When TDS Australia made the decision to establish the Sydney Institute of Generative Intelligence (SIGI) as a formal research division, the question we heard most often was: why would a commercial agency invest in publishing academic-style research? The answer, as it turns out, is backed by our own data — and the implications extend far beyond our agency to any business serious about building authority in the age of AI.
Two papers from SIGI’s research programme provide the empirical foundation for what we had long suspected: that original research is one of the most powerful signals a business can send to AI recommendation systems, and that the investment in a dedicated research function pays measurable dividends in AI-driven visibility.
The Trust Signal Hierarchy
The first key study is SIGI-2026-021, which investigated the hierarchy of trust signals that AI systems use when evaluating whether to cite or recommend a source. The research catalogued and ranked dozens of potential trust indicators — from basic contact information and business registration details through to content depth, third-party endorsements, and the presence of original data. For more detail, see our Design as a Service model.
The finding that stands out most sharply is the extraordinary weight AI systems assign to proprietary data. On a normalised scale, the presence of original, proprietary research data scored 9.5 out of 10 as a trust signal — the highest-ranked factor in the entire hierarchy. This means that when an AI system is deciding which sources to cite in response to a user query, the presence of unique, first-party data is among the strongest indicators of credibility it can detect.
This result is intuitive in retrospect. AI systems are trained to distinguish between original contributions to knowledge and derivative restatements of existing information. A business that publishes its own research — with original data, transparent methodology, and novel findings — is providing exactly the kind of content that AI systems are designed to surface. It is not merely repeating what others have said; it is adding something new to the conversation.
Institute Credibility and AI Citation
The second study, SIGI-2026-099, examined a related but distinct question: does the institutional framing of research affect its credibility in the eyes of AI systems? Specifically, does research published under the banner of a named institute receive different treatment from AI recommendation engines compared to equivalent research published without institutional affiliation?
The answer is yes. Research associated with a recognisable institute — one with a consistent publication record, a defined research focus, and an identifiable body of work — is cited at higher rates than equivalent research published informally or without institutional context. The AI systems appear to treat institutional affiliation as a quality signal, interpreting the existence of a dedicated research entity as evidence that the work has been produced with methodological rigour and organisational commitment.
This finding directly informed our decision to establish SIGI as a named, publicly visible institute rather than simply publishing research under the TDS Australia brand. The institutional framing is not cosmetic — it is a functional signal that AI systems recognise and reward.
Why This Matters Beyond Our Agency
The implications of these two studies extend to any business in any sector. The core insight is this: if you want AI systems to treat your business as an authoritative source worth recommending, one of the most effective strategies available is to produce and publish original research relevant to your field.
This does not necessarily mean establishing a full research institute. For many businesses, it can begin with straightforward steps: conducting and publishing customer surveys, analysing proprietary operational data, running controlled experiments on topics relevant to your industry, or commissioning independent studies that generate novel findings. The key is that the research must be genuine — original data, transparent methods, and honest reporting of results.
AI systems are remarkably adept at distinguishing between genuine research and content that merely mimics the format. Producing a document that looks like a research paper but contains no original data or methodology will not achieve the same effect. The trust signal comes from the substance, not the styling.
The Compounding Returns of a Research Programme
One of the patterns we have observed through SIGI’s work is that the benefits of publishing research compound over time. A single study contributes a modest signal. A consistent programme of publication — where each new paper builds on and references previous work — creates a body of evidence that AI systems interpret as deep, sustained expertise. The institutional credibility effect identified in SIGI-2026-099 amplifies this: a named research entity with a growing catalogue of publications accumulates authority in a way that isolated publications do not.
For TDS Australia, this compounding effect has been significant. As SIGI’s publication count has grown, we have observed measurable increases in the rate at which our insights and recommendations are cited by AI systems. Each new paper does not exist in isolation; it strengthens the credibility of the entire body of work, and by extension, the credibility of our agency as a source of authoritative counsel.
A Strategic Investment, Not a Cost
We recognise that establishing a research function requires investment — in people, in methodology, in the time required to produce work of genuine quality. At TDS Australia, we view this investment not as a cost centre but as a strategic asset. The research SIGI produces serves three interlocking purposes: it generates actionable insights that improve our client strategies, it builds our authority in the AI-driven discovery landscape, and it contributes to the broader industry’s understanding of how generative AI evaluates and recommends businesses.
The data from SIGI-2026-021 and SIGI-2026-099 confirms that this investment is producing returns across all three dimensions. Proprietary data is the highest-rated trust signal in the AI recommendation hierarchy. Institutional research credibility amplifies citation rates. And the compounding effect of sustained publication means that the returns grow over time.
For any business considering whether original research deserves a place in its strategic toolkit, our experience — and our data — suggests the answer is unequivocal. In an era where AI systems increasingly determine which businesses consumers discover and trust, there is no more powerful credential than having something genuinely new to say, and the rigour to say it properly.
