In the second of our Women in Scholarly Communications series, we meet Shilpi Mehra, Product Director at Cactus Communications
Supported by Cactus Communications
For someone helping to shape the future of scholarly publishing, Shilpi Mehra arrived
As Product Director at Cactus Communications, she spends much of her time thinking about research integrity, artificial intelligence and how publishers can manage ever-growing volumes of submissions without compromising quality. But before entering scholarly communications, she had already built careers in dentistry, public health research, and clinical trial management, working across countries including India, the USA, the Netherlands, the UK, and now, Dubai
That breadth of experience has given her an uncommon perspective within scholarly publishing. Having been a researcher herself, she understands the years of work behind every manuscript. Having worked with publishers, she also understands the operational pressures that threaten to overwhelm editorial offices
Throughout our conversation, a key theme emerged: technology is transforming scholarly publishing at extraordinary speed, but trust remains its most valuable currency
LISTEN TO THE WHOLE INTERVIEW HERE:
Seeing publishing through a researcher’s eyes
Shilpi originally trained as a dentist, attracted by the health sciences from an early age. It wasn’t the science that persuaded her to change direction – it was the nature of clinical practice itself
“I generally just like speaking with people,” she says. “I liked helping my patients understand why they had dental health issues and what they could do to prevent them.” But dentistry offered little opportunity for conversation while patients sat in the chair (“they’ve got their mouth open, so they can’t really talk!”), so a move into public health research beckoned. Rather than treating individuals, she found herself studying populations, first in Boston and then in Amsterdam, before moving into research management roles with the NHS and Imperial College London in the UK.
Looking back, she laughs at how little thought she gave to scholarly publishing at the time. “I’d never thought much about scholarly communication,” she says. “You do the research and the end point is you’ll publish it.”
Working in well-established research environments, publication simply seemed like the natural conclusion to a project. It wasn’t until she joined Cactus Communications after returning to India, that she discovered the scale and complexity of the industry supporting that final step
“That was when I discovered that scholarly publishing is a whole new world,” she recalls. “I had no idea that the industry was this large and the amount of work that went into it, was so big.”
Shilpi entered the sector almost by accident. A LinkedIn advert initially caught her eye because it offered freelance scientific writing – an attractive proposition while raising young children. But during the interview process, Cactus recognised that her background in research and trial management would be better suited to supporting researchers directly. Seven years later, she’s still there
Every manuscript represents years of work
Shilpi believes her career outside publishing fundamentally changed how she views scholarly communication today; she has experienced first-hand the effort required to produce research
“Without a doubt,” she says when asked whether that perspective influences her work
She remembers recruiting patients into clinical trials, helping research teams deliver projects on schedule and seeing the enormous commitment required before a manuscript ever reaches a journal
“When you’re on the publishing side, you feel like it’s just one of the hundreds of thousands of manuscripts,” she says. “But when you’re actually there – helping somebody do the research or you’re doing it yourself – then you know how much time, effort, blood and sweat goes into producing a manuscript.”
It’s a perspective that continues to guide her work developing publishing technologies.”These are not just numbers,” she says. “There are people behind each one of those.” That reminder has become increasingly important as publishers grapple with unprecedented submission volumes
Integrity is no longer just an ethical issue
Research integrity dominates conversations across scholarly publishing, but Shilpi believes the debate has evolved. Fraud remains a major concern, she says, yet the industry’s most immediate challenge is now operational. Publishers are no longer simply trying to identify problematic papers. They are attempting to do so while processing record numbers of submissions – many of which require increasingly sophisticated screening. The consequence is that legitimate research is often delayed
“Good research is taking a lot more time to come through and get published,” she says. “That’s the one thing we don’t want. We’re stemming the flow of good science because there’s just too much bad science around.”
The conversation turns to paper mills and fraudulent submissions. Is the problem being driven by a relatively small number of bad actors? “Oh no,” she replies immediately. “It’s systemic.”
A decade ago, misconduct often involved isolated individuals cutting corners. Today’s challenges are rooted far more deeply in the incentives underpinning global research, she says: “The pressures of publishing and all of the rewards that are related to it… promotions, being able to keep jobs, getting funded for more research… it’s now mostly related to publishing and being cited.”
Researchers, universities, funders and governments, all respond to metrics in different ways. Publishing success has become intertwined with career progression, institutional reputation and funding decisions. “It’s everybody at every level,” she says, adding that there is no single technological solution
She points to China as an example of how policy has evolved over time. Encouraging publication in high-ranking journals was meant to improve quality of research, but also created new incentives and avenues for fraudulent behaviour. More recently, penalties for researchers with retractions, is an attempt to deter those behaviours
Neither approach, she believes, tackles the underlying issue: “It’s like… you’re putting a Band-Aid over it. Ultimately, the industry needs to rethink what it actually rewards. It should not only mean an impact factor or CAS quartile or citation count.”
