Post

A tool I always wanted for comparing competitor features with a client's offerings, not just to see where the feature gaps were, but also to understand where the client could catch up or excel. So i've built it. gapnix.com

How does this differ from just using Deep Research in ChatGPT?

Enter a URL, take a few sips of coffee, done?

"deep research" is a feature offered by most platforms now, and does exactly what you describe. Before launching, it asks a few questions to orient the research, and you can come back later with a detailed report along with all the sources it fetched the data from: bsky.app/profile/awak...

Approaching from a different angle with mine. It's more for feature comparison and gaps for product teams as opposed to business intelligence.

I would imagine Dean's approach requires less prompts and supplies faster, more specific intel for SEOs.

Can we agree that no matter how streamlined and nice to use, we're at a point where a lot of SaaS becomes a web app wrapping a complex prompt nicely under the hood? I'm not saying it's a bad idea, or that it shouldn't be done, just that it won't take long before anyone can build their own SaaS

At what timescale though does the early majority of users decide that building their own SaaS, with AI, is easily available to them? Especially as AI usage may steady below what the hype suggests.

I don't think we'll see consumers of SaaS just vibe code their own SaaS, but we'll likely see small companies provide services to develop heavily customized SaaS for just a few users at a fraction of the cost. There will be no moat, as most SaaS will just be web apps wrapping complex prompts nicely

But to have that as a weapon for SMBs where they have historically been either gated, or strapped due to expense? My builds aside, this puts huge power back into their own R&D efforts right?

I'm not talking about LLMs as a fanciful context spinner for text generation, which the vast majority is using it for (me too in some instances), but the breakthrough access to levels of code thanks to LLMs scraping everything from github to StackOverflow and the depraved openweb slurry.

Anybody can do really advanced competitive analysis using ChatGPT Deep Research (or equivalent features available in Claude or Mistral, they all support it). You can ask exactly what you want, it'll search the web and use a wide number of live sources to give you what would have taken hours of work

Here's the funny parallel I can make with the future of SaaS and SEO: Rather than write content or blog posts to try and come up first on Google, you could now vibe code one SaaS per question frequently asked on Google for which nobody offers a product yet

This might be a good strategy to win against AI summaries and "zero-click search": if the AI summary gives the answer taken from your search result, many won't even visit your site. If what you offer is an actual product to solve the problem they're trying to fix, they'll have to click and visit