Nate B Jones demonstrates how to use AI to extract company strategy and career insights from job postings
A solo presentation by Nate B Jones of AI News & Strategy Daily on using AI tools to analyze job postings for strategic intelligence.
Summary
Nate B Jones walks through a practical demonstration of using large language models to analyze publicly available job postings and extract strategic insights that would previously have required enormous manual effort. Using Anthropic as a live example, he shows three different AI-generated analyses — built with a custom Lovable app, Perplexity, and ChatGPT — each surfacing different angles on the same company: product strategy, B2B sales posture, hiring gaps, inferred weaknesses, and career opportunities. He argues that job postings represent a new class of data that was previously considered too voluminous and too tedious to analyze at scale, but which LLMs have now made trivially accessible. He extends this argument beyond job postings to a broader point about data that was once considered "waste data" — including publicly posted selfies — now being analytically useful in ways that carry both opportunity and risk.
Key Takeaways
FULL TRANSCRIPT
Introduction: Reading job postings as a strategic art
Nate B Jones: In the next ten minutes together, we are going to crack the code on reading job postings. It's a lost art, and it's one that we can absolutely transform in the era of AI. And I don't just mean for job seekers, although that's obviously a huge benefit. You can also infer company strategy, B2B sales approaches, and all kinds of other things just from reading job postings.
A year ago, I wouldn't have been able to recommend this. But now, with where LLMs are at, I can actually give you three different examples in the next few minutes that give you a comprehensive approach on how to read an entire company's strategy from just a set of job postings. And yes, I'll be dropping the prompts and everything else in the post, so you'll be able to follow up and do it yourself.
So what are we talking about when we say a job search strategy informed by job postings? In the past, we would have said, "Hey Nate, go and get a thousand job postings — or the last hundred from the last ninety days — and do this yourself. Conduct a manual review. I want you to categorize everything. I want you to identify commonalities, spot weak points, notice what they didn't post. Now look at how the products they offer compare to the job postings." You see how it goes on and on. Not anymore. You can get all of that done in just a prompt. In fact, you can get more done than you could before, because you not only have the volume game, which you can play with LLMs, you also have the strategy game. You can give the LLM a strategic prompt, tell it to reason and infer in a particular way over a set of job postings that it searches, and it will do that and come back and give you a view.
Sometimes just showing it is way easier. So let me show you an actual response that I built about a real job posting situation at Anthropic, the major AI company.
Live demo: Analyzing Anthropic's job postings with a custom Lovable app
All right. So I built this handy little app in Lovable just to showcase what you can do with it. Don't worry about the initial fields here. If you're an engineer, it's really easy to put in an API key and use this yourself — I'll be sharing it. And if you're not, you can follow along, and I'm going to give you some prompts that you can use in other search engines. So if you don't know what an API is, you don't have to care.
But look at what you get. This is analysis results generated today at 9:24 when I'm recording this, and it gives you so many different components to look at. It infers a product strategy — doubling down on their core AI build — and suggests that they have a lack of fresh platform engineering hires, which would indicate that they're focused more on scaling existing tech right now. That aligns with what we see from Anthropic's recent moves. It seems like a solid insight. Meanwhile, alignment science and model welfare roles indicate a willingness to tackle unsolved safety and ethics problems. Again, that aligns tightly with what we see from Anthropic.
We go down to inferred B2B sales approach. It's calling out that this signals a push for rapid enterprise adoption based on startup account management and B2B marketing. And there's little evidence of dedicated sales engineering. One of the things that's really interesting is you can start to infer a B2B strategy from this. You can look at this and say they don't have dedicated sales engineers yet. They don't have post-sales technical support. They're very early in their B2B startup account story. There is an opportunity to come in and offer solutions for a sales team that is probably under stress right now. And you can read that from the job postings. You can infer that.
Now, what if you're a job seeker? What does that look like? Well, they don't have internships or entry-level roles posted right now, and they have very few roles for platform engineering. What's interesting about that is that they are essentially setting themselves up for a potential technical debt risk as they scale. And that is indeed what we see in some of the recent outages and the work that Anthropic has done — to their credit — to talk about why the outages occur. They are struggling to keep pace with scaling demand and they haven't yet invested in platform engineering. So another insight here, if you're putting another lens on this, is that Anthropic may need an additional capital injection in order to start to scale some of these platform pieces out.
It has inferred cultural insights, which seem fair for what they're worth, but it's very easy to get them. It's trivial to tweak the prompt and get what you want. Inferred company weaknesses: it calls out platform engineering, it calls out a miss on PM, QA, and customer support. What's interesting here is that this underlines the research bones of the company — where the company came from. This will not always show up. These are individual insights that you get per company.
Now, this is the part I love the most. You can actually see why the model did this. It will give you a table and it will say, "This is what I read. This is the link to it. This is the reasoning, and this is the claim I'm confirming here." And you can see that it's all recent stuff — this is not old postings it's working from. Is it perfect? No. Does it underline how much you can get out of just looking at job postings? Yes, it does.
Running the same analysis in Perplexity and ChatGPT
Now, I want to give you a couple of other ways to look at this. This is not just something where you have to use a custom Lovable app. You can do it directly in ChatGPT. You can do it in a search engine like Perplexity. I have prompts for that. The key takeaway is that the quality of this assessment depends on your ability to ask very clearly for exactly what is important to you. And that's why I built different versions — a version for job seekers that zeroes in on available roles and what you can get, and a version for folks who are looking for competitive intelligence.
