AI Won't Replace Web Developers — But It Might Make Them Bad at Their Jobs

I Tried to Vibe Code My Own Marketing Site Last Weekend.

As the owner of a digital agency, I felt like I had to test this myself. If this is what clients are going to be doing with their budgets, I need to understand it firsthand.

One animation — just one — took 50 credits and 50 variations of the same prompt. At the end of it, the animation was still slightly off. And the fix broke two other components.

What That Experience Confirmed

AI is effective at producing something that resembles a website.

But there's a significant difference between something that looks like a website and something that actually reflects a brand — something that makes someone feel something when they land on it.

That gap isn't closed by a better prompt. It comes from years of building things and caring enough to notice why 4px of padding feels different from 6px, or why one animation curve feels confident and another feels clunky.

The Homogenization Problem

Every AI-generated site is starting to look the same. Same layouts. Same hero sections. Same interaction patterns. Same color palettes. Because they're all optimized for the same definition of "good enough."

Our clients don't hire us for good enough. They hire us because they want their digital presence to actually reflect the quality of their work — and the two things are not the same.

Where AI Actually Helps

I'm not dismissing it. I use it. It speeds things up and cuts out a lot of the repetitive work. For boilerplate components, documentation, and debugging, it's genuinely useful.

But the risk I see — especially among junior and mid-level developers — is over-reliance. You start prompting for everything. You stop working through problems yourself. You stop growing in the way you used to, because the answer is always one prompt away.

The Amazon Example

Amazon recently held a company-wide meeting about AI-assisted code changes running without proper guardrails. An AWS AI tool tried to make a minor infrastructure update and ended up deleting and rebuilding the entire environment, causing a 13-hour recovery.

Amazon called it an "extremely limited event" — but the response was immediate: junior and mid-level engineers now need senior approval before anything ships.

The companies that win with AI will be the ones who build the right guardrails around it. Not the ones who move the fastest.

The Real Risk

AI won't take your job as a developer. But it might make you bad at it.

Use it where it helps. Build the guardrails. Don't let it do your thinking.

I went into that weekend half expecting to feel threatened. I came out more confident than ever in what our team at DoodleWeb actually does.

I Worked at Agencies for Years Before Starting DoodleWeb.

And there was one thing that kept bothering me the longer I stayed.

A developer would walk into a meeting and say a task was going to take 7 hours. I'd be sitting there knowing — as a developer myself — that this was a 1-hour job.

But I couldn't say anything. It wasn't my agency.

The client gets charged for 7 hours. The work takes 1. Nobody stops it because that's just how things run at bigger agencies: too many layers, too many approvals, a structure that makes it genuinely hard to deliver the way you actually want to.

That's what frustrated me the most. Not the politics or the process — the fact that clients weren't getting the value they deserved, and I could see it clearly, but couldn't do anything about it.

Why That Actually Matters

This isn't just a billing problem. It's a trust problem.

When clients get overcharged on hours, they become skeptical of every agency. They start questioning estimates, building in budget buffers they shouldn't need, and losing confidence in the relationship before the project even starts.

The best clients — the ones with ambitious projects and real budgets — start looking for reasons to trust you immediately. If your pricing model is built on obscuring hours, you're starting from the wrong foundation.

What I Knew Could Be Different

I had the technical background to know when something was being overcomplicated. I knew what good execution looked like. I knew what things actually cost and how long they actually take.

So I figured: if I have that knowledge, I can build something better.

Not a grand plan. Not a 5-year business strategy document. Just the frustration of watching clients get shortchanged and knowing I could do it differently if I had my own place.

That's what pushed me to start DoodleWeb in 2019.

How That Shapes How We Work Today

When a client comes to us, the first thing we ask isn't "what do you want the website to look like?" It's "what business outcome do you actually need?"

Because clients don't come to us for a website. They come because their business isn't growing the way it should — and they think a website might be the answer. Sometimes it is. Sometimes the website is the smallest part of the problem.

We've walked clients through discovery calls and uncovered intake systems that were losing qualified leads before a single human ever touched them. We've found tracking setups that meant the business had no idea what was actually driving revenue. We've found agencies who built sites on platforms convenient for the agency, not the business.

Every time, the conversation is the same: "We can do what you asked. But I think there's a bigger problem here."

That's the practice we built DoodleWeb around. The website is the tool. The business outcome is the job.

And to this day — that's still what drives how we work.

Most Businesses Think Migrating from Drupal to WordPress Is Just Moving Content Over.

It is not even close to that.

We've run dozens of Drupal to WordPress migrations for organizations across Seattle, the Pacific Northwest, and Canada. The clients who came to us after a failed migration with another agency all had one thing in common: their previous agency skipped half the steps.

The result? Broken pages, missing content, and SEO rankings that took years to build — gone overnight.

Here's the exact process we follow on every migration. No shortcuts.

Step 1: Full Site Audit

Before we touch anything, we document every single page, content type, and custom functionality on the existing Drupal site. This isn't a quick browse. We catalog the content architecture, URL structure, custom modules, user roles, media assets, and any third-party integrations.

If you skip this step, you'll discover halfway through the migration that something critical was missed. By then, it's expensive to fix.

Step 2: Migration Plan Document

We create a full written document covering what is being migrated, how each content type maps to WordPress, what needs to be rebuilt, which URLs need 301 redirects, and which custom functionality needs a WordPress equivalent. This document gets client approval before a single line of code moves.

Step 3: Content Freeze

We ask the client to either freeze all content updates while we work — or keep a running log of every change made to the live Drupal site. Working with a moving target is how content gets lost. This is non-negotiable for us.

Step 4: Fresh Complete Backup

A full backup of the existing Drupal site is taken right before we start moving anything. Not the day before. Right before. If something goes wrong — and in complex migrations, something always has the potential to — this is your recovery point.

Step 5: Theme Rebuild from Scratch

The WordPress theme is rebuilt from scratch using proper best practices — not copied over from the old site. This is where a lot of agencies cut corners. They try to replicate the Drupal design by reusing code, which leads to inconsistent behavior, poor performance, and future maintenance headaches. We start clean and build it the right way.

Step 6: Module and Functionality Review

Every Drupal module is reviewed individually. We need to understand what it actually does before we figure out the WordPress equivalent. Some modules have direct plugin counterparts. Some need custom development. Some turn out to be legacy functionality that no longer serves the business.

Step 7: Full Testing on Staging

Everything is tested on a staging environment before it touches the live site. This includes content accuracy, URL redirects, forms, integrations, search functionality, user roles, mobile responsiveness, and Core Web Vitals performance scores. We do not push to production until staging is confirmed and clean.

Step 8: Controlled Live Deployment

We push live during a low-traffic window, with a rollback plan ready. Post-launch monitoring runs for 48 hours to catch anything that didn't surface in testing. We also submit the updated sitemap to Google Search Console immediately to help preserve rankings.

Why Most Migrations Fail

The migration itself is technically straightforward. The process around it is what protects the business.

Agencies skip steps because it takes more time and they're working on a tight budget. The client sees a lower quote, signs, and ends up paying twice — once for the failed migration and once for someone to fix it.

If you're evaluating a Drupal to WordPress migration, ask any agency you're talking to for their exact process. If they can't walk you through each step clearly, that tells you everything.

Someone Searches for Exactly What You Offer. They Find Your Site, Take One Look, and Leave.

You never get an email. You never get a call. You don't even know they were there.

That's the hidden cost of a weak digital presence — and it's costing B2B companies deals they'll never hear about.

We've worked with 120+ businesses across Seattle and the Pacific Northwest. The pattern is always the same: solid service, good track record, competitive pricing — but a website that makes them look small, unclear, or outdated. Potential clients compare them to a competitor with a more polished site and move on.

The Real Problem Isn't Traffic

Most businesses think their website conversion problem is a traffic problem. They spend more on ads, post more on LinkedIn, and push more visitors to a page that still can't respond to what buyers actually need.

A visitor lands on your page. They have two or three specific questions — about pricing, integrations, whether the service fits their use case. The page doesn't answer those questions directly. It explains the product. It highlights features. It points to a "Book a Demo" button.

So the visitor scrolls for a while, gets partial answers, and leaves. They didn't bounce because of your design. They left because the page couldn't respond to what they were actually thinking.

Five Signs Your Website Is Hurting Your Business

1. Your Design Looks Outdated

Your website is your storefront. If it looks like it was built in 2018, visitors assume your business is too. Credibility disappears before you get a chance to speak.

2. Your Leads Aren't Converting

Unclear calls to action, forms that go nowhere, no automation to capture and nurture leads. If traffic is coming in but nothing is converting, it's a system problem — not a traffic problem.

3. Your Mobile Experience Is Broken

More than 60% of traffic comes from mobile. If your layout falls apart on a phone, you're losing the majority of your visitors before they ever read your value proposition.

4. Updates Break Things

You change one thing and something else stops working. That's a sign the site wasn't built correctly from the start. It only gets worse the longer you leave it.

5. You Have No Idea What Visitors Are Doing

No heat maps, no session recordings, no click tracking. If you can't see what visitors are doing, you can't fix what's not working.

What's Actually Happening on Your Site

We worked with a company that had a solid offering and real buyer intent coming to their site. But when we looked at their visitor recordings in Microsoft Clarity, the pattern was clear: visitors landed, scrolled for about 90 seconds, couldn't find what they needed, and left.

The homepage didn't reflect what the company actually did anymore. The team had doubled. The offering had expanded. But the site still reflected the version of the business that existed two years ago.

Their sales team was spending half of every discovery call explaining things the website should have already communicated. One prospect told them: "I almost didn't reach out because your site made me think you were too small."

That's a deal nearly lost before the first call.

What Actually Converts

There are only two things you can actually influence:

Content depth — Make your pages detailed enough that determined visitors eventually piece together their own answers. Most quit before they get there.

Experience responsiveness — Make the page answer their question in real time and move them to the next step. This is what actually converts.

The companies winning right now aren't necessarily the biggest or best-funded. They're the ones whose websites can hold a conversation, answer the question, and move the buyer forward.

What To Do Today

Pull your last 10 closed deals. Ask your sales team: "Did our website help close this deal or not?" If they can't give you a clear answer, your website isn't earning its keep.

The difference between thriving and struggling isn't traffic. It's knowing which parts of your website are actually connected to revenue.