How do I make it prioritize recent papers without ignoring seminal works?

Hi everyone. I've been tasked with conducting a literature review using Elicit for a significant upcoming procurement project. The goal is to map out the current landscape of AI-assisted vendor risk assessment tools. I've found Elicit incredibly powerful for surfacing a wide range of papers, but I'm struggling with a specific balancing act.

My core dilemma is this: I need to ensure our review is grounded in the established, foundational theories and methodologies (the seminal works from, say, the early 2010s onward in this field). However, given the pace of change in AI, it is absolutely critical that we capture the very latest research, models, and applications from the last 12-18 months. I'm concerned that by optimizing for one, I might systematically miss the other.

My current process feels a bit manual and I'm worried I'm missing a more efficient, repeatable method within Elicit. Here’s what I’ve tried so far, along with my observations:

* **Using the default "Relevance" sort:** This seems to mix older highly-cited papers with newer ones, but the mix feels unpredictable. I can't discern the algorithm's logic, which makes it hard to trust for a comprehensive scan.
* **Running separate queries:** One search with no date filters to catch seminal works, and another with a filter for "Last 3 years." This is my primary method now, but it leads to duplication of effort and forces me to manually deduplicate and synthesize two separate lists.
* **Playing with the "Cite count" sort:** While this clearly surfaces seminal works, it almost entirely drowns out recent papers, which haven't had time to accumulate citations. This is a non-starter for my need for current data.

My specific questions for the community are:

* What is your workflow or query strategy to systematically capture both historical foundations and cutting-edge developments in a single, manageable review process?
* Are there specific search operators, or ways to structure the research question, that signal to Elicit to value recency while still respecting citation influence?
* How do you handle the synthesis afterwards? Do you use Elicit's list features or export to another tool to merge and prioritize papers from multiple searches?
* From a total cost of ownership perspective, I'm also considering if this is a limitation where I need to complement Elicit with another tool specifically for alerting on very recent publications, and if so, what has worked for others.

I would be very grateful for any detailed workflows, lessons learned, or even pitfalls to avoid. My team is relying on this analysis to make a substantial platform decision, so I want to ensure the literature review backing it is both deep and current.