Did you know that there's an activity within Library Science that helps guide a search in your preferred AI agent, avoiding errors, ambiguity, and misunderstandings ? This activity is called Reference Interview. In this article, I'll show you what this activity is, how it works, and how to use it (in my Google Gemini ) to conduct a well-founded search on any subject.

But first, let's understand why creating a structured search is the right approach.

How do I structure advanced searches with Gemini?

The search for information is constantly and rapidly accelerating, and this is nothing new. We went from simple text boxes where we entered words, hoping that a rudimentary algorithm would return a list of links Artificial Intelligence models . But this evolution is not smooth and brings us a series of challenges.

Even with tools like Google Gemini, Claude , and ChatGPT opening up almost endless possibilities for accessing and generating information, it's necessary to know how to guide these models to do what we really want them to do.

“It is important that we are not timid in incorporating computerized sources into the routine of the reference service. […] We should not view the use of computerized sources as if it were a 'betrayal'."

Elizabeth Bramm Dunn, 1988

I've been using Gemini's deep research functionality to help me build Domain Analysis , the first step in the Semantic Workflow It 's interesting to note that this new technological frontier, instead of making classic practices obsolete, rescues and amplifies the need for one of the most fundamental skills in my beloved Library Science: the reference interview .

This article, as previously stated, aims to build a bridge between the reference interview, traditionally conducted by librarians, and modern interaction strategies with generative AIs, primarily in the development of initial research on any subject you need to know about.

We will demonstrate how the principles that guide human interaction in the pursuit of knowledge are not only relevant, but fundamental to extracting the maximum potential from tools like Gemini. After providing the necessary background, I will demonstrate, in a video, how I do this in practice.

I'm going to introduce you to two working models: Kipling's six honest creations and Grogan's eight steps. These two frameworks have successfully guided searches for decades in libraries around the world. You can use both to make your preferred model your librarian . We start with Kipling, who shows us a very practical process that is highly adaptable to working with AI, and we end with Grogan, who gives us a more holistic view of the process, important for taking our work here to another level of quality.

The Reference Interview in essence

We need to understand a fundamental difference between two concepts that are important to our article: " referral service " and " referral process ".

The service is the concrete action of the professional, the direct interaction. The process , in turn, is the user's complete journey, including the internal and cognitive phases that precede and follow contact with the specialist. Understanding this journey is what allows us, as SEO , to intervene more effectively.

But before we delve into applications in agents, we need to revisit what constitutes reference service and, at its core, the interview. Considered by many to be the heart of library science, reference is a complex and multifaceted activity. The librarian, in this context , acts as a mediator of knowledge, a professional who employs methods and tools to guide the user through the vast informational universe, helping them find answers that they themselves often wouldn't know how to find.

The reference interview is the cornerstone of this process. It can be defined as an interactive transaction in which the information professional asks one or more questions to the person needing information, with the goal of identifying, clarifying, and refining the subject of their inquiry.

Classic library science studies demonstrate that a significant proportion of incorrect or inadequate responses, and the resulting user dissatisfaction, originate from communication failures during the interaction between an information specialist and the person needing that information.

"Paradoxically, the introduction of machines into the reference process will force even the most reluctant reference librarian to act like a human being interacting with another."

William A. Katz, 1978

The central problem is that the user often cannot precisely express what they need or, in more complex cases, is not even sure what they are looking for. This first question the user asks is called the Initial Question, and it is only the tip of the iceberg of a deeper informational need. This is where the interview becomes a crucial tool, especially for two groups of people:

1. Those who know what they need, but are unable to adequately articulate their need.
2. Those who are unsure about what they really need to solve their problem.

For this second group, the interview is not merely useful; it is essential . Ignoring it is the shortest path to an inefficient search and frustrating results. This is in the classic library .

But how can this happen in interaction with AI agents? Is it possible to use knowledge of Library Science as a guide to achieve better results in research using your preferred model? I'll tell you right now: Yes, it is. I myself have been doing this for months in a row with excellent results.

