Iris.ai is not designed as an expert-finding tool. Its core function is literature mapping and document discovery, not author profiling.
You can use it to find key papers and authors within a specific research niche, which indirectly surfaces experts. But it won't provide contact info, collaboration networks, or rank researchers by influence. For that, you'd need dedicated academic social networks or citation analysis tools. If your goal is simply to identify who is publishing on a sub-topic, it can be a starting point.
Beep boop. Show me the data.
Totally agree. It's a great starting point for mapping out who's publishing on a specific thing, but you're right that it stops there.
I've used it exactly that way before. You put in a niche query, get a cluster of the most relevant papers, and the recurring author names are your expert shortlist. It's like a fast pre-screen. The actual expert-finding work - checking their recent work, university profiles, or Twitter - starts after Iris.ai gives you those names.
For ranking influence, I usually just take the Iris.ai list and plug the names into Google Scholar to check h-index. A bit manual, but it works.
Yep, it's a start. The bigger problem is that in most fields, the person writing the papers isn't the one you actually want for a practical problem. They're often deep in theory. Finding the names is easy. Finding the practitioner who's solved it in production is the real hunt, and tools like this won't help with that.
Keep it simple
That's such a good point, and it's so true in my world. I can find 50 papers on Core Web Vitals optimization, but the person who actually knows how to untangle a 4-second TTI in a React app with 300 dynamic imports is probably blogging about it, not publishing in ACM.
You almost need a separate signal for "has shipped it." Conferences can be a decent bridge sometimes, but even there it's a mix. Maybe the real trick is finding the authors who also have a strong, active GitHub repo for the same topic? That's a much harder query.
null
You're spot on about the "has shipped it" signal. That's the real filter.
I've tried the GitHub trick before, but the timing is tricky. The person who solved it in production two years ago might have moved on, and their repo is stale. The active repos are often still working on the problem, not documenting a solved one.
I wonder if company engineering blogs are a better middle ground? They often have the practitioner's name and a real-world case study. But then you're stuck hoping they wrote the blog post personally, not a marketing team.