There is no shortage of AI tools promising to transform research. Barely a week goes by without another launch – another platform powered by a large language model, another interface built on public data, another chatbot that will supposedly revolutionize how we work. And there is no doubt that they can be very useful tools; this blog, for example, is written in partnership between a human author and AI.
We want to explain why we think Dimensions Research Strategy is different to other AI tools in research.
The problem with most AI tools for research
The challenge with general-purpose AI in research strategy isn’t capability: it’s the lack of provenance and domain-level depth. Ask a large language model a question about your institution’s competitive position in oncology research, and it will give you a confident, plausible answer. But, it won’t give you an answer that shows an understanding of how actors in the research environment think about these problems; it won’t be able to tell you the assumptions that are implicit in how it arrived at its analysis and, most importantly, it won’t give you provenance or reproducibility.
It can’t tell you which data it used, whether that data are current, how it weighted different signals, or whether its methodology reflects how a skilled research analyst would actually approach the question. The answer sounds right at first sight, and it’s confidently given. You just can’t defend it when someone asks you to.
This is a problem we’ve all encountered with chatbots.
The minor errors that slip into analyses – unchecked or unfounded assumptions – perhaps don’t matter too much when you’re just asking for a recipe or collecting actions from a daily meeting you missed, but they are foundational when determining a research strategy, portfolio analysis or when considering a tenure-track proposal. In these situations, AI must be used carefully, in a limited way. AI has tremendous promise to help in situations of cognitive overload, and so the foundations on which it is built matter enormously when the question has deep consequences for a person, a department, a strategy, a research program or even a research field.
Why Dimensions Research Strategy is different: every finding traceable to source
This is the thing we think will matter most to the people who will actually use this, and the thing we’ve seen really resonating in our discussions with the community: the provenance of every piece of data.
When a Vice-Chancellor receives a Dimensions Research Strategy analysis during a meeting with senior colleagues to discuss their research strategy, they need to know that every number in it can be challenged and defended. So we’ve made sure that every finding in Dimensions Research Strategy links back to its primary data source. Every analysis is auditable and repeatable, giving you an evidence chain you can check – rather than a “thin” report built on potential hallucinations without a route to go deeper to understand the underlying facts.
That standard of analytical rigour used to be available only to institutions with the budget for dedicated research analysts. We think every research organisation should be able to access the best strategic intelligence.
What a decade of real analytical work looks like, encoded
Digital Science has spent over a decade doing this work for the research community by hand.
Our research analysts have delivered bespoke strategic intelligence for some of the world’s leading organisations: impact evaluations, research landscape analyses, strategic benchmarking, faculty recruitment intelligence. We have experience working with clients around the world to solve some of the hardest research analytics problems in the sector.
What Dimensions Research Strategy encodes is the thinking behind that work: not the outputs themselves, but the patterns and commonalities which surfaced over time, aligned with the accumulated judgment of domain experts who understand not just how to query data, but know the personality, caveats and – yes – even the failure-points of that data intimately.
The result is not a large language model trained on public text and pointed, unseeing, at a research question. Nor has any of the analysis our clients have received been used in training an LLM. It is AI built from the inside out, from a decade of real analytical practice in the research intelligence domain, by the people who have been doing this work for the world’s leading research organizations.
Research strategy doesn’t fit in easy categories
We also designed Dimensions Research Strategy to solve a frustration that comes up in every conversation we have with the research community. Existing tools impose their own topic structures: fixed classification systems that don’t map to how institutions actually think about their research.
Rather than asking institutions to work within our categories, Dimensions Research Strategy lets you define the research areas that matter to you, and reasons from there. The analytical framework bends to the question, not the other way around.
The underlying data matter too – and it’s these data that are Digital Science’s heritage. Dimensions Research Strategy is built on Dimensions and Altmetric data combined: publications linked to grants, patents, clinical trials, policy documents, and real-world attention signals.
