workflows and tools that don’t affect delivered work, disclosure is less necessary. If an agency helps a client implement AI into their own operations, then transparency and parallel setups make sense. Ethically, he prefers being upfront when AI influences delivered outputs.
Why now?
The pace of AI development is rapid and the gap is widening between agencies that use AI and those that don’t. This doesn’t mean jumping in blindly — it means evaluating where AI makes sense for your agency. Because basic execution is easier and cheaper, agencies should refocus on strategy and processes where they can offer unique value. If competitors use AI to create better client touchpoints and faster responses, your agency may need to respond to maintain an edge.
An AI vision document
Matt recommends creating an AI vision document: a deliberate, written plan that outlines where AI will and won’t be used in the agency. It should map processes, identify repetitive tasks ripe for automation, and define guardrails for areas requiring human judgement. This prevents haphazard “YOLO” adoption and keeps teams aligned on boundaries and responsibilities. You can even use AI to analyze your time-tracking data (e.g., from tools like Everhour) to spot patterns and candidate processes for automation.
AI as a core service offering
Rather than simply applying AI internally, agencies can productize AI expertise as a service for clients. Many small businesses don’t want the label “AI”; they want improved workflows, cost savings, or revenue-driving automation. Agencies can offer automation, internal web apps, and SOP-driven solutions — sometimes building custom tools that would previously have been cost-prohibitive. Be cautious about security and proper engineering practices, and weigh whether to build in-house versus integrate existing SaaS.
Using data and AI-driven research
AI excels at synthesizing large datasets. Agencies can provide new services by analyzing support tickets, user behavior, geography-based market insights, or competitor data and presenting affordable, actionable recommendations. This opens up offerings that were previously too expensive for small clients.
AI inside agency operations (low-hanging fruit)
Many practical wins are already accessible: automated meeting notes and summaries, drafting proposals and SOWs, and producing initial scope suggestions based on discovery calls and past proposals. These outputs should be reviewed by humans, but they dramatically reduce time to produce polished client-facing documents and allow agencies to propose more thoughtful solutions faster.
AI for support workflows
Agencies can feed help tickets and site data into automation platforms (n8n, Make, etc.) so that AI gathers context—error logs, prior tickets, relevant PM items—and surfaces likely fixes or suggested next steps for support staff. The aim is not to replace human agents but to enrich them with fast, contextual suggestions so issues are resolved more quickly. This is a sensitive area: using AI as a frontline, fully automated responder risks frustrating customers, especially those who want a human touch or have accessibility needs. Guardrails and escalation paths to humans are essential.
AI-assisted debugging and WordPress management
Connecting AI agents to WordPress sites (with careful security and permissions) can reduce time spent on one-off debugging tasks. AI can analyze REST APIs, error logs, plugin conflicts, and server data to propose likely causes and remediation steps. Page builder and block-aware tools are emerging so agents can make targeted edits, but Matt stresses the need for human review. He notes a cultural tendency to forgive AI mistakes more readily than a human developer’s errors; human oversight remains crucial until trust and reliability are proven.
WordPress as the scaffolding for AI-driven workflows
Nathan and Matt observe that WordPress is evolving to be AI-ready: richer APIs, block-level semantics, and integrations that let AI agents interact with sites as infrastructure while humans retain oversight. In many cases WordPress will act as the platform while AI tools, combined with human review, manage content, edits, and operations.
Part two coming
With half their show notes covered, the hosts decide to split the conversation into a two-part series. Part one sets the stage: why agencies should consider AI now, how to plan its adoption with an AI vision document, practical internal uses, ethical considerations, and new client offerings. Part two will continue the deep dive into remaining topics, risks and examples.
If you run a WordPress agency or are curious about how agencies are adapting to AI, this episode offers pragmatic guidance: be intentional, document where AI belongs, keep humans in the loop at critical moments, and explore productized AI services that add measurable value to clients.