How to Use AI to Write Blog Posts 5x Faster (My Real Workflow)

I started using speech-to-text in 2014. The tools were rough back then — clunky, needed constant correction, didn’t understand context. But even then, speaking a first draft out loud instead of typing it cut my article writing time roughly in half.

I’m dyslexic. For me, that wasn’t just a productivity hack — it was a lifeline. The gap between having an idea in my head and getting it into written form used to be enormous. Speech-to-text collapsed that gap. If an article normally took me two hours to write, I could get a solid first draft in one hour by speaking it.

That was 2014. Now it’s 2026, and AI has transformed every single step of that workflow. The time savings aren’t 2x anymore. They’re closer to 5x on a good day. And for someone with ADHD who used to avoid content creation specifically because of the boring admin work involved — the change is genuinely life-altering.

This is my real workflow. Not a polished framework I invented for a blog post. The actual tools and steps I use to publish on Techtippr.

“I used to avoid content creation because of the boring admin work. AI took that away. Now I can focus on the part I actually love — the ideas.”

Step 1: Consume Faster to Produce Better Ideas

Before I can write anything good, I need to know things. For a long time I thought I was a slow learner because reading was slow and effortful. Turns out, the format was the problem — not my brain.

Here’s how I consume information now:

• YouTube at 1.5–2x speed — I’ve done this for years. Most people talk slower than we can comprehend. 1.5x is the sweet spot for most content; I push to 2x for familiar topics. You adapt within a few days and normal speed starts feeling painfully slow.
• Speechify for articles — I paste or upload articles into Speechify and it reads them aloud while I do something else — make coffee, walk around, sort my desk. I absorb the content without being glued to a screen. The AI voice quality has improved massively; it no longer sounds robotic.
• AI-curated reading — instead of manually searching for articles to read, I now let AI surface what’s worth my time. I ask Claude or Perplexity to find the most relevant, recent, high-quality pieces on a topic. It filters out noise and brings me signal. I don’t scroll Twitter or RSS feeds anymore.

The result: I consume significantly more useful information per hour than I used to, while spending less time actively sitting and reading. That directly improves what I write — more depth, better angles, fresher takes.

Step 2: Speak Your First Draft — Don’t Type It

This is still the single highest-impact change I ever made. And in 2026, it’s dramatically better than 2014.

Back then, speech-to-text had maybe 85–90% accuracy. You’d speak a paragraph and spend half the time correcting mistakes. Still faster than typing cold, but not by a huge margin.

Now, with tools like WhisperFlow, MacWhisper, or even the built-in dictation on iPhone and Mac (powered by the same Whisper model), accuracy is 98–99%. It understands context, handles punctuation intelligently, and formats as you speak. You talk — it writes. Cleanly.

My process:

1. I open a note or a new document
2. I speak the first draft out loud, stream-of-consciousness, covering the key points I want to make
3. The AI transcribes and lightly formats it in real time
4. I then paste that raw draft into Claude with a simple prompt: “Clean this up, improve the flow, keep my voice and the key points. Don’t add fluff.”
5. Claude returns a polished version in about 10 seconds

What used to take two hours — researching, outlining, typing, editing — now takes 30–40 minutes. The speaking removes the blank-page problem. Once the raw ideas are down, the AI handles structure.

Step 3: Let AI Do the Formatting You Hate

One of the most tedious parts of blog writing — especially for a tech blog — is formatting. Comparison tables. Bullet lists. Headers. Code blocks. Callout boxes. Every one of these takes time to build manually in WordPress.

I don’t build them manually anymore.

I tell AI exactly what I want:

• “Build a comparison table for these 5 apps — columns: price, best feature, platform, free tier.”
• “Turn these bullet points into a proper numbered step-by-step guide.”
• “Add a FAQ section at the bottom — 5 questions a beginner would ask about this topic.”
• “Write a TL;DR summary box — 5 bullet points, plain language, goes at the top.”

Each of these used to take 10–20 minutes. Now it’s a 30-second prompt. I focus on the ideas and the real information. AI handles the scaffolding around it.

For Techtippr specifically, I’ve built a full design system (the tt-* CSS classes you see on every article). Once that existed, I could tell AI to format articles using those components — hero banners, callout boxes, comparison tables, internal link blocks — and it just does it. Every article looks polished and consistent without me touching a line of CSS.

Step 4: AI-Powered Video Editing Workflow

Video used to be the most time-consuming part of my content pipeline. Record → sit through the footage → cut manually → add captions → export → upload. Hours per video.

Here’s what that looks like now:

Text-Based Editing with Descript

Descript transcribes your video automatically and lets you edit it like a Word document. Delete a word in the transcript — that section of video disappears. It automatically removes silence and filler words (um, uh, like) in one click. Bad takes are easy to spot and cut because you’re reading, not scrubbing a timeline.

For someone with ADHD who finds timeline editing deeply tedious, this is the difference between dreading video editing and actually doing it.

Automatic Captions

Captions used to be manual or expensive to outsource. Now every major editing tool — Descript, CapCut, YouTube itself — generates accurate captions automatically. One click, done. Captions also dramatically improve view duration on social media where most people watch without sound.

One-Click Graphics and B-Roll

Adding lower thirds, titles, and simple graphics used to require design skills or templates. AI tools now suggest and insert these contextually — you approve, they render. I spend almost no time on visual design for video anymore.

Step 5: The Automated Publishing Pipeline

This is the part that made the biggest difference for my ADHD. The admin work of publishing — writing YouTube descriptions, choosing tags, filling in metadata, scheduling posts — I genuinely used to avoid content creation partly because of this. Not the creative work. The paperwork.

I built a pipeline that handles it:

1. Drop the finished video into a folder
2. An automation (built with Make/Zapier + AI) detects the new file
3. It transcribes the video automatically
4. AI generates: YouTube title (3 options), description, tags, chapters, and a thumbnail text suggestion
5. The video gets uploaded to YouTube as a draft with all metadata pre-filled
6. I review and hit Publish — usually takes under 5 minutes

For blog posts, a similar pipeline pushes drafted content to WordPress with the right category, tags, and SEO metadata already filled in. I review, adjust if needed, publish.

The creative work is mine. The admin work is automated. That’s the whole game.

The Tools I Actually Use (My Stack)