Skip to content
Home » The Effortless Blog » Identify Novel Research questions with ResearchKick

Identify Novel Research questions with ResearchKick

Asking the right question is half the answer in academia. Questions need to be novel (to be published) but also outside the sphere of what is known (to be able to answer them). They are also key to an engaging discussion and future research sections. In this deep dive, we are looking at ResearchKick, a tool built specifically to discover new questions.


Join the Effortless Newsletter. Receive free tips on note-taking, literature review, AI tools, and other productivity topics specifically tailored for academics and students.


ResearchKick is as easy to use as ChatGPT. You start with a topic and receive a number of questions, refine each question until you arrive at something that fits your research. Lastly, Researchkick can help you identify papers that have been published around this question. Here is a schematic overview of the process:

The developer behind ResearchKick is, by the way, Dr Mushtaq Bilal , who is known on Twitter for his ambassador-like role between academics and the AI community. So let’s get started with finding a research question.

Signing Up For ResearchKick

To sign up for ResearchKick go to the website and click the “Get Started” button at the top.

Signing up for ResearchKick using Google or any Email Address.

However, to use ResearchKick you will need to buy credits. See the pricing section on how it works. (The introductory video points out that there is only a small team behind ResearchKick, and without investment or going into debt, this is the only way to do it.)

Describe your Research Topic

To get started with ResearchKick, you will need to define a research topic. When using any AI software, you have to assume that it is completely impersonal and does not know where you are in the scientific journey. Therefore, it is necessary to describe the topic in some depth. On the other hand, overly describing your niche makes it improbable to find questions of broad interest.

Think of your research topic as a dinner party conversation—interesting enough to engage your peers but not so specialized that only a few experts can contribute. Aim for that sweet spot where familiarity meets innovation.

Choosing a topic in a ResearchKick conversation.

Notice that in this example, I use a concise but targeted topic description. I am interested in range shifts (which is ecological jargon for plants changing their locations/range), but describe very specifically that I am looking for woody plants (i.e., trees, not herbs, grasses, mosses, or bushes!). Adding “climate change” relation immediately narrows it down from paleo-archaeological questions to more present-day anthropogenic climate change related questions. While the description is concise, it allows for interpretation where needed without being overly strict. 

Once you have described your topic, ResearchKick will suggest a number of questions you might want to explore. Not all of these are necessarily novel questions. But if you are writing a paper on a specific topic, you can use these questions to: 

  • Direct your research and literature review efforts
  • Discuss them in your discussion chapter
  • Answer them for yourself to build understanding of the big picture of your research topic
  • Prepare for them, as these are questions an audience would ask you at a conference

Lastly, you will likely find one question that is closer to your research. We can dive into it in the next step.

Refine a Research Question with ResearchKick

In the previous step, we received a set of questions, one of which likely resonates. If none do, you could try to rephrase your topic in a new conversation. When refining a question it is enough to say “Refine Question N” but the results are generally better if you copy and paste the question into the follow-up prompt as it helps the AI to maintain the context of your question:

You can continue refining your questions recursively. In my experience, just 1-2 iterations of refinement narrow your questions fairly well. 

Analyze the Literature on the Question

The last step is to check the literature on this question. AI is becoming progressively capable of finding literature based on a plain text question. This is called semantic search. You can learn all about semantic search and various tools in my latest literature review workflow. ResearchKick, luckily, already has such a search engine integrated. Just ask it if a question has already been addressed:

The results are several studies with summaries and links. Because ResearchKick is using a search under the hood, it does not make those studies up, and you can click the links to read the study in detail. 

It is always a great idea to center your literature review around a topic. So, instead of saving the paper in your Zotero (or Mendeley) folder, add it to a reading list and specify why exactly you want to read this paper. Refer to the question you just found. This way, if you come back to it months later, you will still know if something is relevant to you at that time. This simple technique is one of three I recommend to optimize your literature review.

Here is how such a reading list can look over time:

Using this method, your ResearchKick conversations double as a literature review, and the effort is not wasted!

ResearchKick Databases

In the previous step, ResearchKick generated a number of citations. These are found in one of the three databases you can use: Semantic Scholar, Scite AI and Data, or Biomedical and Life Sciences (likely PubMed). All of these databases have access to over 200 million papers, but some are better suited for certain topics.

You can switch between the databases by clicking the “Change Database” button at the top left of every conversation:

Is there a best database? In my personal experience, the differences were quite small, and even repeated conversations would lead to similar results. However, it is worth trying it out at the beginning to see which one works best for you. Then, commit to a single database for the rest of the process and forget about it.

ResearchKick Pricing

Every conversation with the ResearchKick AI consumes credits, which are used depending on the length of the conversation. In my experience, I used just under 1000 credits for a conversation to find a research question over multiple steps, as outlined above. Here is how to buy credits:

You can buy credits directly (to try it out) or get a subscription. A subscription (1200-1100 credits/$) is 5 times cheaper than directly purchasing credits (208 credits/$). Credits in the pay-as-you-go package expire after one month.

Summary

Using ResearchKick is as simple as using ChatGPT (and indeed, it is built on top of ChatGPT-4). The tool is specifically designed to find compelling and novel research questions (or even come up with things to ask during a webinar). To get to a novel question, use these three steps:

  1. Describe your topic concisely and clearly but without becoming too narrow.
  2. From the list of questions, ask to refine the one(s) that align with your interests/capabilities.
  3. Ask ResearchKick if the question has already been addressed for a mini-literature review.

Write down the papers ResearchKick finds for later reading.