It was the second evening at WordCamp Asia 2026. I was standing in a hotel lobby in Mumbai with three other agency founders. We’d been talking for forty minutes. Not about Gutenberg. Not about block themes. Not about the latest WordPress release.

We were talking about survival.

One of them turned to me and asked, point-blank: “Varun, what’s your backup plan?”

I paused. Not because I didn’t have an answer, but because I’d already been asked that question, in different words, by different people, at least a dozen times since landing. At breakfast. Between sessions. Over beers. In the hallways.

And I realized something that stayed with me the entire trip: 90% of every meaningful conversation at WordCamp Asia 2026 was about the same thing. AI. How to use it. Whether it would replace us. Whether we were already too late.

I’d come to Mumbai with my wife and three of my team members. We had two extra days blocked after the event to explore the city, capture Mumbai, and decompress. But during the conference itself, I let my team loose. Go talk to people. Sit in on conversations. Ask other developers what they’re doing. Absorb the full vibe. I wanted them to feel the shift firsthand, not hear about it secondhand from me.

They felt it. We all did.

This post is my honest download from those conversations. No motivational fluff. No “AI will save us all” optimism. Just what I heard, what I saw, and what I think it means for anyone running a WordPress business right now.

90% of Every Conversation Was About AI

I’ve been to WordCamps for years. The conversations usually follow a pattern: plugin drama, client horror stories, which hosting is fastest, who’s launching what. There’s always one big topic that dominates. Last year it was the WordPress governance drama. Before that, it was full site editing.

This year? AI wiped everything else off the table.

And not in the way you’d expect. It wasn’t on the main stage, not really. The official talks were about WordPress-as-usual topics. But in the hallways, at the dinners, during the after-parties, at every coffee break, the conversation was AI, AI, AI.

Founders were asking each other: Are you using it? How? What tools? How much have you automated? Is it actually working or just hype?

But here’s the thing that hit me hardest: AI is not optional anymore. The conversation has moved past “should we use AI?” That question is dead. The real conversations now are about HOW you’re using it. What flows have you built? What automation is running? How are you segmenting tasks between AI and humans? How extensive is your setup? The founders who are ahead aren’t talking about whether to adopt AI. They’re comparing workflows, sharing MCP configurations, debating which tasks to automate first.

Nobody asks ‘are you using AI?’ anymore. That’s like asking if you use the internet. The question now is: how deep are you? What’s your stack? What did you automate this week?

– Overheard at WordCamp Asia 2026, hallway track

The energy was different from any WordCamp I’ve attended. It wasn’t excitement. It was anxiety mixed with urgency. People weren’t asking about AI because they were thrilled. They were asking because they were scared of being left behind, and because the ones who’d already started were visibly, undeniably ahead. I’ve written before about working 16 hours and still feeling behind. At WordCamp Asia, I realized that feeling has spread to the entire community.

And the ones who were calm? They were the ones who’d already started.

Three Types of Founders I Met

After three days of these conversations, a pattern became obvious. Every founder I talked to fell into one of three categories. And where they fell determined how worried they were.

The Hands-On Founders

These were the ones personally using Claude Code, building MCP servers, writing skills and SOPs for their workflow. They could talk specifics, not just “we use AI” but “I built an MCP that connects our Basecamp to our WordPress staging site and auto-triages support tickets before my team even sees them.”

These founders were calm. Not because they had everything figured out. But because they were IN it. They were learning by doing. They understood what AI could and couldn’t do because they’d hit the walls themselves.

When someone asked them “what’s your backup plan?” they’d say: “I don’t need a backup plan. I need to keep going deeper.”

The Delegators

These founders had one person on their team, usually a senior developer or a tech lead, who “knew AI.” The founder themselves hadn’t touched a terminal. They hadn’t written a prompt longer than two sentences. They’d seen a demo, thought it was cool, and assigned it to someone.

These founders were nervous. And for good reason. They were making decisions about their company’s future based on secondhand information from someone who was also still learning. They couldn’t evaluate what their team was telling them because they didn’t have the hands-on experience to know what was real and what was wishful thinking.

