Congratulations on this prestigious recognition for your outstanding research and innovative contributions. Your achievement is a testament to your dedication and excellence.
Plastic and chemical waste accumulation presents significant environmental challenges worldwide. Polycarbonate materials and organophosphate ester compounds are widely used in industries, creating substantial waste streams that require sustainable management solutions.
Synergistic co-upcycling utilizes chemically complementary reactions to transform these waste materials into valuable products. This innovative approach maximizes resource recovery, reduces landfill disposal, and promotes a more efficient circular economy.
By converting waste into useful chemicals and materials, co-upcycling supports environmental sustainability and industrial innovation. Such advanced recycling strategies help reduce pollution, conserve resources, and contribute to greener manufacturing practices globally.
Xiaoguang Huang is affiliated with Guangzhou Baiyunshan Tianxin Pharmaceutical Co., Ltd in China and contributes to research activities within Pharmacology. His academic profile demonstrates participation in scientific publications, citation visibility, and pharmaceutical research engagement. These scholarly activities support recognition within the Global Scholar Awards and reflect continued involvement in pharmacological and healthcare-related scientific advancement.
Global Scholar Awards π
The race to deploy AI agents is accelerating, but most organizations are still building on sand. A new Gartner report suggests that the key to building reliable AI agents is a “context layer”.
According to Gartner’s latest research, 42% of enterprises plan to deploy AI agents by the end of 2026, and AI agent spending is expected to grow from 22% to 31% of total AI budgets in just one year1. Despite this wave of investment, only one in five organizations report that their GenAI tools are delivering significant value. Hallucinations, limited impact, and unpredictable behavior remain stubbornly common.
The problem, Gartner argues, isn’t the models, but rather what surrounds them.
The Missing Layer
Behind every reliable AI agent is something Gartner now calls a “context layer”— a dedicated architectural component that curates, organizes, and delivers the knowledge an agent needs to act intelligently. Without it, agents are left processing noisy, poorly prioritized data, making expensive errors and producing outputs that can’t be trusted or traced.
Gartner is unambiguous about the stakes: by 2027, organizations that prioritize semantics in AI-ready data could increase their agentic AI accuracy by up to 80% and reduce costs by up to 60%. The context layer is no longer an optional refinement — it is the necessary foundation.
And yet this layer cannot simply be purchased. No vendor offers it out of the box. It must be engineered, assembled from services, capabilities, and custom modeling that together transform an organization’s tacit knowledge into something AI agents can actually use.
Three Components, One Foundation
As stated in the report, there are three interlocking components that make up this ‘context’ layer: semantics, operational state, and provenance. Together, they form a pipeline that allows agents to retrieve the right information, organize it coherently, and act on it with accountability.
Semantics: Meaning, Not Just Data
Semantics is the component most organizations are missing, despite it being the one with the greatest leverage. Gartner finds that organizations implementing semantic modelling such as ontologies and knowledge graphs, are 2.2 times more likely to achieve high effectiveness in AI data engineering, however, only 40% of organizations have done so.
Semantics means representing your organization’s knowledge—business entities, rules, policies, relationships, metrics—in machine-readable form. This allows AI agents to interpret what something means in context and execute an action based on that context, not just pattern-match on keywords. Without this layer, even the most sophisticated agent is, in effect, guessing.
This is precisely the domain where metaphactory brings long-standing proven capability. metaphactory by metaphacts, a Digital Science solution, is a knowledge graph platform enabling organizations to build and maintain rich semantic models for over a decade—connecting business glossaries, ontologies, and data products in ways that AI agents can directly leverage. For organizations serious about agentic AI, a robust semantic foundation isn’t a future aspiration; it is a prerequisite.
Operational State: The Right Information at the Right Time
While semantics provides meaning, your operational state provides situational awareness. AI agents need access to current, accurate information about the entities and processes they’re acting on beyond just snapshots, such as up-to-date information on customers, datasets, experiments, publications and suppliers.
For research-intensive organizations, this is particularly acute. The ‘operational state’ of a research environment spans live datasets, ongoing experiments, researcher expertise, institutional repositories, and the evolving landscape of published science. Digital Science’s portfolio—including Dimensions, Altmetric, and Figshare—represents exactly this kind of curated, continuously updated operational knowledge. Rather than building this knowledge from scratch, organizations working in research and innovation already have access to a pre-assembled foundation.
Gartner also highlights the Model Context Protocol (MCP) as the emerging standard for connecting agents to operational state efficiently and securely. Dimensions, Altmetric, and metaphactory already support MCP, reflecting a broader conviction that research infrastructure should be designed to meet agents where they are, not retrofitted after the fact. As adoption of the protocol grows across the industry, having well-structured knowledge accessible through it will matter more, not less.
Provenance: Trust Through Traceability
The third component—provenance—is what makes agentic AI governable. It encompasses the systematic tracking of data lineage, agent decisions, actions, outcomes, and feedback across the full lifecycle of AI operations.
For research organizations, publishers, and funders, provenance isn’t merely a governance checkbox. It is central to the integrity of the work itself. Reproducibility, accountability, and the ability to audit AI-assisted conclusions are not simply peripheral concerns; they are defining ones. Gartner notes that 74% of organizations recognize that data governance tools are essential to operationalizing AI governance, yet robust provenance mechanisms remain rare in practice.
Digital Science’s longstanding commitment to open, traceable research infrastructure, including persistent identifiers, transparent data lineage and open metadata, gives research organizations a natural head start on this component. The challenge is connecting these capabilities explicitly into the agentic architecture, so that every AI-assisted decision can be traced back to its sources and reviewed.
Research Intelligence as a Context Layer
There is a broader framing worth making explicit here: for organizations operating in research, science, and innovation, the context layer is not merely a technical architecture problem. It is, at its core, a research intelligence problem.
The tacit knowledge Gartner describes—the organizational understanding that must be made machine-readable for AI agents to function—is, in a research context, the accumulated intelligence of a scientific community: what has been discovered, by whom, with what methods, validated how, and applied where.
We have spent over a decade building infrastructure that captures precisely this kind of knowledge at scale. The shift to agentic AI doesn’t make that infrastructure less relevant—it makes it more so. The question is no longer just “can researchers find the right information?” but “can AI agents, acting on researchers’ behalf, find, interpret, and act on that information reliably and accountably?”
The answer depends entirely on the quality of the context layer underneath.
What This Means in Practice
For R&D leaders and data and analytics leaders, the practical implication is this: before asking which AI agent to deploy, ask what context layer you have in place to support it. Gartner’s advice is to start with high-value use cases rather than attempting a comprehensive build all at once—iterate, demonstrate outcomes, and expand. That is sound counsel. But iteration without a semantic foundation, without right-time data access, and without provenance mechanisms will simply produce faster failures.
The organizations that will lead in agentic AI are not those that move fastest to deploy agents. It is the organizations that invest earliest in the knowledge infrastructure that make agents worth deploying.
Digital Science is working with research organizations and data-intensive enterprises to build the context layers their AI strategies require.
In the first blog in this series, we explored engagement and impact readiness for the Research Excellence Framework (REF) 2029. Here, we turn to the second element of assessment: Contribution to Knowledge & Understanding, and what it takes to approach it with evidence and confidence.
Contribution to Knowledge & Understanding (CKU) sits at the core of REF assessment. For institutions preparing submissions, the task is not simply to present strong individual outputs, but to show how research collectively advances knowledge within and across disciplines and how that work was enabled and supported within the institution’s research environment.
As REF 2029 approaches, most institutions will find that they are not short of high-quality research. The task they now have is to present that research as a coherent, representative, and defensible account of contribution, traceable back to the people, grants, and infrastructure that enabled it.
That distinction matters more than it might first appear.
Beyond completeness
It is tempting to frame CKU readiness as a data completeness problem. If all outputs are captured, the argument goes, selection can proceed with confidence.
But completeness is not the same as representation.
REF panels assess whether a submitted body of work reflects the range and diversity of a unit’s research activity, not simply whether a record exists for every output. A dataset can be complete and still produce a submission that is narrow, uneven, or poorly contextualised.
