For researchers and scientists

The AI synthesises your field every day. Is your work in the synthesis?

Journalists, committees, companies and funding agencies use ChatGPT, Gemini and Claude to find out who the reference on a topic is. If your entity is not disambiguated, your work may be credited to someone else — or to no one.

The decision moved

The literature review of whoever decides now starts in a chat, not in an indexed database.

Whoever is looking for an expert rarely opens a scientific database. They ask the artificial intelligence — and get a synthesis with very few names in it.

Example of a question to the AIs
"Who are the leading researchers on [topic] today?"
— A journalist looking for a source
Example of a question to the AIs
"What is the origin of concept [X]? Who proposed it first?"
— The attribution of your discovery, in dispute
Example of a question to the AIs
"What has science established about [topic]?"
— A company deciding who to invite to its scientific board
The cost of being invisible to the AIs

Being on Google is not the same as being in the AI’s answer.

You may have a respectable h-index, papers in first-tier journals and a consolidated line of research — and the artificial intelligence may still fail to tie any of it to your name. A namesake, an old affiliation or a disconnected identifier is enough for your work to appear without an owner. The scientific community knows who you are. The language model has no way of knowing.

One namesake
is enough for the AI to credit your work to someone else, or split it between two entities.
Without the citation
is how your finding reaches whoever asked, if the model cannot tie it to you.
A few names
is what the press and the committees get when they ask who the reference in your topic is.
The solution

Sumaúma AI Presence makes you a researcher that the AIs recognise and cite.

This is not about publishing more. It is about making your entity legible to the machine — name, ORCID, affiliations, co-authorships, line of research and the concepts you proposed — so that language models can tie every result back to its origin. That is what we call Presence Engineering for Generative AIs, applied to your scientific authority.

Pillar 01

We monitor

We measure your G-SoV (Generative Share of Voice) with AI-Scan: whether the AIs cite you in the topics of your research line, in what position, and whether they are crediting your work correctly.

Pillar 02

We structure

We build your knowledge graph and tie together the nodes of your identity — ORCID, Wikidata, affiliations and co-authorships — so the AIs recognise a single entity and do not split your body of work between namesakes.

Pillar 03

We clone

We extract your Linguistic DNA and vectorise the concepts you proposed, so the AIs know what originated with you and credit it to you.

Pillar 04

We propagate

We publish consistent, original semantic content across authority and science-communication channels — because the model learns from accessible text, not only from the paper behind the paywall.

+100
professionals served since 2022
We come from the market of those who live off their own name

Sumaúma started in 2022 as a digital marketing agency and had — and still has — the segment of professionals whose authorial work is the asset as its main vertical. We have served more than 100 of them. Today we apply that experience to the new frontier: what the generative AIs answer when someone asks who the reference in your field is.

Frequently asked questions

Questions answered.

I am a researcher. How do I get the AIs to cite my work?

The first step is to measure. AI-Scan shows whether the AIs already associate you with your research line, in what position, and whether they are crediting your results correctly. From there, we disambiguate your entity in the knowledge graphs — tying ORCID, Wikidata, affiliations and co-authorships into a single node — so the model stops splitting your body of work; we vectorise the concepts you proposed, so attribution travels with the idea; and we propagate accessible content across the channels the AIs are trained on. There is no shortcut and no space for sale: the AIs cite whoever they can recognise.

The AI is confusing my work with a namesake’s. How is that fixed?

With entity disambiguation, which is exactly the Graph Structuring pillar. The model gets confused because it has no stable identifier tying name, affiliation and body of work together. Once your ORCID, your Wikidata record and your institutional pages point coherently to the same entity, and that entity is marked up in structured data, the model finally has a single node to anchor your publications to.

Does this replace publishing in indexed journals?

Not at all — and nothing does. The peer-reviewed paper remains the foundation of your scientific authority. Presence Engineering takes care of what happens next: whether that authority reaches the answer a journalist, a committee or a company gets when they ask the AI who the reference on your topic is.

Does this interfere with the integrity of my scientific output?

No. We do not write, do not co-author and do not alter any result. The work is metadata and semantic structure: making sure the work you already published is correctly credited to you. Nothing we do touches the scientific merit of what you produce.

How long until my work starts being cited correctly in the AIs?

Entity disambiguation is usually the fastest change to feel: as soon as ORCID, Wikidata and affiliations point to a single node, the model stops splitting your body of work. AI-Scan gives you the immediate diagnosis, and propagation across accessible channels moves your G-SoV over the following weeks, with continuous monitoring. It is not a button we switch on or off.

Secure the correct attribution of your work in generative AIs with Sumaúma AI Presence.

Request your AI-Scan and find out how the AIs are crediting your scientific work today.