If you want to understand a company's real strategy, don’t just listen to their marketing. Instead, analyze their job postings. That's exactly what we did for OpenAI—and the insights were mind-blowing. In this post, I'll show you how we performed this analysis using Researchly—and then break down how you can replicate it with an N8N agent (see download at the end of the post).
Why Analyze Job Posts?
This particular post was inspired by Deedy Das, Partner at Menlo Ventures:
In general, Job posts are a strategic goldmine:
- They reveal which products and initiatives are prioritized
- You'll see which teams are expanding (or shrinking)
- You can spot markets or industries a company is targeting before they make announcements
For OpenAI, for example, we discovered they're looking for talent across hardware, B2C, B2B, developer platforms, and even global expansion, based just on their hiring data (see below)
How to analyze job postings using N8N, Firecrawl and AI Agents
(You can download the full N8N template at the end of the post).
To build your job posting scraper in N8N follow this approach:
First Part: Get the job posting links
The steps below refer to the first part of the above screenshot.
- Get the link to the job portal. In this case https://openai.com/careers/search/
- Use firecrawl to parse the links (they have an option to return “links” specifically). Depending on the target site, you might need custom scraping (for instance, if the job page has a lot of paginations)
- Save the jobs in a Google Sheet. Other options like Airtable work just as well, but I prefer Google Sheet for the simplicity.
Second Part: Get the data for the job post
The steps below refer to the second part of the above screenshot. Once you have all the links in the Google Sheet, you need to extract Job Title and Summary.
Simply use Firecrawl again, but this time request a summary of the page as well. You can extract the Job Title from the meta data scraped by Firecrawl.
Third Part: Analyze the data
Once you have all the data, you can use your LLM of choice and do your analysis. In my experience, GPT-5 provides the most insightful answer, but returns a lot of jargon as well. I use 4.1 to get a simplified version of the GPT-5 response. If you are going to do these analysis regularly, I suggest you setup a pipeline in N8N.
Insights from the OpenAI Job Post (22. 09. 2025)
Now to the insights:
Global expansion with APAC focus
- Signals: AI Adoption/APAC roles (Tokyo, Singapore); Strategic Partnerships Lead, Japan; Marketing Lead, Korea; Sales Strategy & Ops APAC; GTM Enablement Singapore; User Ops Tokyo; FDE roles in Tokyo, Paris, Munich; Support Lead EMEA; IT London; Senior Recruiter Tokyo.
- My take: Chinese AI Models (think Deepseek) are strong, the competition will be interesting
Developer platform and deployment-first motion
- Signals: Solutions Architects (GenAI Deployment), Partner SA, Forward Deployed Engineers in multiple hubs, Developer Experience Engineers, Customer Education Program Lead, Product Ops (Developer Platform).
- My take: With tools such as OpenRouter making LLMs a real commoditiy, OpenAI wants to ensure their API is super easy to use. That’s also how Strip won.
Safety, integrity, and trust as product pillars
- Signals: Dozens of roles in Safety Oversight, Scaled Abuse, Integrity Foundations, Youth Well-Being, Child Safety, Misalignment Research, Adversarial Model Research, Model Policy, Quant Threat Forecasting, T&S Operations.
- My take: Obvious, but it also shows that content moderation (like it was with Social Media) is hard
Owning supercompute: Stargate build-out
- Signals: Large set of “Superintelligence Infrastructure – Stargate” roles across power, land development, mechanical/electrical R&D, facilities ops, physical/security, cost control, design, commissioning, structural engineering.
- My take: Vertical integration in datacenters/energy to secure capacity and reduce unit economics for training and inference. This relates to their latest $ 300 billion deal with Oracle.
Hardware and edge device initiative (consumer-facing)
- Signals: Camera SW, Perception, Mechanical (Sensing), Embedded SW, Linux Kernel, System SW Manageability, Controls SW, Product/Manufacturing Quality, Secure Manufacturing & Stealth, Technical Hardware Sourcer (Consumer Products), Hardware Security (Trusted Computing/Crypto).
- My take: There are working on a hardware consumer device. Very similar to META’s glasses
Growth engine across B2C and B2B funnels
- Signals: Numerous Growth roles (SEO/CRO, Emails/Lifecycle, Performance Marketing, Paid Marketing Platform Eng), Comms/Marketing for developers, Content for ChatGPT.com, GTM Innovation.
- My take: Inbound (SEO, content) remains an important channel despite AI search.
N8N Templates
- Here is the link to the N8N template to analyze job postings.
- Note: It is a bare version, you will need to adjust if you want to scrape job posts en masse
