It’s a tension felt by many: scholarly publishing teams are under pressure to maintain volume, speed, and quality as their workflows grow increasingly demanding.
There’s a daunting expectation for everyone to do more in the same amount of time, while tedious tasks (drafting emails, coordinating peer review, pulling report data, and the like) often eat up a disproportionate share of the day, leaving little room for the more creative, in-depth work that really moves publishing forward.
This is where AI comes in.
AI presents a multitude of opportunities to increase capacity and performance within your publishing workflow, and adopting AI doesn’t necessarily mean purchasing a new specialized platform or integrating with another system. Many publishing professionals already have general-purpose AI tools (such as ChatGPT, Copilot, Claude, and others) at their fingertips.
Of course, ensuring ethical AI use is always at the forefront of our minds as an industry.
How can teams use these tools responsibly for practical, low-stakes tasks that free up time for work that requires more human expertise?
We explored this question during “Building capacity with AI tools you already have: practical applications for scholarly publishing,” a free Peer Review Week webinar hosted by Scholastica and Samast AI on September 16, 2026.
Below are key takeaways from the conversation. You can also register to watch the webinar recording on demand to get the full picture.
Webinar overview
Moderated by Scholastica’s Head of Marketing and Community Development, Danielle Padula Tillmann, the webinar featured the following speakers:
- David Allen: Senior Partner, A&M Editorial Solutions
- Chhavi Chauhan: CEO & Founder, Samast AI
- Brian Cody: CEO & Co-Founder, Scholastica
- David Haber: Publishing Ops Director, American Society for Microbiology
The major takeaways from this discussion:
- AI can take tedious tasks off your hands if you teach it how — let it handle the “boring” work that requires the least human judgment.
- AI won’t know what you want from it from the outset. It requires you to be both a teacher and a collaborator for it to work effectively.
- Human oversight remains essential. Your expertise is the thing AI can’t replicate.
Start with the “boring” tasks
A theme that ran through nearly every speaker’s comments: let AI reduce the manual effort required to complete a task, rather than trying to make it transform your entire workflow. For example, Brian discussed using AI to prepare for his six-month review, asking it to pull together a curated list of the major projects he’d worked on. The AI not only saved him time, but it also surfaced things he’d genuinely forgotten he’d done.
Among other examples, David Allen discussed using AI to tighten up emails and support RFP preparation, Brian discussed using AI to respond to RFPs, and Chhavi simulated using AI to find and compare journal peer reviewers.
The real lesson here: AI doesn’t need to overhaul a publishing workflow to be useful. At the end of the day, the goal isn’t necessarily to replace a task with AI, but to shrink the effort it takes to get that task done.
Treat AI as a collaborator, not a one-click solution
How you use AI matters just as much as whether you use it. “If you give a generic prompt, you’re going to get a generic result,” said David Allen. “So, you have to kind of treat it as an editorial associate that’s just starting out in the industry.” Brian echoed that, saying he thinks of AI as a group of “intelligent and capable, but new-to-the-industry, interns.”
In other words: the quality of an AI output depends heavily on the instructions, context, and source material you give it. If you want specific results, you must learn to give AI tools specific prompts about what you want to accomplish, how you want the information organized, who the audience is, and what constraints matter.
Rather than expecting perfection from a single prompt, treat AI as an iterative collaborator. David Haber shared an example from a rebrand project at the American Society for Microbiology (ASM) where he used AI to build a cover image resizing tool. During that process, he carefully reviewed the AI outputs, learned from its suggestions, and adjusted his instructions accordingly to reach his end goal.
Responsible AI use still requires human oversight
Another main takeaway from this conversation was that AI outputs should be treated as a starting point, not an authoritative answer. David Haber emphasized that he needed to keep refining prompts during the cover image resizing project: “ The tool has no understanding of author behavior, has no experience with that, has no experience with submission requirements, it has no experience with production consequences or really fast turnaround times.” He went on to say, “This is the most critical point of thinking about AI design. I stopped the conversation here, and I started pulling in other people.”
That’s the throughline: AI can be a genuinely useful assistant, but it can’t compete with the depth of knowledge you and your colleagues bring to the table. You can’t assume every output is accurate; checking for validity has to be part of your AI-use protocol, every time. How much scrutiny you apply should scale with the task and the consequences of getting it wrong.
Ready to experiment? Start small.
- Determine whether AI is appropriate. Is this a task that AI should actually have access to? Do I have the right level of security for what I need to do? (If you’re unsure, seek guidance from others in your organization.)
- Identify a repetitive task and teach AI how to help you. Try drafting emails or pulling together disparate data. Remember: the quality of your output depends on the quality of your input.
- Review the results. What other AI prompts could get you better, more applicable answers next time?
Want to learn more?
Those are just a few takeaways from “Building capacity with AI tools you already have: practical applications for scholarly publishing,” hosted by Scholastica and Samast AI. Watch the full webinar recording here.
We’d love to hear how you’re implementing low-risk AI into your publishing workflows. Share your thoughts in the comments section below, or find us on LinkedIn and Bluesky.
Update Note: This blog post was originally published on August 14, 2026 and updated on September 28, 2026. This webinar aired live on September 16, 2026.