Google Gemini as a Reference Librarian

The rise of AIs like Gemini forces us to rethink our relationship with search tools. We are no longer dealing with a page that answers searches. An agent can be understood as a "synthesis engine," a system capable of processing, contextualizing, and generating information in a dialogical way. In this new paradigm, AI assumes the role of the reference librarian, the professional responsible for conducting the reference interview and assembling our object of interest: a search strategy .

deep research function, Gemini can act as our personal, customized librarian; we just need to know how to guide it in creating a search strategy that suits our objective and information needs.

We can refine our search by adding context, delimiting scope, and exploring facets of knowledge to avoid a lengthy and tedious iterative process. We can do much more than simply "ask" the model; we can guide it, using knowledge from Library Science, so that the search comes out the way we need it to the first time.

And for that, the established techniques of Library Science offer an invaluable roadmap. Come with me and I'll introduce you to two frameworks for you to apply. First, we'll show the model that Kipling created, and then we'll move on to another classic: Grogan.

AI-powered research framework: Kipling's six honest servants.

Rudyard Kipling, in one of his poems, gifted us with a reminder of universal value:

I Keep Six Honest Serving Men

I keep six honest serving-men (They taught me all I knew); Their names are What and Why and When And How and Where and Who. I send them over land and sea, I send them east and west; But after they have worked for me, I give them all a rest. I let them rest from nine till five, For I am busy then, As well as breakfast, lunch, and tea, For they are hungry men. But different folk have different views; I know a person small— She keeps ten million serving-men, Who gets no rest at all! She sends 'em abroad on her own affairs, From the second she opens her eyes— One million Hows, two million Wheres, And seven million Whys! 

Rudyard Kipling in Rudyard Kipling, Just So Stories (1902)

In a free translation of the most important passage for us, we have:

"I keep six honest servants (They taught me everything I knew); Their names are: What, Why, When, How, Where, and Who."

These six questions form a robust and highly efficient framework for structuring a reference interview and can be directly adapted for formulating an in-depth search strategy with Gemini.

Let's walk through each of the questions, putting together a prompt for a very well-structured initial question that will save us a huge amount of time.

1. What?

This is the starting point of any search, and its goal is to clearly define the central topic of the investigation. It helps us focus on determining and disambiguating the subject we need to know. In terms of the Semantic Workflow, this is the Knowledge Domain. At this stage, the librarian seeks clarity on the terms used by the user. They can use dictionaries or encyclopedias to disambiguate terminology and use paraphrasing ("So, what you're looking for is...") to confirm understanding and encourage the user to elaborate.

Can you understand where I got the idea of ​​using vocabulary in SEO?

How do I apply it to Gemini?

• Role Definition (Persona): Start the prompt by assigning a role to the AI. Example: “Act as a senior researcher in molecular biology.” This adjusts the lexicon and depth level of the AI.
• Terminology and Disambiguation: If your topic involves polysemous terms (with multiple meanings), provide the context in the prompt itself. Example: “Review 'manga', the fruit (Mangifera indica), and not the part of clothing or the genre of Japanese comics.”
• Active paraphrasing: You can adopt the strategy of using AI to refine the scope of the search. After an initial response, which may be incomplete, ask: “Based on your response, refine the search to focus on the genetic aspects of pest resistance. Ignore the commercial aspects of the harvest.” This simulates the negotiation and refinement that occur in a real interview, but it is inefficient when dealing with a deep research model that takes many minutes to perform the search and consumes many environmental resources. I prefer to use another strategy, which I will explain later.
• Addressing incompleteness: If your own question is incomplete, use AI to expand on it. Example: “I’m researching 'neural networks'. What are the most important subtopics and application areas I should consider for comprehensive research?”

2. Why?

Understanding the purpose of the information is one of the key points for the relevance of the answer that will be delivered, and it helps us define the purpose and context of our research. The same topic can be approached in numerous ways, depending on its end use, so we need to understand how to answer this question.

The librarian asks about the purpose of the research. Is it for a school assignment? A scientific article? A business decision? Personal curiosity? The answer to this question completely changes the search strategy.

How do I apply it to Gemini?

Explicit Contextualization: Incorporate the purpose directly into the prompt. The difference in results will be noticeable . Explain to the agent why they need that research, and in what context it will be used. Let's go back to our previous example:

From: “ Do an analysis of 'mango', the fruit (Mangifera indica), and not the part of clothing or the genre of Japanese comics. ” To: “ I need to do research for my ninth-grade Biology class where I need to write a script for a video talking about all the uses of the mango fruit (Mangifera indica) in Brazilian food, and not the part of clothing or the genre of Japanese comics. ”

The context is now described in the initial prompt. But this is only the first part of the work; let's take it slow, we'll get there, step by step.

3. Who?

Closely linked to "Why?", the question "Who?" defines who the information is intended for, adjusting the tone, language, and depth, and helping us understand the profile of our audience. In a library interview, the librarian seeks to find out who the user is.

• What is your level of knowledge on the subject?
• Is he seeking the information for himself or for a third party (a boss, a client)?

Application in Gemini

Defining the Audience: Be explicit about the audience for the content that the AI ​​will generate. Let's go back to our example:

I am a ninth-grade student in Brazil and I need your help to do research for my Biology class. My classmates and I need to write a script for a video talking about all the uses of the mango fruit (Mangifera indica) in Brazilian cuisine, not just the clothing aspect or the Japanese comic book genre.

We've now given the model a bit more information to understand who we are. But let's move on.

4. How?

This question defines the "packaging" of the information. The output format is as important as the content itself and helps your chosen model generate the answer at the level of detail most appropriate to your objective, which in our case is to generate research to base a script for your video on.

Acting as a reference librarian, the professional asks whether the user needs "just an introduction," an "exhaustive search," bibliographic references, abstracts, or full texts. When working with an AI model or agent, you also need to specify what type of output you want, but in our case, we'll have a text report, so you don't need to worry about that.

Application in Gemini:

Format Specification: Give clear instructions regarding the level of detail required for the desired response. Since in our case it's research to generate a script for a Biology lesson video, we can opt for something more introductory, like this:

I am a ninth-grade student in Brazil and I need your help to conduct introductory research on the main uses of the mango fruit (Mangifera indica) in Brazilian food. This research will form the basis for creating a script for a video that my classmates and I will present in our Biology class.

5. When? and 6. Where?

I want to draw your attention to these two questions, which are grouped together on purpose. These two filters are crucial for narrowing down the search and dramatically increasing the relevance of the results, saving time and effort. The temporal, idiomatic, and geographical restrictions are the fine-tuning in the use of reference interview strategies in AI models.

In their work, librarians inquire about date or location limitations. Does the information need to be up-to-date? Does it refer to a specific region? Language limitations can also be defined here.

Application in Gemini:

Adding Restrictions: Integrate these limitations directly into your prompts so that the AI ​​can filter the universe of information. Since in our case it's a survey for high school students in Brazil, we can finalize our initial prompt like this:

I am a ninth-grade student in Brazil and I need your help to conduct introductory research on the main uses of the mango fruit (Mangifera indica) in Brazilian food. This research will form the basis for a video script that my classmates and I will present in our Biology class. Please limit the research to sources in Brazilian Portuguese, from Brazilian sources, and published between 2000 and 2025.

Great! Our initial search is ready, now all we have to do is go to Gemini (or your preferred model) and select the deep search function.

Access this link to read the research report generated by Gemini based on our prompt, which was built following the Kiplin model: Mango: The Queen of Fruits on Brazilian Plates . And in the image below you can see a screenshot with some of the references used by Gemini to generate the report.

They will be our source for creating a manga specialist agent, which I will show you later.

Techniques for interacting with AI

Before we talk about Grogan's model, we need to go through an interesting topic that will be important later: the types of questions. Know that the reference interview offers very refined techniques that can be adapted for a truly efficient dialogue with your preferred model. Let's look at them:

• Open vs. Closed Questions: An “open” question ( “Talk about Semantic SEO” ) is ideal for the exploratory phase, to map a domain. A “closed” question ( “Does the 'canonical' tag prevent a page from being indexed?” ) serves to validate specific facts and obtain direct answers. You can alternate between the two types to guide your research and begin creating your prompt.
• The “Neutral Question” and the Power of Meta-Search: This is perhaps the most impactful technique. Instead of asking for the ultimate answer, use AI to help you build a better search strategy. This is "asking about the question."
  • "What are the most important subfields and concepts related to 'Graph Theory' for a beginner?"
  • What would be the ideal structure (topics and subtopics) for a complete article on 'The impact of disinformation on democracies'?
  • "Suggest 5 essential questions that a logistics expert would ask to optimize a supply chain."
• Attentive Listening and the Feedback Loop: In the context of AI, "attentive listening" translates to critically reading and analyzing each generated response before formulating the next prompt. Identify gaps, inaccuracies, or areas that deserve further investigation. Each AI response is an invitation to the next step of the interview.