Publishers have a unique responsibility
However, Shilpi is equally clear that publishers cannot view integrity as someone else’s problem. They are the “gatekeepers”, she says, and that responsibility extends beyond editorial workflows
Publishers also have the influence to shape industry-wide conversations and collaborate with universities, funders and technology providers to improve research practices
She has seen that collaboration accelerate in recent years. Reflecting on successive World Conferences on Research Integrity, she recalls that publishers were once a relatively small presence. Recently, they have become more central participants, presenting posters, sharing new fraud detection methods, and joining mixed panels to brainstorm solutions, alongside researchers and institutions. “That’s exactly the step in the right direction,” she says
Shilpi also believes the industry’s appetite for experimentation has changed dramatically. Only a few years ago, solving major problems often meant months of discussion before any practical work began. Today, organisations are prepared to pilot ideas quickly
“Everyone’s saying, ‘Yes, let’s start a pilot… let’s start a trial… let’s see how we can do this.’”
That willingness to experiment gives her confidence that publishing will eventually move from reacting to integrity challenges towards anticipating them. “I’m confident about it because there’s no other way to deal with it.”
The pace of change is fast, she admits, but necessary. “We’re not used to working with things changing so quickly.”
But, she notes, publishers, technology companies and industry organisations are all investing heavily in solutions – and the ecosystem appears to be moving in the same direction
AI will redefine editorial work – not replace it
Few subjects generate more debate than artificial intelligence, yet Shilpi takes a measured view of its impact. She rejects simplistic predictions that AI will replace editors: “I don’t think that that’s the way that it’s going at all.”
Instead, she expects people and technology to divide work differently. Routine tasks – screening documents, identifying patterns and surfacing evidence – are increasingly well-suited to AI. Context, judgement and interpretation remain distinctly human
“AI can be very good at working on routine, mundane stuff – it’s usually much faster and more accurate than humans at that.”
But concepts such as novelty and contextual judgement simply cannot be automated: “We’re not going to be able to say (anytime soon!)… run your paper through this tool and we’ll be able to tell you whether this research is novel or not.”
For publishers, she believes the greatest opportunity lies in using AI to support decisions rather than make them. She says the best use of AI is for evidence presentation: it should gather relevant information, identify potential concerns and reduce repetitive work, allowing editors to focus on applying their expertise to “good” science
Human judgement, she argues, remains essential: “The final responsibility or accountability always lies with the person who’s signed it off.”
From detecting bad science to supporting good science
While much of the discussion around AI focuses on identifying fraud, Shilpi’s longer-term vision is considerably more ambitious. She hopes integrity checking will eventually become almost invisible
In many respects, Shilpi compares AI to modern writing assistants. Just as language models can draft emails, humans still decide whether those emails should be sent. Publishing, she believes, will work in much the same way
“My hope… is that integrity checks and trust markers would run in the background. Nobody consciously thinks about spelling software any more; it quietly improves documents while they’re being written.”
Rather than waiting until submission to identify problems, publishing systems could help researchers produce stronger manuscripts from the outset – flagging missing references, highlighting copied text or prompting authors to strengthen evidence while they are still writing
Increasingly, she believes the conversation is shifting away from asking how publishers identify bad science and towards asking how the entire research ecosystem encourages good science. Publishers, universities, funders and organisations such as the STM Integrity Hub, are beginning to focus on trust markers rather than simply fraud detection. She believes that is a healthier direction for the industry
Finding perspective outside publishing
Away from work, Shilpi’s interests couldn’t be much further removed from AI and editorial workflows. Gardening has become her escape
After spending a decade living in Britain, she developed an enduring appreciation for the UK’s enthusiasm for gardens despite the weather. “It’s amazing to me how despite the bad weather, everybody works so much in their gardens.”
Whether pruning, weeding or simply tending plants, she finds the process deeply restorative: “It brings me a lot of joy.”
Asked what career she might have pursued had scholarly publishing never appeared, she smiles. Having already enjoyed several professions, she suspects she would have returned to dentistry – this time specialising in paediatric care. Helping children overcome their fear of dentists proved just as rewarding as the clinical work itself: “I really loved working with the kids.”
Trusting in trust
Shilpi is clearly excited about the potential of AI and believes technology will reshape scholarly publishing; she expects editorial workflows to look very different in 10 years’ time. But, for her, none of those advances will matter unless they make the research ecosystem more trustworthy
Her optimism isn’t driven by technology alone. It’s based on the belief that publishers, researchers, funders and technology providers are willing to work together
Having seen research from many different angles – as a clinician, researcher, trial manager and now a product leader – she knows that every manuscript represents years of work by real people
Technology can help publishers handle growing volumes of research, spot patterns, and make processes more efficient – it can speed up good science and reduce routine tasks. But it cannot replace the human judgement, accountability and shared responsibility that scholarly publishing relies on
Interview by Tim Gillett
LISTEN TO THE WHOLE INTERVIEW HERE:
