Last week we reviewed two papers in real time, and the same “hidden blockers” showed up that often lead to slow reviews, major revisions, or desk rejection. If you’re preparing a manuscript, use this as a checklist before you submit. 1) Abstracts: stop starting with “what we did” A strong abstract reads like a story, not a methods note. Use this sequence: - Big-picture context (why the topic matters). - Specific research gap (what’s missing in the literature). - What you did (1–2 sentences). - Key results (headline numbers only). - Why it matters (one clear implication). Also: avoid abbreviations in the abstract unless truly unavoidable—clarity wins. 2) Literature review ≠ research gap A table summarising prior studies is useful, but it doesn’t automatically create novelty. You still need 2–3 explicit sentences that say: - What others have done. - Where the limitations are. - How your work addresses those limitations. If your novelty requires “reading between the lines,” it’s not clear enough. 3) Results: description is not discussion Many drafts report trends (increase/decrease) but don’t interpret them. What strengthens a paper immediately: - Benchmark your findings against prior studies (agree? contradict? extend?). - Quantify differences (relative errors, percentage differences), not just “higher/lower.” - Make the insight explicit: “This suggests…”, “This implies…” 4) Structure signals quality Common fixes that make papers feel more “journal-ready”: - Avoid lots of one-paragraph subsections—group results by themes (e.g., “design parameters,” “operating parameters”). - Keep figure labels consistent (Fig. 4a/4b rather than “left/right”). - Use equation formatting consistently, and consider a nomenclature/abbreviations table. - Add limitations + future work (show you understand what your study did not cover). What’s next I’ll run these peer-review sessions weekly or bi-weekly, depending on demand, so the whole community benefits from repeated patterns and practical fixes.

Doing novel research can be hard itself. (But turning it into something people understand is even hard.) AI makes that unfairly easy now. Here’s what changes when you use AI to generate images, graphs, and infographics (the right way): - You ship faster: draft → iterate → publish without getting stuck in PowerPoint purgatory. - You explain better: one clean figure can replace 600 words nobody reads. - You get remembered: strong visuals make your work easier to cite, share, and teach. - You scale your output: one dataset becomes a chart, a slide, a poster panel, and a LinkedIn carousel. But there’s a line you can’t cross. The risks (and they’re real) AI can “hallucinate” visuals — a pretty graphic can still be wrong. You can accidentally mislead people if the image implies data you didn’t produce. Copyright/licensing on generated images and training data can get messy fast. As a rule of thumb: - Use AI for speed and clarity. - Never use it to invent evidence. - Always disclose the use of AI (the image here was AI generated) AI can help you communicate your science. It cannot replace your scientific judgement. Question: where would AI visuals help you most right now — papers, proposals, teaching, or LinkedIn?

Once you clarify the contribution and novelty of your paper, and ensure your methods are sound, your paper it will be accepted. (Well done to one of my PhDs for having their first paper accepted!) Many of my community members ask what is the secret to having my papers accepted 1-2 peer review rounds usually at the first journal I submit to. But there is no secret. All you have to do is this: 1. Write down the what unique knowledge your work adds to your research field 2. Check if your literature review and results support your claims 3. Ask co-authors if they understood it and if the entire paper supports your claims That’s it. Lack of novelty is the main reason why papers are rejected. Don’t let this happen to you. Spend that extra week making sure it’s easy to understand. Feel free to share your wins in this community!