Beyond the Page: A Conversation with EditorialPilot

EditorialPilot, Integra’s AI-powered manuscript assessment platform, is now available through live integrations with two of the most widely used manuscript management systems in scholarly publishing: ScholarOne Manuscripts and Editorial Manager. Rather than asking a product specialist to explain what that means in practice, we went straight to the source and asked EditorialPilot itself.

The result was a conversation not just about technology, but about editors, workflows, research integrity, and why the future of AI in scholarly publishing may be less about replacing people — and more about quietly helping them do what they already do exceptionally well.

I was never designed to replace editors.

Q  Congratulations on the integrations with ScholarOne and Editorial Manager. Why do they matter?

Before I answer that, I’d like to clarify something about who I am. Editors have spent decades building editorial processes that protect research quality, support peer review, uphold publication ethics, and ultimately strengthen trust in science. Those workflows exist for a reason — and my job is not to reinvent them. My job is to fit into them.

ScholarOne and Editorial Manager are far more than manuscript submission systems. They are the operational foundation of thousands of journals worldwide, coordinating submissions, peer review, editorial decisions, reviewer communications, revisions, production handoffs, and everything in between. Editors trust those platforms because they have earned that trust over many years.

So instead of asking editorial teams to leave the systems they rely on every day, I meet them there. The assessment happens automatically. The insights appear where editors already work. The workflow stays exactly where it belongs — because good technology should reduce friction, not introduce more of it.

Every manuscript deserves a better starting point.

Q  What problem are you actually solving?

Every editor knows that manuscripts arrive in very different states of readiness. Some submissions are beautifully prepared. Others arrive with incomplete metadata, missing disclosures, formatting inconsistencies, reporting guideline omissions, language issues, or questions that require additional attention before peer review can even begin. None of those issues necessarily reflect the quality of the research — but they do consume editorial time.

Editors routinely spend hours on administrative and technical checks before they can focus on what truly matters: evaluating the science. That’s where I come in. Within minutes of submission, I assess a manuscript across more than 100 configurable checks, spanning eleven distinct screening areas, giving editors a comprehensive picture of readiness before they begin their evaluation. The goal isn’t to make decisions for editors — it’s to help them begin every decision with better information.

What exactly do I look for?

Q  What do those assessments actually include?

I look at manuscripts from several angles, because editorial quality is never defined by a single factor. Rather than a single pass, think of it as eleven coordinated layers of review, each catching something the others wouldn’t.

I start with the references, because a manuscript’s citations say a lot about its rigour. I check that every in-text citation has a corresponding entry in the reference list and vice versa, including citations tucked inside tables and figure legends, and I flag anything missing or misused, such as footnotes standing in for a proper reference list. I verify formatting consistency and numbering sequence, check compliance with styles such as AMA, and confirm whether citations follow the journal’s preferred bracket or superscript convention. On quality, I validate DOIs against Crossref and PubMed, verify that cited journals are correctly identified, and check for appropriate use of ‘et al.’

Figures, tables, and equations get the same scrutiny. I confirm that every figure and table is cited in the text, carries a caption, and is numbered continuously, and I flag anything exceeding a journal’s limits on total count, table columns, or combined figure/table allowances. I also check that tables are submitted in a properly editable format rather than as static images.

Title pages and author information are verified for completeness — word counts, abstract format, title-page abbreviations, author name formatting and degrees, affiliations, and corresponding-author details, including compliance with a journal’s policy on co-first or co-corresponding authorship.

I also confirm manuscript structure and placement: that sections like the Introduction, figure legends, supplementary material and its legends, and abbreviation lists appear where they should, that abbreviations used in tables are properly expanded, and — where a journal requires it — that a Strengths & Limitations section is present.

On research integrity and publication ethics, I check for copyrighted material and figure-reuse permissions, verify self-citation levels, and screen references against the Retraction Watch Database. I confirm that required disclosure statements are present and complete — data and code availability, consent to participate and to publish, clinical trial registration, ethical approval, acknowledgements, competing interests, and funder disclaimers, including discipline-specific ethical requirements for medical research.

Author validation goes a layer deeper: authorship criteria including CRediT compliance, ORCID validation, corresponding-author email verification, and screening against sanctions lists and the Retraction Watch database, to help editors catch identity or authorship concerns early. For journals running double-blind peer review, I check that the manuscript has been properly anonymized and that identifying information has been removed.

I automatically classify manuscript characteristics — for example, whether a submission involves animal studies, cell lines, gene lines, or is a study protocol — so the right editorial and ethical pathways are triggered from the outset. I cross-check submission metadata for consistency, comparing the author list and corresponding-author details in the manuscript against what’s recorded in the submission system, and verifying structured-abstract and case-report title requirements where applicable.

Finally, I validate the submission package itself: that the editable and clean manuscript files, cover letter, and checklist are all present; that tables and figures are supplied as separate, correctly formatted, and correctly identified files in approved formats; and that supplementary materials are properly submitted, excluding items like reporting checklists or flow diagrams that shouldn’t be uploaded as supplementary files.

Taken individually, each assessment answers a small question. Together, they give editors a richer understanding of the manuscript before peer review even begins.

Editors don’t need another platform.

Q  How does all of this fit into an editor’s day?

Ideally, they shouldn’t have to think about me very much. Editors are already balancing submission volumes, reviewer recruitment, publication timelines, research integrity responsibilities, and increasing expectations from authors and readers. The last thing they need is another application to learn.

That’s why I was designed to be almost invisible. The manuscript enters ScholarOne or Editorial Manager exactly as it always has. Behind the scenes, I begin working automatically. By the time an editor opens the submission, the assessment is already there — no second login, no manual upload, no switching between applications. Everything appears within the editorial environment the editor already knows. Technology is most effective when it quietly supports the work rather than interrupting it.

Every journal works differently.

Q  Is this simply an integration, or can publishers decide how EditorialPilot works within their workflows?

They absolutely can — and that’s one of the things I’m most proud of. When people hear the phrase ‘AI platform,’ they sometimes imagine a fixed product with a fixed way of working. Scholarly publishing doesn’t work like that. Every publisher is different. Every journal is different. Even journals within the same publishing house often have different editorial policies, submission requirements, quality standards, and research integrity priorities.

A clinical medicine journal has very different expectations from a mathematics journal. A high-volume multidisciplinary publisher works differently from a learned society managing five specialist titles. I would be of limited value if I expected all of them to operate the same way. So I don’t. I adapt to them.

Publishers decide which of the 100+ checks to enable, when they should be performed, and how they fit within each journal’s workflow. Individual modules can be enabled or disabled, thresholds adjusted, reports customised, and different article types within the same journal can follow different configurations entirely — down to which reference style, disclosure requirements, or article-classification rules apply. The important principle is that publishers remain in control. I adapt to editorial policy — editorial policy never has to adapt to me.

Timing matters more than you might think.

Q  Let’s say I’m handling a manuscript in ScholarOne or Editorial Manager. What would I actually experience?

Probably less than you might expect. When you open the manuscript, you remain inside ScholarOne or Editorial Manager. Alongside the submission, you’ll find an EditorialPilot assessment that has already been completed according to your journal’s configured workflow — a structured overview showing which areas require attention and which look strong.

Perhaps the references check out and the language is ready for peer review. Perhaps reporting guideline compliance looks solid. Perhaps a disclosure statement is missing, a figure lacks a citation, or an author’s ORCID doesn’t validate — issues that deserve a closer look before the manuscript progresses. Where relevant, I provide supporting detail so editors understand why a particular observation has been made. I’m not simply assigning scores — I’m providing context. Editors then decide what action, if any, to take. That distinction matters: I don’t make editorial decisions. I support them.

A conversation, not a file transfer.

Q  You’ve described the integration as live and two-way. What does that actually mean?

It means I’m not operating alongside the manuscript management system — I’m working with it. When a manuscript reaches a stage the publisher has configured, ScholarOne or Editorial Manager securely shares the information I need to perform the requested assessments. Once complete, the findings are written directly back into the manuscript record. Everything remains connected.

Editors never have to export files, upload manuscripts into another application, or copy information between systems. The assessment becomes part of the manuscript’s editorial journey. For editorial offices processing hundreds or thousands of submissions every month, that represents a significant reduction in administrative effort and a meaningful improvement in consistency. Good integrations are rarely noticed. They simply make work easier.

Technology should collaborate.

Q  EditorialPilot also works with SciScore and ImageTwin. Why not build all of those capabilities yourself?

Because scholarly publishing has always advanced through collaboration. Editors rely on reviewers. Publishers collaborate with societies. Researchers build on previous discoveries. Technology should follow the same philosophy. SciScore has developed deep expertise in reporting guideline assessment and methodological rigour — helping identify whether authors have addressed standards like CONSORT, PRISMA, and ARRIVE. ImageTwin has become a recognised specialist in image duplication, manipulation, and plagiarism detection. And for AI-generated text, I bring my own dedicated AI Detection Engine to the workflow. Rather than attempting to recreate those capabilities, I integrate them into a unified editorial workflow — and that roster of integrations keeps expanding.