One founder told me: “My lead developer says we’re using AI for everything now.” I asked: “What specifically?” He couldn’t answer.

The Deniers

And then there were the founders still running the same workflow they had in 2023. Same tools. Same processes. Same manual everything. Some of them genuinely didn’t know what had changed. Others had heard about it but thought it was overhyped, or that it would “settle down” and they could catch up later.

I’m not going to pretend this is fine. It’s not. These founders are already behind, and the gap is widening every month. Not because AI is magic, but because the founders in category one are compounding their advantage daily while category three stands still.

The founders who were calm at WordCamp Asia weren’t calm because they had answers. They were calm because they’d already started asking the right questions, with their own hands on the keyboard.

Developers Are Getting Fired. But Not for the Reason You Think

This one came up in almost every conversation with agency owners, and it needs to be said directly: developers are losing their jobs. Right now. At agencies across the WordPress ecosystem.

But not because AI replaced them.

Because they refused to evolve.

One agency founder I spoke with, a 30-person team, told me he let go of four developers in the last six months. Not juniors. Mid-level developers with three to five years of experience. His reason: “I gave them Claude Code access eight months ago. Three developers tripled their output. Four developers never opened the terminal.”

The gap between an AI-augmented developer and a manual one isn’t 10% or 20%. It’s 3x to 5x in raw output, and closer to 10x when you factor in the quality of planning, debugging, and shipping. When one developer on your team is shipping three features a week and another is still manually writing every line of a single feature, the math doesn’t work anymore.

And it’s not about intelligence or talent. Some of the smartest developers I know are the ones resisting the hardest. They’ve built their identity around being the person who writes perfect code by hand. They see AI as a threat to that identity, not a tool to extend it.

I get it. I really do. But the market doesn’t care about our identities. The market cares about output, quality, and speed. And if you’re delivering at 2023 speed in a 2026 world, you’re not being principled. You’re being left behind.

I gave them Claude Code access eight months ago. Three developers tripled their output. Four developers never opened the terminal. I had to let the four go.

– Agency founder, 30-person team, WordCamp Asia 2026

3x to 10x

The output gap between AI-augmented developers and manual-only developers. The gap is widening every month.

If you’re a developer reading this: it’s not too late. But the window is closing. Open a terminal. Install Claude Code. Spend a weekend building something with it. The learning curve isn’t as steep as you think. It’s just unfamiliar. And unfamiliar is not the same as hard.

One Year Changed Everything

Let me put the timeline in perspective, because I think a lot of people don’t realize how fast this moved.

Twelve months ago, April 2025, this is what AI-assisted development looked like for most WordPress developers:

  • ChatGPT was a chat window you copy-pasted code into
  • GitHub Copilot was autocomplete on steroids, helpful but limited
  • Nobody talked about MCP because it barely existed
  • Claude was powerful but you used it through a browser
  • “AI workflow” meant asking ChatGPT to write a function and then manually fixing it

Now look at today, April 2026:

  • MCP servers connect your AI directly to WordPress, Basecamp, Slack, GitHub, databases, everything
  • Claude Code runs in your terminal with full codebase context, reading and writing files, running commands
  • Skills and agents let you codify your expertise into reusable AI workflows
  • Autonomous pipelines can triage support tickets, publish blog posts, audit plugins, run QA, without you watching
  • “AI workflow” now means an AI that understands your entire codebase, follows your coding standards, and ships PRs you can review and merge

This isn’t incremental improvement. This is a phase shift. The kind of change that happens slowly at first and then all at once. And if you weren’t paying attention during the “slowly” phase, the “all at once” feels like getting hit by a truck.

That’s what I saw at WordCamp Asia. A lot of people who just got hit by the truck. They’d been vaguely aware that AI was advancing, but they’d mentally filed it under “future thing to worry about later.” And suddenly it was April 2026 and the developers next to them were doing in a day what used to take a week.

