Jorge, you're taking the first crack at this in-house. Good. That's the right instinct. Here are the eight tests I'd run with you in the room, written plainly so you can run them on your own draft, today, with no one else involved.
Take your homepage draft, your one-liner, your category claim. Run each one through the tests below. If it passes, keep it. If it stalls, you've just found the work. None of this needs me. It needs you to be honest about what a stranger actually sees.
In the first five seconds on your homepage, can a stranger answer three things? Who is this for, what problem do you solve, and what's your point of view. If they can't, you're asking the buyer to do work that he won't do. He'll just leave.
Cover your logo and show the page to someone outside the company. Ask them who it's for. If your words would fit any nearshore engineering shop, they're describing the category, not Jalasoft. The page has to be unmistakably yours with the name hidden.
Read your hero line out loud. Does it name the person you're trying to reach, the PE-backed CTO who carries the delivery risk personally? Or does it describe what you do? "Talent built on trust" is about you. The buyer wants to see himself on the page before he reads a word about you.
Would your ideal buyer recognize his own frustration in your words before you ever mention what you sell? Lead with the problem he feels at 11pm, the one keeping him up. When you open with your solution, you're talking to a mirror, and he can tell.
Think of a deal you lost to a shop with worse engineers than yours. Look at how fast they made that buyer feel understood. That speed is usually the real gap, and it shows up in the words long before it shows up in the engineering.
If you're going to name a category, say it in one sentence: what it is, who it's for, and what it replaces. Then ask three people to repeat it back to you the next day. A category nobody can repeat is still just a phrase you like. The repeat-back is the whole test.
Open ChatGPT or Claude and ask it: "What does Jalasoft do, and who is it for?" Whatever it answers is your real positioning, because that's the answer every buyer researching you is going to get. If the machine gets it wrong, the fix isn't the machine. It's the source it's reading, which is your site.
Go line by line and ask, "so what ... why does this matter to the CTO?" Whatever survives is signal. The rest is parmesan on a weak narrative, sprinkled on to sound impressive. Cut it, and watch how much sharper the page gets.
Every test here asks you to see your own words the way a stranger does. That's the hardest thing to do from the inside, because you know too much. You know what you meant, so you read the meaning into words that don't actually carry it. That's not a Jalasoft problem. It's true of every founder, me on my own work included.
When you've got a draft you're proud of and you want a second set of eyes with no agenda, I'm a phone call away. No pitch, no clock. I want you to win this one whether I'm in the room or not.
Go build it. You've got the truth of what Jalasoft is. This is the part where you tell it.