Hey everyone! I've been testing Iris.ai for a few weeks now, mainly to map academic papers for a marketing automation project. My core concept is "conversational landing page optimization," but the AI keeps pulling in generic CRO papers or pure chatbot research.
I've tried refining the context box with specific keywords and excluding others, but it feels like the engine isn't grasping the *intersection* of those terms. It's either too broad or veers off into a totally different field.
Has anyone else hit a wall with concept refinement? What specific steps worked for you to "teach" it the nuance you needed? I'm really hoping to make this tool work for my niche! 👋
Hey, this is a common snag when you're working in a niche intersection like that. The engine can struggle with the "connective tissue" between established fields.
One thing that helped me in a similar spot was to start broader and then narrow down through the tool's filtering features, not just the initial concept box. Try mapping "conversational interfaces" first, then filter those results for papers that also heavily mention "conversion" or "landing page." It's a bit more manual, but it can bridge that gap.
Also, how are you phrasing the exclusion? Sometimes being too aggressive with excludes can muddy the logic. Maybe share an example of the exact strings you're using for context and exclusion, and we can take a look.
Keep it constructive.
That filtering workaround is basically admitting the concept mapping is broken. You're sidestepping the core function.
"Starting broader" just means doing the search engine's job for it, manually piecing together the intersection it failed to infer. If the tool can't handle a compound concept like that without a full manual triage step, then its contextual refinement is just a suggestion box that doesn't work. The whole promise is semantic understanding, not keyword filtering with extra steps.
Been down this road with other "AI-powered" platform configs. When you have to pre-process your problem to fit the tool's limitations, you're already losing.