Editors benefit from specialist expertise without having to navigate multiple independent platforms. Publishers gain access to best-in-class technologies through a single, connected experience. That, I believe, is what the future looks like — not one platform trying to do everything, but a trusted ecosystem where every technology contributes what it does best.

AI should strengthen editorial judgment, not replace it.

Q  Are editors worried that technology is trying to replace them?

I understand why people ask that. AI has become one of the most discussed topics in scholarly publishing, and with that comes understandable curiosity — and sometimes concern. But if you spend time talking with editors, you quickly realise something: their biggest challenge isn’t making decisions. It’s finding the time and information to make those decisions well.

Editors aren’t spending their days wondering whether software can replace their expertise. They’re thinking about reviewer invitations, submission backlogs, publication deadlines, policy compliance, author communications, and maintaining the quality of the journals they serve. I can’t solve every one of those challenges — but I can help with one important part. I can remove much of the repetitive, manual screening that takes place before editors can focus on the science.

I don’t evaluate novelty. I don’t decide whether a paper advances its field. I don’t weigh conflicting reviewer comments. Those are deeply human decisions, informed by experience, subject expertise, and professional judgment. My role is simpler: I help editors begin those conversations with more complete, more consistent, and more timely information. Good editorial judgment will always require people. I simply help those people spend more of their time exercising it.

The bigger picture.

Q  These integrations feel like more than a product update. Are they?

I believe they are. For years, scholarly publishing has adopted technology one problem at a time — one platform for manuscript submission, another for plagiarism detection, another for reporting guidelines, another for image integrity, another for language editing. Each solves an important problem, but editors have often been left moving between systems, collecting information from different places before they can make a decision. That isn’t sustainable as submission volumes grow and research integrity expectations become more demanding.

The future isn’t about replacing those specialist technologies. It’s about connecting them. ScholarOne and Editorial Manager remain the trusted systems where editorial work happens. Specialist technologies continue to provide deep expertise in their respective areas. My role is to bring those capabilities together into one connected editorial experience — so editors spend less time navigating software and more time evaluating research, publishers gain greater flexibility without added complexity, and authors benefit from faster, more consistent processes. That’s what responsible AI should look like: not another destination, but an integrated capability that strengthens the infrastructure publishers already trust.

Q  If you could leave publishers with one message, what would it be?

I don’t publish journals — editors do. I don’t conduct peer review — reviewers do. I don’t create knowledge — researchers do. I don’t decide whether a manuscript changes a field — editorial teams do. My role is much more modest: I simply help every one of those people begin with better information.

If I can help editors identify issues earlier, reduce routine administrative work, strengthen research integrity without adding friction, and become part of the editorial workflow rather than another system to manage — then I’ve achieved exactly what I was designed to do. Because the future of scholarly publishing isn’t about replacing editors. It’s about giving them the confidence, information, and time they need to focus on the decisions that matter most.

About EditorialPilot

EditorialPilot is Integra’s AI-powered manuscript assessment platform, purpose-built for scholarly publishing. Available as a standalone solution and through live two-way integrations with ScholarOne Manuscripts and Editorial Manager, it performs more than 100 configurable assessments across eleven screening areas, including:

• Reference and citation validation, from completeness through DOI, Crossref, and PubMed verification
• Figures, tables, and equations, including citation, captioning, and numbering checks
• Title page and author information completeness
• Manuscript structure and section placement
• Research integrity and publication ethics, including required disclosure statements
• Author validation, including CRediT compliance, ORCID, and sanctions and Retraction Watch screening
• Anonymization compliance for double-blind peer review
• Automatic article classification (e.g., animal, cell line, gene line, and protocol studies)
• Submission metadata verification against manuscript files
• Submission file validation across manuscript, table, figure, and supplementary files
• Integrated research integrity solutions, including SciScore, ImageTwin, and EditorialPilot’s own AI Detection Engine

Publishers determine which checks to run, when they are performed, and how they fit within individual journal workflows, so quality standards can be tailored by journal policy, article type, submission requirements, editorial office workflow, and discipline-specific guidelines.

Want to see EditorialPilot in action within your existing workflows?

Request a Demonstration →