“The challenge for universities is not simply about capturing and submitting quality research outputs to REF, it is about demonstrating the full diversity and breadth of the research outputs. In choosing which outputs to submit, universities are expected to demonstrate the diverse range of staff contributing to the outputs; the diverse range of disciplines, research methods and output types whilst also ensuring that contributions from inter- and multi-disciplinary collaborations are represented,” says Natalie Dallat, Head of Research Performance, Ulster University.
This distinction has practical implications. In a decoupled framework, where submitted outputs do not need to be linked to specific individuals, institutions still need to demonstrate a substantive connection between research and the environment that enabled it. That requires not just complete records, but well-contextualised ones.
Three areas of risk are worth examining in turn.
Output visibility: what institutions know and what they can prove
In practice, most significant outputs are already known to institutions. Academic workflows, open access deposit requirements, and internal review processes mean that the majority of relevant publications are captured somewhere.
The more common challenge is not absence but unevenness, gaps in coverage that accumulate over time through staff mobility, inconsistent author affiliations, publications linked to grants but not captured locally, and interdisciplinary outputs that fall between Units of Assessment (UoA).
Figure 1: University of Oxford, all publications 2021-present
These are rarely major gaps in institutional systems. But in aggregate, they can affect the completeness and credibility of a submission, particularly in disciplines where research activity may be systematically underrepresented relative to its actual volume.
Addressing this requires two complementary layers. Research information systems such as Symplectic Elements provide structured output capture, validation workflows, and linkage between researchers, publications and grants, creating the audit trail that REF governance demands. An independent, interconnected data layer such as Dimensions then enables cross-checking: surfacing missing outputs, highlighting metadata discrepancies, and providing a broader view of publication activity beyond local records.
“What Dimensions allows institutions to do is essentially hold a mirror up to their own systems. Not to replace internal records, but to ask: is what we’re seeing internally representative of what’s actually out there? For some disciplines or research groups, that comparison can be revealing,” explains Ann Campbell, Director Research Impact & Comparative Analytics at Digital Science.
Together, structured capture and independent validation strengthen confidence in completeness before output selection begins and provide a more defensible evidence base for the decisions that follow.
Understanding performance within fields
Once institutions have confidence in the completeness of their records, a second challenge emerges: interpreting performance in a way that is fair and defensible across disciplines.
Raw citation counts rarely tell the full story. Citation norms vary significantly across fields; what constitutes a well-cited output in a fast-moving biomedical discipline looks very different from the equivalent in history or architecture. A paper with 20 citations might be considered relatively modest in one field, but well above average in another.
While output selection is typically led by discipline experts within UoA, decisions are often informed by broader portfolios and mixed indicators. Without appropriate field-level contextualisation, there may be tendency to overvalue some outputs that align with readily interpretable patterns of performance (i.e., citation counts) and undervalue others particularly where interdisciplinary research is involved. This can have consequences both for selection and for the narrative presented to panels.
The scale of this variation is visible in the data. Across UK institutions, raw citation counts for outputs in Units of Assessment such as Clinical Medicine or Physics far exceed those in disciplines like History or Art & Design, and yet when performance is measured relative to field norms, the picture shifts substantially. Units that appear modest on raw citations often demonstrate strong or above-average relative contribution when field-normalised indicators are applied. For institutions making selection decisions across multiple UoAs, this difference is not academic: it directly affects which outputs are recognised as genuinely competitive, and which risk being undervalued simply because they sit in lower-citation disciplines.
Figure 2: Average citation counts vary substantially across Units of Assessment, while field-normalised performance (FCR) highlights strong relative contribution in disciplines where raw citation accumulation may be lower. *
* Average citation counts and field-normalised citation performance across REF Units of Assessment (2014–2021, UK Institutions, articles only using Dimensions UoA Classification) Figure 3: Average Citation Count per Publication
Field-normalised indicators and disciplinary benchmarking support a more accurate and defensible reading of performance. Dimensions enables field-normalised citation analysis, benchmarking against peer institutions, collaboration pattern analysis, and trend tracking across time.