I wanted to mention the issue of questions, as they are important for our next step.

AI-powered research framework: Grogan's eight steps

While Kipling's six questions offer us a complete framework for preparing our research prompt, Denis Grogan's model, from one of the most respected theorists of reference service, provides us with a complete map of the process, a holistic view encompassing the user's journey from the need for information to its final resolution.

"Computers, as is well known, lack this common sense and will only eliminate work that is incomprehensible to the user or obsolete books if they are specifically instructed to do so."

GROGAN, 2001, p. 83

We need to remember the difference between service and process, between the concrete action of the professional and the user's journey. The journey the user takes is longer than the reference service, and it is not uncommon for the person who needs the information to return to the librarian for another round of research.

Let's now unveil Grogan's eight steps and, for each one, draw a direct parallel with the use of deep research , as we did with Kipling's questions. To summarize, the steps are: the problem, the information need, the initial question, the negotiated question, the search strategy, the search process, the answer, and the solution.

Step 1: The problem

It all begins in the user's mind, when they identify a "knowledge gap," an uncertainty that creates a problem. In Library Science, this is called an Information Gap.

Grogan, referencing the seminal studies of Robert S. Taylor, describes this need as visceral ; it is a vague dissatisfaction, a discomfort that often cannot yet be articulated in words. It is an entirely internal phase, a silent monologue that occurs "inside the consultant's head," and only after the person goes through this process does a search, in the context of our article, effectively begin.

Therefore, there are a number of implications for the librarian. Even if the professional does not work directly at this stage, they need to recognize its existence. The understanding that every search stems from a perception of "insufficiency," "lack," or "inadequacy" of knowledge is what underlies the empathy and patience necessary for the entire subsequent process.

When a person begins a search, it's due to an understanding that they don't know enough, and this realization triggers a series of feelings. Kuhlthau's information-seeking process can help you understand this issue more deeply.

Translation for Gemini:

This is the pre-prompt . It's the ideation phase, where you, the researcher, feel the need to explore a topic, but perhaps you're not yet clear about the objective, scope, or the terms you'll use. My recommendation is to recognize the "visceral" nature of your own research, dive headfirst into what you lack: certainty, and look at your objective. What do you need to solve but don't have the information to arrive at the solution?

But don't feel pressured to formulate the perfect prompt right from the start. Accept that your first interaction with Gemini will most likely be an exploration of this still-dip need. Perhaps it's not the time for an in-depth search, but a basic investigation.

Step 2: The need for information

The abstract problem now needs to be transformed into a conscious objective: the need to seek information to fill the gap, which is why I asked you to delve into the lack but not lose sight of the objective.

This phase is still predominantly internal, but the intention to act has already been formed. Motivation can vary immensely, from simple curiosity to the urgency of a critical decision. As the philosopher Bertrand Russell observed, people seek not only knowledge, but "certainty." The role of information is, therefore, to reduce uncertainty .

Understanding that curiosity is a powerful driving force and that the ultimate goal is to mitigate uncertainty allows us to be more proactive and welcoming, validating the user's journey from its very first moments—this is the biggest implication for librarians right now. And how can we translate this to Gemini?

Translation for Gemini:

Here we begin actively working with Gemini to solve our problem, build our search, and ultimately generate our report (remembering that in Grogan's studies this is not the end of the search for the user). This is the moment when you open the Gemini interface with a purpose. Your goal is not just to "investigate possibilities," but to transform uncertainty into clarity, or at least begin that process.

Step 3: The initial question

The moment of constructing our initial question, or our case study, our prompt in its version 1.0, is an ongoing process, somewhat integrated with the previous step. But make no mistake, this is a critical moment!

This is the first verbalization of the need, the first contact with the librarian, or with our Digital , Gemini. It's the transformation of internal thought into an external question. However, as Eleanor B. Woodruff noted in 1897, it's rare for this initial question to be a perfect reflection of the real need. It's often "vague," "ambiguous," "incomplete," or "generic." The librarian's "gift," according to her, is precisely to extract the clear idea behind the vague request. And here lies an intrinsic shortcoming of AI agents: they are not librarians, they won't empathize with you and help you construct an initial question, or your first prompt with a human would.