That means:
- 165M+ publications
- 41M+ researcher profiles
- 8M+ grants
- 256m+ mentions of over 24 million research outputs
- 180M+ patents
- 5bn+ connections between research objects
Of course, data means nothing without the connections between it. When you are asking complex questions that will drive critical decisions across research strategy, those connections are where the insight lives. Dimensions and Altmetric bring together more connections in a single, verified database than any other provider – so whether you’re asking about the impact of your research (connections from publications to patents, clinical trials, and policy) or the efficacy of your funded research (connections from grants, through publications to impact), the full picture is there. No other system brings those connections together at this scale and with this level of verification, in one centralized platform.
So when you ask Dimensions Research Strategy how your institution’s translational medicine capability compares to your peer group, it’s drawing on a picture that includes not just what researchers have published, but what they’ve been funded to do, what’s made it into clinical trials, what’s been cited in government policy, and what’s generating public attention. The whole research landscape, navigated with context.
Until now, assembling that picture – across publications, grants, clinical trials, patents, and policy documents and all the links between them – for a single strategic question can take a skilled analyst weeks, months or not even be possible. Dimensions Research Strategy aims to change that so that anyone can create reproducible, interrogable analyses which an organization can leverage with confidence – because they can see every step.
What it will do
At launch, Dimensions Research Strategy will deliver two workflows designed for senior research leaders.
- The first is Benchmarking.
What we’ve heard from early development partners is that the most important feature isn’t the comparison itself – it’s being able to define the research areas that matter to your institution, rather than working within the fixed topic structures that existing tools impose. So rather than measuring your institution against someone else’s metrics, Research Strategy lets senior leaders define their own peer group – the institutions they actually compete with – and run a multi-dimensional comparison across research output, citation impact, collaboration reach, funding competitiveness, and societal engagement. The analysis is interrogable and iterative: ask a follow-up question, adjust the comparison, drill into a specific discipline. The value isn’t in knowing where you rank in absolute terms; it’s in understanding your relative position against the institutions that actually matter to your strategy.
- The second is Expertise.
Understanding who is working in a specific research area – globally, at peer institutions, or within your own institution – is a question that comes up constantly in strategic planning, and one that currently takes days of manual work to answer with any confidence. Dimensions Research Strategy surfaces that picture in a single session. Organizations can map researcher expertise across output profile, collaboration network, field standing, and real-world impact, to understand the full landscape of who is working in areas that matter to their strategy, and identify expertise that complements their existing capabilities. The insights inform the process, while the decisions remain with the humans who hold the true strategic judgment.
These are the first two workflows on our roadmap – but they are far from the last. Research impact, funding intelligence, grants preparation, research integrity and security: each of these is a challenge we hear raised by the research community, and each is a space we can see tremendous opportunities in. The sequence and the timing of these will be shaped by ongoing conversations with our Development Partners and Early Adopters. But the direction is clear: a platform that gives every research institution the strategic intelligence it needs, across every dimension that matters.
If you’re interested in becoming an Early Adopter, sign up to our launch newsletter here.
Do we see this replacing analysts?
In a word: No.
Every organization we’ve spoken with so far has a highly capable analyst working at the centre of this – and in every case, the bottleneck isn’t skill. It’s capacity.
In the same way that this blog was written in collaboration between human and AI, we see the role of the human analyst as evolving from data wrangler to strategic partner. The human co-author of this article has spent the time allocated to this task editing, shaping, sense-checking, cross-interrogating. Her role has changed, but remains essential to the success of the task.
Whether you’re a head of department, research manager, or librarian taking your first steps in research analytics, or a team of experienced analysts looking to go further and faster – Dimensions Research Strategy is built for both. In the first case, it puts rigorous, evidence-backed analysis directly in your hands, built on a decade of expertise you wouldn’t otherwise have access to. In the second, it removes the weeks of manual work that stand between a question and an answer, freeing your team to focus on what only they can do.
With Dimensions Research Strategy, the analysis that used to take a week can now take minutes. What you do with the time that frees up becomes the next interesting question.
Strategic expertise across every dimension that matters. On demand, for every institution.
Stay informed – sign up to our launch newsletter
If you’d like to stay informed about availability and early access opportunities, sign up here.
The post Research strategy deserves better than a chatbot: Why Dimensions Research Strategy is different from other AI tools in research appeared first on Digital Science.
from Digital Science https://ift.tt/a8Wkuf6
