Many people use ChatGPT as a secondary search engine, but there are times when you’ll want to go deeper than a quick Q&A prompt. You might be looking to learn a new skill, tackle something for work, help with your child’s homework or simply explore your curiosities.
Now you can do it all on one AI platform, instead of jumping between Google, ChatGPT, YouTube, podcasts, books, journals, encyclopedias and more.
Manual research and exploration will always have a place — it’s how we learn and how we double-check that AI is providing accurate information. That said, having a deep research assistant is useful. I decided to test one out using Perplexity.
Getting set up with Perplexity
The setup is simple. Go to Perplexity’s website and log in. I’m using the free version, but if you want access to the latest models, deeper sourcing and no usage limits, go for the Pro version at $15/month.
Make sure to select the Deep research tab. It looks like I only have one query available, so let’s see how far free gets me.
What’s really cool is that you can control where it sources information from. It pre-selects the web, but I toggled that off and used Academic. This is great for work and study purposes, as Perplexity says it «prioritizes results from scholarly databases, journals, and reputable academic publications, filtering out non-academic or general web sources.»
As a journalist who is going through IVF, I’ve been writing more about fertility and reproductive technology. Reporting on anything health-related requires a great deal of research and fact-checking. I tend to do all of this manually, looking at scientific studies and stats from reputable industry associations.
Being a 37-year-old aspiring parent, I’m interested in finding out the root cause and reason for the term «geriatric pregnancy.» Is it just that egg quality starts to decline after 35? If a woman creates genetically normal embryos, does age still matter?
So, I asked Perplexity.
It generated an in-depth report with 128 sources, which you can download as a PDF. My report came in at 10 pages, with all the references hyperlinked.
This would’ve taken me hours to do manually. At first glance, I can see the reference materials are all relatively recent (unlike when I use ChatGPT for the same kinds of tasks).
Now, let’s dig into the information to see whether it came from reputable sources. I spent about 10 to 15 minutes scanning the report to see if it could tell me something I didn’t already know.
It allows you to pull out interesting lines to expand further, like this one:
This line says, for example, that even with a genetically normal embryo, age still plays a role.
And another interesting line reveals a sharp drop in success at 38, not 35:
This was fascinating to me. Perplexity found research from 2024 stating that a personalized embryo transfer protocol with further testing in older women achieves the same outcomes as younger women (so long as the embryo is normal):
Perplexity’s summary was also very helpful. It drew conclusions with supportive stats.
I also learned that endometrial receptivity declines with age, but it’s not irreversible.
I loved the closing, which described «parallel aging across the reproductive system.»
We hear a lot about egg quantity and quality, as well as uterine receptivity, but less about placental aging. In fact, I’d never heard that term before, so I knew I was onto something.
In my follow-up prompt, I asked Perplexity to explain placental aging and whether it’s a new scientific finding.
Plus, Perplexity populates follow-up prompts like this, which helped guide my research:
I learned that placental aging shows up more commonly in IVF pregnancies. I’ve got my story nugget.
The verdict on Perplexity for deep research
The entire process took about an hour, from opening Perplexity to identifying new information for a research topic I’m working on. Not bad for a free tool.
My favorite part was being able to select only academic sources. It saved me a lot of time, and I liked how the chatbot made academic research much more fun and interactive.
Just be sure to click those hyperlinks to the sources and double-check that the AI tool has accurately reflected the academic findings, since AI tools are known to hallucinate or draw the wrong conclusions.