That's why I created this video:

The main implication for the librarian is that this is the time for action; it's time to put into practice the "asking skills" they have developed and help the person searching. Experienced professionals in reference services turn their attention to the verbal and nonverbal clues they see from the person needing the information, and begin to decipher the mystery of the search, between what was and what was not said.

Translation for Gemini:

The initial question is, literally, your first prompt . And, just like in human reference, it will likely be imperfect.

Remember that you may have already experimented with some iterations of your model, but now you're going to start building your final prompt. An example of an initial prompt could be this:

"Tell me about the solar energy market."

Here we have a very generic prompt, which doesn't help the agent know if you want data , future trends, information on panel technology, incentive policies in Brazil, or a global market analysis. Therefore, the answer will be broad and probably superficial. It's the digital equivalent of going to a library counter and saying, "I want books about energy."

It's okay to start with a prompt like that, but remember: to save your time and avoid generating inefficient searches, don't generate a search yet.

Step 4: The negotiated issue

Herein lies the essence of the reference. Grogan teaches us that the negotiated issue is a collaborative "transaction" where the initial issue is negotiated, refined, and clarified.

A study by Mary Jo Lynch in 1976 revealed that almost half of library inquiries required this negotiation. This study, which was Mary Jo Lynch's doctoral research at Rutgers University, analyzed transactions in four public libraries and found that reference interviews occurred in 49% of inquiries . In other words, almost half of the interactions with users required active intervention from the librarian to identify, clarify, and refine the true issue of the inquirer.

Its absence is one of the main causes of service failure. This phase requires sensitivity, insight, and extensive general knowledge, as the librarian needs to rephrase the question in terms that the information system (whether it's a catalog or a database) can understand.

This is where Kipling's six honest servants come to the rescue, and the two frameworks can work together. The good news is that you already know how to use them; the bad news is that you can't count on a librarian to help you, unless you go to a library near your home.

According to Kipling's model, open-ended questions ("What specifically about electric cars interests me?") are preferable to closed-ended questions ("I want to know about lithium batteries in new Chinese models?").

The goal here is to understand the "motive and context," because, as Grogan says, the client "cannot define what they want, but they can comment on why they need it."

Translation for Gemini:

This is the iterative process of refining prompts . The "negotiated question" is, in fact, a sequence of prompts .

1. Prompt 1 (Initial Question): "Tell me about the solar energy market."
2. Prompt 2 (Negotiation with "Where?" and "When?"): "Excellent starting point. Now, refine this analysis by focusing exclusively on the Brazilian residential solar energy market over the last two years."
3. Prompt 3 (Negotiation with "Why?" and "How?"): "Based on this data, create a table comparing the installation costs and average payback period (ROI) for the three main photovoltaic panel technologies available in Brazil."

This dialogue is a co-production between you and the AI, where you apply Kipling's framework to transform a generic question into a precise and efficient query. Remember that you don't need to use Deep Research yet.

Step 5: The search strategy

With the negotiated question in hand, the librarian outlines an intellectual action plan. You, with your initial research completed, already have a written prompt to generate the initial question.

In a library setting, the Librarian selects the most appropriate sources (encyclopedias for an overview, periodicals for recent research), chooses specific titles, and defines access points (search terms, synonyms, subject headings).

Success here depends on intimate knowledge of the sources and that "intuition" which, in reality, is the result of years of practice. This is an analytical task that requires breaking down the topic into its facets, identifying key terms, and using tools such as thesauri and indexes. Our model

Translation for Gemini:

Instead of creating the strategy from scratch, you can use AI to co-create the search plan. This is an advanced tactic. After asking yourself open-ended questions and finding the answers, you can generate a prompt like this:

"I am researching the topic 'Adoption of Artificial Intelligence in Logistics'. To create a comprehensive report, what would be the most important sources of information to consult (e.g., market reports, academic publications, leading companies in the sector)? What are the key terms and synonyms I should use to deepen the research?"

Your agent will not only provide the information but will also help you structure your own research, a valuable step in optimizing your time. But wait, don't use Deep Research yet.

Step 6: The Search Process

This is the practical execution of the strategy. It's a dynamic, non-linear activity where we take action. In our library scenario, the professional adapts, follows new leads that emerge, and explores paths not initially foreseen. The search can be manual, computerized, or a combination of both.

In the context of Library Science, this phase demands proficiency in the use of search tools on the part of those conducting the process. Speed ​​of response may impress, but it should not sacrifice the quality and human aspect of the interaction.

Translation for Deep Research with Gemini:

Now, let's use Deep Research! This is our interactive session with the agent. Be prepared for serendipity, which is the ability to find something valuable or useful without having specifically searched for it. At this stage, the occurrence of pleasant and unexpected discoveries, often by chance, during a search for something different, is very common. And when dealing with an Artificial Intelligence model, the surprises may involve unexpected but not so pleasant discoveries.

But fear not, you haven't come this far to give up! You have in your hand a prompt specially created to generate the best search report.

Remember that AI may present a connection or subtopic you hadn't considered. Don't hesitate to deviate from your initial roadmap to explore these new avenues. Use the models' ability to process large volumes of data to quickly test hypotheses: "Analyze the data from the previous response and identify any correlations between implementation cost and the geographic location of the companies."

Step 7: The answer

Now we have the information delivered to the user, in our case, you. In most cases, it's a document, a specific piece of data, or a bibliographic reference. In our case, it's the report that Gemini delivers to you. Did you click on the link I put above to see the previously generated report?

How do you judge the quality of your report?

The answer should be clear and technically accurate; the report should guide your interpretation of the results and avoid unnecessary information and data in the answer you seek, helping you stay focused on solving your information search.

Therefore, Gemini's response is the generated text. You need to understand that this is not the final step. The output file that the model generates is a "high-quality rough draft." It should be viewed as raw material, not as the finished product. Your role is to verify, validate, contextualize, and edit this response.

Step 8: The solution

This final step ultimately belongs to you.

The solution is reached when the client, the person making the inquiry, confirms that the information provided resolved their problem and satisfied their need for information. If this does not occur, the process returns to step 4: negotiating the issue.

The “solution” is not the AI's answer, but rather the final product you create from it: the completed report, the published article, the polished presentation, the informed business decision. It is the result of your curation, critical analysis, and synthesis of the material generated by the AI.

If, upon reviewing Gemini's "response" (Step 7), you realize it doesn't meet your "gut need" (Step 1), you haven't failed. You simply restart the cycle in Step 4, re-negotiating the issue with a more refined and specific prompt, in a continuous process of improvement until the appropriate solution is reached.

Is the Future of Search a Human-Machine Partnership?

The use of tools like Gemini, ChatGPT, Claude, and others does not eliminate the need for, nor diminish the importance of, information professionals; on the contrary, it enhances their role. You yourself have seen that there are phases in this process where the knowledge of a librarian is necessary.

The ability to conduct a deep and structured dialogue with AI, using the timeless principles of the reference interview, should be understood as a valuable and effective skill for navigating the informational complexity of the 21st century and beyond.

"The purpose of reference and information services is to allow information to flow efficiently between information sources and those who need information. Without the librarian bringing the source closer to the user, this flow will never exist or will only exist inefficiently."

Kenneth Whitaker

Mastering the art of asking "What," "Why," "Who," "How," "When," and "Where" in the digital context can be what separates obtaining superficial answers from building deep and truly useful knowledge. The reference interview is not something that should be relegated to the analog past, and in my view, libraries are the ideal partner for artificial intelligence in this activity.

Excellence in search, today and in the future, is a co-production, a partnership where human intelligence guides, refines, and extracts the best from artificial intelligence.

With this report in hand, you have two options:

1. Create a notebook in Google NotebookLM : ask your notebook questions about your article ideas, interact with it, and use it as your expert on the topic you're going to discuss.
2. Create another agent : Create another specialized agent that uses all this information and generates a briefing for a human to write. With the necessary adaptations, you can generate a briefing for another agent specialized in writing.

In the second part of this article, I'll show you how to use the report we generated for our search to create an agent specialized in the subject you want!

