The developer sends the staging link on a Thursday afternoon and it looks great. Clean layout, buttons that click, a mobile menu that folds up neatly. Then you open the About page and it’s three paragraphs of lorem ipsum in a nice font, the team photos are gray boxes with a little mountain icon, and the homepage headline still says “Headline Here.”
The site is done. The site is also empty. Most of the launches I’ve watched slip their date didn’t slip because of code or design — they slipped because nobody had written the words or found the pictures yet.
Nobody warns you about this part. You budget for design and build. You don’t budget for the awkward six weeks afterward when someone has to write forty pages of copy and find images that don’t look like a dentist’s waiting room. So let’s talk about the parts that actually get people stuck.
Text: What to Write Yourself, and What Not To
Here’s an unpopular opinion: you should write your own first drafts. Not because you’re a good writer — you might not be — but because you know your product better than any copywriter you’ll hire in the first month. You know why a customer picked you over the cheaper option. You know the one feature people bring up on every sales call. That knowledge is the raw material, and it lives in your head, not in a brief.
So write the ugly version. Open a Google Doc, dump everything you’d say if a friend asked what your company does, and don’t fix the grammar. Then hand that to an editor. Fixing a founder’s messy, true paragraph takes an afternoon. Inventing one from a two-line brief takes a week and usually comes back wrong.
Where does AI fit? Somewhere, but not where most people put it. As a blank-page tool for outlines and rough phrasing, fine. As a machine you paste straight onto the homepage, no. I’ve read a lot of AI-written landing pages and they share a tell: they sound confident about nothing. “Seamless solutions.” “Empowering journeys.” Not one detail a competitor couldn’t also claim. If your copy would sit on a rival’s site with just the logo swapped, it isn’t copy. It’s filler.
Visuals: Where to Get Photos and Graphics Legally
This is where the legal landmines hide, and where I watch small businesses gamble without knowing they’re at a table.
Free stock libraries are the obvious start. Unsplash and Pexels both let you use images commercially, for free, with no credit line required on a normal website download. That part is real and generous. But read one layer deeper. Unsplash’s own terms grant a broad worldwide license, and in the same breath admit they can’t police every upload — so if a photo shows a recognizable face and you drop it into an ad, you’re trusting the photographer’s unverified claim that a model release exists. The free tier comes with no indemnification. A dispute over that image lands on you, not on them.
Pexels reads about the same. Free for commercial use, no credit needed, but their terms spell out that when you’re using a photo to sell something, deciding whether you need a person’s or brand’s permission is your job. They put the disclaimer in writing so there’s no arguing about it later.
So no, I’m not telling you to avoid free stock — I pull from it every week. The thing to get right is matching the source to the stakes. A desk-and-coffee shot behind a blog post nobody will scrutinize? Grab it off Pexels and move on. A grinning “customer” on your pricing page? You’ve just implied a real human vouches for you, and a stranger’s stock face is a lousy way to make that promise. Free versus paid was never the real question. Reputational risk was.
Generative AI images are the newer temptation, and worth a caution. They’re handy for backgrounds and abstract texture. The catch is legal: in the US, the Copyright Office said in its January 2025 report that a purely AI-generated image can’t be copyrighted, since typing prompts doesn’t make you its author, and the D.C. Circuit reached the same conclusion later that year. So that slick AI hero image you generated is one a competitor can lift and you can’t do much about it. For a blurred background texture, shrug. For a logo, that’s the kind of thing that ends up in a lawyer’s email eighteen months from now.
The Content You Can’t Just Google

Here’s the split that separates a brochure site from one that does actual work.
Some content you can round up in an afternoon — team headshots, a stock texture, a rough intro paragraph you’ll rewrite anyway. Other content doesn’t exist anywhere until someone builds it for you. The case study of the project you shipped last quarter. A thirty-second clip of your product doing the thing instead of a paragraph describing the thing. Illustrations drawn to match your actual interface, not a stock marketplace’s guess at what “a tech startup” looks like. A little demo a visitor can click around in.
None of that is downloadable. It has to be made, and making it well is a craft.
This is the point where hiring help stops being a luxury and turns into the reason the site works at all. When what you need is custom visual storytelling — the product film or the bespoke graphic that carries a brand rather than dressing it up — that’s the kind of work the top web design agencies exist to make, precisely because you can’t pull it off a shelf. Outsource the building of things that don’t exist yet. What should exist in the first place — that judgment stays with you, and no vendor can hand it back.
Think of a restaurant. You can buy the plates and the tablecloths and the glasses from a catalog. The dish people drive across town for, you can’t. The catalog stuff is table stakes; the dish is the whole reason anyone shows up.
How Not to Get Stuck: A Working Order for Gathering Content
Most content bottlenecks aren’t a shortage of content. They’re a shortage of sequence. People try to do all of it at once, freeze somewhere in the middle, and the launch quietly slides a month to the right.
Start by digging through what you already own. It’s usually more than you’d guess — old proposals, three decent slides from last year’s pitch deck, that explainer email you’ve now typed out forty times, photos from a shoot everyone forgot about. Get it into one folder before you pay anyone to make something new.
Then rank your pages instead of your file types. The homepage, whichever product or service page does the selling, and the contact page are the ones carrying opening week. The blog you’ll be filling all year is not. I’d push those few pages to genuinely finished before touching anything else — a live site with three strong pages does more for you than a flawless plan for thirty that never ships.
Now take the gaps that are left and be honest about which kind each one is. Some you can handle yourself: founder drafts, the assets from that folder, a two-minute Loom you record on your phone between meetings. Some you just buy: a Pexels image, an icon set, a Canva template you tweak. And some you have to commission — the case studies, the product video, the custom graphics that need a real professional. Where people go wrong is shoving all three into one pile and handing it to whoever quotes the lowest number. A $40 icon pack and a $4,000 product film are not the same purchase, and pretending they are is how you end up with a cheap-looking film.
Set a “good enough to launch” line, out loud, before you start. Your About page can get better after you’re live; nobody’s deciding against you because of it. Your pricing page can’t — get that one right the first time. Then quit polishing the pages that were already fine.
So what’s actually holding up your launch? Nine times out of ten it isn’t the writing or the photos. It’s that nobody’s decided which content is worth paying for, and there’s usually one page people keep avoiding — the one that forces them to say plainly what the company is even for. An empty staging site looks like a content problem. Every so often it’s a much more uncomfortable one.
Building Autonomous AI & Data Agents with Python
Hands-on engineering course with ReAct agents, MCP tool integrations, and LangChain/CrewAI production patterns.