About Bibliomatica

The hallucination was the architecture

A language model is built to produce plausible sentences, not to count. Asking it "how many times was this paper cited?" is asking a novelist to audit your accounts. It will always give you a beautiful answer, and it will sometimes be fiction.

The failure everyone blamed on "AI not being ready" was really the absence of a hard wall between the tool that reasons in language and the tool that computes the math.

Then I added up the months we were losing

Before anyone could get to analysis, they burned weeks merging Web of Science, Scopus, Lens, and Dimensions exports that don't agree on a single column. Hand-reconciling "J. Smith," "John A. Smith," and "Smith, J." into one author 400 times over. Re-running VOSviewer until the clusters looked right, with no record of the settings that produced them.

Half my students stalled here. Some never finished. It wasn't the thinking that broke them, it was the plumbing.

The two tools that could have saved them didn't speak to each other

Bibliometrix and VOSviewer compute rigorously but say nothing in plain language and punish you with an R learning curve. The chatbot writes beautifully but makes the numbers up.

Every researcher was ferrying data by hand between a tool that was correct but mute and a tool that was fluent but unreliable, absorbing every transcription error in between.

So I built the wall, and made it the whole product

In Bibliomatica, deterministic software (Bibliometrix, OpenAlex, ROR, ORCID, VOSviewer) computes every number. AI is allowed to do only one thing: write the narrative around verified values. Never a digit of its own.

Each figure the text cites is a placeholder bound to a real cell in a real CSV. Anything unverified is blocked before it can reach the page.

Seven disciplined phases carry a researcher from "I want to do a bibliometric study" to a submission-ready manuscript, where every claim is traceable, every step reproducible, and no number is ever guessed.

Who I am

I am Adam Linoby, a Senior Lecturer at Universiti Teknologi MARA. Before Bibliomatica, I spent years teaching, supervising postgraduate research, leading research teams, securing grants, and presenting work nationally and internationally.

My academic foundation spans sport science, physiology, biochemistry, human performance, and health technology, including a PhD pathway in Applied Human Physiology and Biochemistry, an MSc from Bond University, and earlier UiTM degrees.

I work inside the research pipeline. Systematic reviews, meta-analyses, intervention studies, validation studies, grant-funded projects, manuscript preparation, indexed publication. Scopus and Web of Science publications, reviewer roles, Cochrane manuscript translation, research grant review, postgraduate supervision awards, invited expert panels.

I also build. Patented and award-winning research, health, and exercise technologies. Mobile applications. AI-for-academia workshops. AI-driven startups aimed at academic and health outcomes.

Bibliomatica is what happens when someone who lives inside the research pipeline gets tired of watching students lose months to plumbing.

The umbrella

Bibliomatica is one product inside Workcyte, an AI-powered digital academy that builds tools, training, coaching, and IT solutions for academic productivity. If Bibliomatica handles the guided bibliometric-study pathway, Workcyte handles the broader ecosystem: courses, AI training, one-to-one coaching, e-guides, micro-apps, and consulting.

Workcyte exists because researchers need more than one app. Bibliomatica is the first specialized product inside a larger mission: helping academics move from research questions to publishable work through verified software, guided learning, personal coaching, and applied AI solutions.

The promise

Bibliomatica exists to give every researcher a bibliometric study they can defend line by line. I hand the language work to AI and the arithmetic to deterministic science, so citation counts, networks, and metrics come from verified computation, never a model's guess. I carry you from a first question to a submission-ready manuscript through seven reproducible phases.

My promise isn't that AI will do your thinking. It's that it will never fabricate your evidence.