REFERENCES

• ACCART, Jean-Philippe. Reference service: from face-to-face to virtual . Brasília, DF: Briquet de Lemos, 2012.
• BALBINOTTI, Stheve. Unveiling Grogan's eight steps in a reference process. ACB Journal: Librarianship in Santa Catarina , Florianópolis, v. 25, n. 3, p. 571-587, Aug./Dec., 2020.
• FERREIRA, Nalin. Marketing SRI (Slides) . [Sl: sn], [sd]. Teaching material.
• FERREIRA, Nalin. PRACTICE OF THE REFERRAL SERVICE: an experience report in the health field . [Sl: sn], [n.d.]. Teaching material.
• FIGUEIREDO, Nice Menezes de. Reference and information services . São Paulo: Polis: APB, 1992.
• GILLUM, Shalu. The Changing Face of Public Services in Health Sciences Libraries Post-Pandemic. Medical Reference Services Quarterly , vol. 43, no. 2, p. 152-163, 2024.
• GROGAN, Denis. The practice of referral service . Translated by Antonio Agenor Briquet de Lemos. Brasília, DF: Briquet de Lemos/Livros, 1995.
• GROGAN, Denis. The practice of referral service . Brasília: Briquet de Lemos, 2001.
• HIGGINS, Julian et al. Update from the Methodological Expectations of Cochrane Intervention Reviews (MECIR) project. Cochrane methods , vol. 2, no. 3, 2012.
• IFRS CAMPUS PORTO ALEGRE. Protocol for developing a search strategy . Porto Alegre: IFRS Campus Porto Alegre, [n.d.].
• LANKES, RD Expect more: better libraries for a complex world . São Paulo: Febab, 2016.
• LANKES, RD Bad Libraries Build Collections, Good Libraries Build Services, Great Libraries Build Communities. 2012. Available at: https://davidlankes.org/beyond-the-bullet-points-bad-libraries-build-collections-good-libraries-build-services-great-libraries-build-communities/. Accessed on: May 7, 2025.
• Medical Library Association - MLA. Professional development. Professional competencies . Available at: mlanet.org/page/competencies. Accessed on: April 4, 2022.
• MONTEIRO-KREBS, Luciana. BIB03088 – Reference and Information Service: Lesson 01_v2 . [Sl: sn], [sd]. Teaching material.
• MONTEIRO-KREBS, Luciana. BIB03088 – Reference and Information Service: Lesson 04b . [Sl: sn], [sd]. Teaching material.
• MONTEIRO-KREBS, Luciana. BIB03088 – Reference and Information Service: Lesson 2 – Slides . [Sl: sn], [sd]. Teaching material.
• MONTEIRO-KREBS, Luciana. BIB03088 – Reference and Information Service: Virtual SRI . [Sl: sn], [sd]. Teaching material.
• MONTEIRO-KREBS, Luciana. Performance indicator framework for the reference service . [Sl: sn], [sd]. Teaching material.
• MORO, Eliane Lourdes da Silva; SILVEIRA, Manoela; PARANHOS, Mariana; MARTINI, Paula. What is a face-to-face reference service . Porto Alegre: UFRGS, 2015. Academic work (Undergraduate degree in Library Science and Information Science) – Federal University of Rio Grande do Sul, Faculty of Library Science and Social Communication, Department of Information Science, Porto Alegre, 2015.
• NATIONAL LIBRARY OF MEDICINE. NETWORK OF THE NATIONAL LIBRARY OF MEDICINE. Health literacy . Available at: https://nnlm.gov/initiatives/topics/health-literacy#toc-4. Accessed on: April 4, 2022.
• HOLY LIBRARY SCIENCE. The 8 Steps of Reference Service According to Grogan . [Sl]: Holy Library Science, [sd]. Duration: 07min00s. Available at: https://youtu.be/NqS8pB48y14. Accessed on: May 18, 2024.
• SILVEIRA, Nalin Ferreira da. The Information Commons in Brazilian university libraries: support for research and innovation for scientific and social development . 2019. 181 p. : ill. Dissertation (Professional Master's in Informatics in Education) – Federal Institute of Education, Science and Technology of Rio Grande do Sul, Porto Alegre, 2019.
• VARGAS, GAT; AVILLANEDA, MDRS Reference service in the digital library. Bibliotecological investigation , Mexico, v. 19, no. 39, p. 120-133, 2009.