That's not something we've talked about yet in this video, but as someone who has had to run competitive intelligence in the past, this would have been a lifesaver a year ago. It would have been huge. Because all you have to do is plug this thing in and you get a full competitive readout on your competitor just based on their job listings that they have publicly shared. You're not doing anything inappropriate. You're just looking at their job listings.
Here it is. Same exact company — Anthropic. It runs the query. You can actually see the query here — I'll be sharing it in the post. It thinks for a bit and it goes and runs it. It gives you a sense of what it looked at and the signals it pulled out. This is a slightly different order from what you saw in the Lovable app. This is an order that emphasizes proving how it got there. So these are the grounds or the inputs it's using, and it wants to show you those first. If you want to just scroll, you can scroll down to insights and see where they're investing.
This one is absolutely aiming at the career side, so it brings out more of the career piece than I have in the Lovable app. Although Lovable makes it really trivial to remix these — so when I publish this, anybody is going to be able to just remix it and make it what they want. You can make it a career-focused one really easily. I'll include this prompt so the career folks are going to have plenty to work with.
It has a Seattle office that's growing rapidly and an NYC hub. It talks about the comp, which is of course insane AI comp. And then it gives you the receipts to show you how it's thinking about it. It's also talking about competition, which isn't surprising, but it's nice that it pulled that out. It calls out automation risk. It calls out less emphasis on consumer features, so you sort of know where they're at — which aligns with what the Lovable prompt found. So this is like a lens on the same company from a different camera angle, where you're looking just at careers and obsessing about it. And as you can see, it's not a fancy web page, but it's lots of information you can use. You don't have to have an API key or anything. You just run the prompt. And by the way, if you don't have Perplexity or don't use it, ChatGPT has its own search engine and will also run this prompt.
The "company radar" prompt for product managers and analysts
Let me give you one more peek. I love this one. This is a company radar that's more for the product manager or someone who wants to do an overall analysis, and I think it's really cool. It gives you a sense of what's in the box.
All right. So it's going to go through and look at all the signals, prove its way forward, and then get into product strategy. It's going to talk about how Anthropic is investing in Claude Code, what MCP looks like as far as a moat goes — which is something I've been calling out as an engineering move for them to build that ecosystem. This one talks a little bit about how they're doing B2B sales. And this one does catch a sales position piece around healthcare and financial services, which the Lovable prompt didn't get.
If you're looking at reconciling that, what I'd suggest is you pull all three and then start to hybridize and harmonize them a little bit and pull out the specific insights you're looking for. It's almost like getting 3D vision — you get different perspectives on the same company, and they're roughly aligned, but you get different nuances that come out.
You have a callout on how engineers work, which I really love. You have a callout on anti-hierarchy signals, which is another great one. And you have some interesting inferred weaknesses: Are there too many engineering manager positions with no teams built? Is there euro chaos because they're acqui-hiring teams? Scaling fractures — this feels like it's really big and really fast. TPU dependency, which is frankly a really interesting piece of intelligence. I think this is a phenomenal overall perspective on the company. If I were in any kind of competitive intelligence role, this would be really exciting.
The broader lesson: LLMs are unlocking formerly "trivial" data
One of the things I want you to take away as you look through this is that this is not hard to do. I'm going to share the prompts, I'm going to share how I worked through it, but what you should be thinking is: where is there data that I want to get a hold of that would previously have just been really hard to analyze? How can I get a hold of that data and make use of it?
And that is one of the larger lessons I want to call out here. We are in a world where there is an entire new class of data that was previously considered trivial — data that wasn't worth hiding, wasn't worth securing, because nobody had the time to analyze it. It's now open season. This data is now available for analysis, available for strategic understanding. If you're an investor trying to invest in a company, why wouldn't you run a query like this on open job postings and cross-check that against what the company's principals are telling you? Of course you would. That makes perfect sense.
So this is not just something that job seekers are interested in. This is something where if you need the full picture on a company, it is now easy to get. And I want you to ask yourself: what other classes of data are like that? What other classes of data out there have been trivial for a long time, and we're now thinking maybe that's not trivial anymore?
A privacy warning: location inference from public photos
I'm going to give you another example, and this is actually a safety tip in the age of AI. Think about when you last posted a publicly available selfie outside. The reason I say that is because with the advent of reasoning models — especially the ChatGPT image recognition models — they are extremely good at knowing where a photograph was taken in the world. So if you have your Instagram feed set to public and you're taking a bunch of selfies outside, even if you don't reveal your location, your location can be inferred from that information.
There are other examples as well, but I think that gives you a picture. We are entering a world where LLMs are making a whole new class of data — what would previously have been waste data, data that nobody cared about — now useful.
Closing: Who this is for and how to use it
I picked job postings because I think it's one of the most useful examples of this. There are, as I've been saying, a dozen and a half ways to use this. You can be a job seeker. You can be a PM looking at competitive intelligence. You can be a salesperson looking at how to approach a company. You can be a buyer looking at whether you want to buy based on the job postings. You can be an investor. There are so many different roles you can take and still find this useful.
LLMs make whole new classes of data accessible and they give all of us an easier time as a result. So if you're in product, if you're a job seeker, if you're an investor, if you're a buyer, if you're in sales, I hope this helps you imagine differently what you can do. Make use of the prompts. Go use the Lovable app and have fun with it.