Peer review remains central to CKU assessment. But contextual data helps institutions approach that peer review better prepared with a clearer sense of where their research sits within its field, and a stronger basis for the interpretive narrative they are expected to provide.
From individual outputs to coherent thematic narratives
CKU submissions are strongest when outputs form a coherent intellectual narrative. Panels respond to thematic depth and sustained advancement of knowledge not isolated high-performing items, however well-cited they may be.
That makes output selection a genuinely strategic exercise and the scale of the choices involved is considerable. Analysis of REF21 submission patterns shows that the typical institution produced eligible research across 33 of the 34 UoA, but submitted to just 20. In nearly one in five cases where an institution had a meaningful body of research within a UoA, that UoA received no submission at all. Even within the UoAs that institutions chose to submit, the median coverage rate was under 7%.
Figure 4: Research breadth vs submission breadth
The submitted profile, in other words, represents a deliberately selective slice of a much broader underlying research base. That selectivity is appropriate as REF rewards quality over volume, and strategic narrowing is both permitted and expected. But it means the submitted body of work must tell a coherent story about where an institution’s research genuinely lies. Getting that story right requires a clear view of the full landscape: understanding where depth is concentrated, where disciplines connect, and where gaps might undermine the coherence of what is presented to panels.
Thematic clustering and citation network analysis can help identify areas of concentrated strength and the interdisciplinary bridges that connect them. These analytical approaches surface patterns that may not be visible when outputs are reviewed individually, and support the kind of coherent story that distinguishes a strong CKU submission.
That coherent story, however, increasingly needs to account for more than publications alone. As REF increasingly recognises diverse outputs, datasets, code, preprints, and other research artefacts alongside traditional publications, institutions also need infrastructure that makes that breadth visible and accessible.
The evidence from REF21 illustrates how far there is still to go: of the 4,000 non-traditional outputs submitted, almost three quarters had unknown or unresolvable locations, and only 244 had DOIs. REF21 Main Panel D assessors noted the wide variety, inconsistent quality and uneven preservation of practice-based outputs with many hosted on fragile, short-lived platforms that were difficult to navigate.
Platforms such as Figshare support persistent access, DOI assignment and the presentation of these materials as part of a coherent research record, ensuring that the full range of contribution is available for assessment.
Scholarly visibility: useful context, not a proxy for contribution
While CKU is fundamentally about intellectual contribution, the broader circulation of research can provide supplementary context. Where outputs are being cited in policy documents, taken up in professional practice, or discussed in specialist communities, those signals can help situate the reach of a body of work, particularly in applied or interdisciplinary fields where impact pathways are diverse.
Altmetric can surface where outputs are being referenced beyond traditional citation indexes, from policy and clinical guidelines to media and public discourse. These signals do not measure contribution to knowledge and understanding, and should not be presented as a substitute for bibliometric evidence or peer judgement. But as additional context, they can help round out the picture, particularly for outputs whose significance may not be fully reflected in citation metrics alone.
The important distinction is that scholarly visibility supports interpretation. It does not replace it.
From reactive selection to confident CKU readiness
CKU readiness is about planning, not last-minute correction. Institutions that approach it most effectively don’t wait until selection is imminent. They build the evidence base over time, ensuring completeness, contextualising performance, and constructing the thematic narrative that panels expect to see.
“What we often see is that institutions feel more confident in REF preparation when they’ve been building the picture gradually over time. It becomes easier to understand where strengths are emerging, how research sits within its field, and how to present that contribution coherently,” says Campbell.
REF readiness is about leading, not lagging. For CKU, that means investing in the evidence and infrastructure and contextual understanding that supports selection throughout the cycle.
Institutions preparing for REF29 are increasingly focusing on areas such as:
Together, these form the building blocks of a CKU submission that is traceable, representative, and defensible.
Digital Science supports this readiness through interconnected solutions that strengthen evidence and decision-making, while leaving judgement firmly with institutions and REF panels.
Whether you want to audit output visibility and identify gaps in your publication record, benchmark your CKU evidence within disciplinary context, or map the thematic strengths that will anchor your submission narrative, Digital Science can help.
.png)