The speed of change broke people’s mental models. They couldn’t comprehend how someone could build an MCP server that auto-processes bug cards from Basecamp, runs PHPCS and PHPStan, writes the fix, tests it in a browser, and posts a verification screenshot, all without human intervention. But I showed them. Because that’s what we do. Every day.

The Backup Plan Panic

The question that haunted every dinner conversation: “What else can we do besides WordPress?”

I heard it all:

  • “Maybe we should learn Astro. Static sites are the future.”
  • “Laravel is more robust. We should pivot to Laravel apps.”
  • “What about building a SaaS? WordPress agencies are a dying model.”
  • “My friend started a print-on-demand business. Maybe we should diversify.”
  • “What about headless WordPress with Next.js?”

I listened to all of this. I understand where it comes from. When the ground shakes, your first instinct is to run somewhere stable. And for a lot of WordPress founders, AI feels like an earthquake. So they’re looking for solid ground: Astro, Laravel, SaaS, anything that feels different from what’s shaking. But as I wrote in a previous piece, AI won’t kill web development agencies. It will kill bad ones. The platform isn’t the problem.

But here’s what I told every single person who asked me about backup plans:

WordPress isn’t the problem. Whether YOU are keeping up, that’s the problem. AI doesn’t care what platform you’re on. If you’re not using it, you’re behind on WordPress AND Astro AND Laravel AND everything else.

Running from WordPress to Astro because of AI is like changing cars because you can’t drive. The issue isn’t the vehicle. It’s the driver. I’ve written about this before: AI code quality is a driver problem, not a car problem. And the same applies to AI adoption.

The founders who are thriving right now aren’t thriving because they picked the right framework. They’re thriving because they learned to use AI deeply, regardless of the stack. A WordPress founder using Claude Code with MCP servers is shipping faster than a Laravel developer who’s still writing everything by hand.

I’m not saying don’t learn Astro or Laravel. Learn whatever makes you better. But don’t learn them as an escape from AI, because AI is everywhere. Learn them as additions to your toolkit while you go deeper on AI in your primary domain.

We spent three months evaluating whether to switch to Laravel. Then we spent one weekend building an MCP server for our WordPress stack. The MCP gave us more leverage than the entire Laravel evaluation.

– Agency founder, WordCamp Asia 2026

I only do WordPress, even though AI lets me do everything. That hasn’t changed. What changed is how deep I go with AI inside WordPress.

The FOMO Is Real. And It’s Justified

Let me be honest about something: the fear of missing out that founders are feeling right now? It’s not irrational. It’s not hype-driven anxiety. It’s a legitimate response to a real shift.

If you’re a WordPress agency founder and you haven’t built your first MCP server yet, you ARE behind. Not terminally behind. You can catch up. But every week you wait, the catch-up gets harder. Not because the technology gets more complex, but because your competitors are compounding their advantage.

Think about it like this: a founder who started using Claude Code six months ago has six months of prompting experience, six months of workflow optimization, six months of building skills and SOPs that make their operation faster. When they wake up tomorrow, they start from that compounded position. You start from zero.

That compound effect is what makes this different from previous technology waves. CSS frameworks came and went. You could learn Tailwind in a weekend and be productive. But AI isn’t a tool you learn once. It’s a practice you develop over time. The expertise is in knowing how to think alongside AI, how to scope for AI, how to review AI output, how to build systems that leverage AI at scale.

That takes time. And time is the one thing you can’t buy back.

The FOMO isn’t about missing a trend. It’s about watching someone do in a day what takes you a week, and knowing they built that advantage while you were deciding whether to start.

– My reflection after three days of hallway conversations

The Opportunity Nobody Is Talking About

Here’s the flip side that almost nobody at WordCamp Asia was talking about, because they were too busy being scared.

If most founders are behind, and they are, that means the ones who move NOW have a massive, disproportionate advantage. This isn’t the kind of opportunity that comes often. It’s a window where effort compounds faster than it normally would, because your competition is still figuring out what’s happening.

I’m going to be specific about what “moving now” looks like. Not abstract advice. Concrete actions you can take this week.

1. Use the Terminal

Stop being afraid of the command line. Claude Code, the most powerful AI coding tool available today, lives in the terminal. Not in a browser. Not in a GUI. In the terminal.

If you’re a founder who’s never opened Terminal on your Mac or WSL on your Windows machine, this is your first step. Not because the terminal is inherently better, but because that’s where the leverage is right now. The best AI tools are CLI-first.

Install Claude Code. Open your project directory. Start a conversation. Ask it to explain your own codebase to you. You’ll learn more about your project in thirty minutes than you’d learn in a week of reading documentation.

2. Create SOPs

Standard Operating Procedures. Document your workflows. Every recurring task your team does (support ticket triage, plugin updates, client onboarding, QA testing, deployment), write it down step by step.

Why? Because once a workflow is documented, AI can execute it. An SOP is essentially a prompt. It tells AI what to do, in what order, with what quality checks. Without SOPs, you’re giving AI vague instructions and getting vague output. With SOPs, you’re giving AI a complete playbook and getting consistent, reliable output.

I have SOPs for everything. Blog publishing (16 steps, never skip one). Plugin QA audits. Bug triage from Basecamp. Support ticket responses. Content calendar planning. Each one is codified, version-controlled, and executable by AI.

3. Build MCPs

Model Context Protocol servers are the connective tissue between AI and your actual tools. An MCP server lets Claude talk directly to WordPress, to your project management tool, to your database, to your hosting dashboard, anything with an API.

This is where the real leverage lives. When Claude can read your support tickets from Zoho Desk, check the related plugin code on GitHub, query the WordPress database on your staging site, and draft a response, all in one conversation, you’re not just using AI. You’re building an AI-powered operation.

Start small. Build one MCP that connects AI to the tool you use most. Then add another. Then another. Each connection multiplies the value of every other connection.

4. Write Skills

Skills are reusable AI instructions that codify your expertise. Think of them as the layered knowledge system I wrote about. Instead of putting everything in one giant prompt, you break your expertise into focused, reusable skills that AI can invoke when needed.

I have skills for WordPress plugin development, bug fixing, security audits, SEO optimization, content publishing, PR reviews, and dozens more. Each one encodes years of my experience into a format AI can follow consistently. When I invoke a skill, AI doesn’t just write code. It writes code the way I would write it, with my standards, my patterns, my quality checks. If you want to understand what skills actually matter in 2026, skills like these are at the top of the list.

You don’t need fifty skills on day one. Start with one. Your most common task. The thing you do every week that follows a pattern. Write it as a skill. Refine it. Then write the next one.

5. Work a Little Harder Than Before

I’m not talking about 80-hour weeks. I’m not talking about burnout. I’m talking about intentionality.

The founders who are winning right now are the ones putting in focused effort to learn new tools, build new workflows, and stay ahead of the curve. Not frantically. Not desperately. Deliberately. An extra hour a day learning Claude Code. A Saturday morning building an MCP server. A Sunday evening writing a skill for your most tedious process.

It’s a good time to work. Not because the world is ending, but because the return on effort right now is unusually high. Every hour you invest in learning AI tools today saves you ten hours of manual work next month. That ratio won’t last forever. Eventually everyone will catch up and the advantage will normalize. But right now, the early movers get an outsized reward.

What I’m Doing About It

I don’t want to just tell you what to do. Let me tell you what I’m actually doing, right now, to stay ahead.

I manage over 100 WordPress plugins. Every single one has a CLAUDE.md file that teaches AI about that plugin’s architecture, conventions, and patterns. When I or anyone on my team opens a plugin with Claude Code, the AI already knows the codebase. It doesn’t start from scratch. It starts from expertise.

I have MCP servers connected to: WordPress (all our sites), Basecamp (project management), Zoho Desk (support tickets), GitHub (code), Slack (team communication), our documentation website, our content calendar, and Google Search Console. Claude can reach into any of these systems, pull data, make changes, and report back.

I have skills for every major workflow: plugin development, bug fixing, QA auditing, content publishing, support triage, security reviews, performance optimization, release packaging. Each skill is a codified version of my best practices, executable by AI on demand.

My support triage runs semi-autonomously. Blog posts go through a 16-step publishing pipeline. Bug cards from Basecamp get auto-processed with code fixes and browser verification screenshots. Plugin releases go through automated QA with WPCS, PHPStan, Playwright browser tests, and accessibility checks.

I’m not saying this to brag. I’m saying this because none of this existed twelve months ago. All of it was built in the last year. If I can build this while running a business, managing a team, and shipping 100+ plugins, you can build your version of it too. You just have to start.

A Message to WordPress Developers

If you’re a developer, not a founder, a developer, and you’re reading this, I want to talk to you directly.

Nobody is coming to save you. Your agency isn’t going to wait for you to be ready. Your clients aren’t going to lower their expectations because you’re uncomfortable with new tools. The market has moved, and it’s moving faster every month.

But here’s the good news: the developers who DO embrace AI are more valuable than ever. Not less. A developer who can use Claude Code to ship features 3x faster, who can build MCP integrations, who can write skills that encode team knowledge, that developer is worth their weight in gold. Every agency founder I talked to at WordCamp Asia said the same thing: “I can’t find enough developers who know how to use AI properly.”

The demand is there. The pay is there. The opportunity is massive. But only if you move.

Don’t wait for your company to train you. Don’t wait for a course. Don’t wait for someone to make it easy. Open the terminal. Install the tools. Break things. Build things. The learning happens in the doing, not in the waiting.

A Message to WordPress Founders

And if you’re a founder: this is on you. Not your team. You.

You don’t have to become a developer. You don’t have to build MCP servers yourself. But you DO need to understand what’s possible. You need to have enough hands-on experience to evaluate what your team tells you. You need to know the difference between “we’re using AI” (vague) and “we’ve built an automated pipeline that does X, Y, and Z” (specific).

Be a category-one founder. The hands-on one. Not because you need to do everything yourself, but because you need to lead from understanding, not from ignorance. Your team can’t take you somewhere you can’t see. And if you can’t see where AI is going, you can’t lead your company there.

The MIT research on AI task performance shows that the gap between “using AI” and “using AI well” is enormous. The “good enough” problem is real, but it’s only a problem if you don’t know what “great” looks like. Get your hands dirty. Learn what great looks like. Then hold your team to that standard.

Stay Positive. Stay Hungry. Start Today.

I came back from WordCamp Asia 2026 with a strange mix of feelings. On one hand, the anxiety in the community was palpable. Real businesses run by real people genuinely worried about their future. That’s heavy.

On the other hand, I came back more optimistic than I’ve been in years. Because the opportunity is genuinely enormous for anyone willing to put in the work. The WordPress ecosystem isn’t dying. It’s transforming. And transformations create more winners than they create losers, as long as you’re willing to transform with them.

The founders who were calm at WordCamp Asia weren’t calm because they had everything figured out. Nobody does. They were calm because they’d started. They’d opened the terminal. They’d built their first MCP. They’d written their first skill. They’d gone through the uncomfortable phase of not knowing what they were doing and come out the other side with something that worked.

That’s all it takes. Start. Today. Not next month. Not after you finish that client project. Not after you take a course. Today.

Open the terminal. Install Claude Code. Point it at your project. Ask it a question. See what happens.

Then do it again tomorrow. And the day after that. And the day after that.

That’s how you build the future. Not with backup plans. Not with panic. Not by running to a different framework. By going deeper where you already are, with tools that are already available, starting right now.

The best time to start was six months ago. The second best time is today. Open the terminal.

I’d love to hear from other founders and developers who were at WordCamp Asia 2026. Did you feel the same shift? Are you building with AI or still on the fence? What’s your first step going to be? Drop a comment or reach out. This conversation is bigger than any one of us, and we’re all figuring it out together.