A modern approach to research: AI-powered Research System

Artificial Intelligence is inevitably becoming an integral part of the research process. One of the ground-breaking tools EDISONDA offers its clients is the AI-powered Research System. An AI-powered Research System that enables organisations to collect, analyse and share knowledge from ongoing research more effectively.

In this article, we take a closer look at what the AI-powered Research System is, the elements that make it up, and how it can help companies with their decision making.

What is the problem?

In a traditional research cycle, research findings are usually presented in the form of reports. These summaries include a description of the study, examples of user responses, and conclusions and recommendations for implementation. Although these documents are prepared with great care, their life span usually ends at the presentation stage. Long reports are difficult to read and the knowledge they contain is difficult to access. Each time we want to read the observations, we have to find the right report and then search it for the information we want. The knowledge contained in the reports is unstructured and very difficult to manage and distribute. This traditional model of collecting research knowledge suffers from several problems:

Researchers’ limited memory

One of the most common problems is losing conclusions or forgetting the results of previous studies. This situation is known as ‘limited researcher memory’. In many organisations, research is often not properly archived, or access to it is hampered by a dispersed system of files and disks, resulting in duplication of research or under-utilisation of existing knowledge.

Knowledge silos

In large organisations, the lack of a publicly accessible knowledge repository means that departments often work independently of each other.

This leads to a lack of communication and knowledge sharing. This division hinders collaboration and blocks the flow of information between key departments.

Cumbersome data format

As mentioned above, research reports often take the form of long and complex documents that are time-consuming to produce and difficult to review. The knowledge encapsulated in the document makes it difficult to use the research, especially by those who were not directly involved in producing it.

Generality and timeliness of data

In traditional research systems, research findings often become outdated or lose context over time. Unfortunately, the conclusions collected in reports cannot be updated or overwritten. This can lead to the use of outdated data or outdated recommendations.

What does the Research System change?

First and foremost, the Research System organises knowledge and research results. This is done through the Knowledge Repository, which is at the heart of the Research System and allows research results to be categorised and stored in a consistent manner and made easily accessible to anyone seeking knowledge. Such a publicly accessible knowledge base greatly facilitates cross-departmental collaboration, allowing knowledge to be centrally managed and shared across the organisation.

The AI-powered Research System takes this a step further. It is a more advanced solution that uses artificial intelligence to enhance the research process. The AI Research System harnesses the power of large language models, which differentiates it from traditional approaches in its ability to process large amounts of data and find patterns in it.

The capabilities provided by the LLM configuration enable intelligent assistance in finding research results and clustering knowledge to create interactive representations of users, called AI Personas.

What is the basis for an AI-powered Research System?

To better understand the construction of an AI research system, let’s look at the elements that underpin this way of collecting and using research knowledge. There are four pillars on which a modern research system should be built: the research strategy, the aforementioned knowledge repository, and the AI-assisted elements, namely AI Assist and AI Personas.

Research strategy

The research strategy is a key element of the AI-powered Research System. It is the plan that enables research to be conducted with real users on a regular basis. It provides the system with a cyclic portion of current data, which is the basis for analysis. 

When creating a Research System supported by AI, we must remember that artificial intelligence is not human. It is not designed to be, and never will be. Therefore, it is extremely important that when we build an AI-based research system, we build it on the data we get from real users.

Knowledge Repository

The Knowledge Repository is a central database that stores research, analysis, and other relevant information. It is the place where structured research results are stored. This structure makes it easy to browse and search the data, often using tags, search engines, filters and browsing modes. The repository acts as a ‘second brain’ where knowledge is categorised and stored in an organised way. Repositories are created using Excel. Microsoft List, Airtable or Dovetail.

AI Assistant 

The AI Assistant is a virtual assistant that searches the database and generates answers to user questions. It is a suitably configured voice model, thanks to Prompt Engineering, with which we interact via a chat interface. Such an assistant can direct us to specific documents or provide comprehensive answers to general questions about the results of the research.

AI Persona 

An AI Persona, also known as a synthetic user, is a virtual representation of a typical representative of the target audience. It is an advanced feature of the Generative AI configuration that allows interactions with a linguistic model based on a specific set of survey data and other related sources, that match the characteristics of the target audience. The AI Persona is created by restricting the model’s knowledge to a set of data corresponding to a particular segment. The model representing the AI persona is trained on group-specific inferences, interview transcripts, survey results or product reviews. The AI persona has a broader application because we can ask it evaluative or predictive questions. It can help us test hypotheses, evaluate new ideas and simulate the behaviour of different customer segments.

How does the process of developing an AI-powered Research System work?

Designing an AI-powered Research System is a step-by-step process:

  • We begin our work with the client with a workshop to learn about the research cycle, the customer segment and the purposes for which research is conducted in the organisation.
  • We then look at the structure and formats of the data available to the client. In many cases, this data needs to be made consistent and structured in order for the language models to work effectively.
  • The next step is to select the ecosystem and model that will support our knowledge base. Depending on whether the company is working in a specific ecosystem of tools, e.g. Microsoft, Google, or has a flexible software approach, we will choose the appropriate base for the repository and one of the available language models.
  • As we have already mentioned, the creation of the AI-powered Research System begins with the construction of the Knowledge Rrepository, which is the foundation of the system. At this stage, we will feed the repository with existing data and prepare procedures for updating the database with new research results. We will determine the format of the data and how it will be categorised and described.
  • In the next step, as part of the proof of concept, we will configure the language model so that it becomes a very useful tool for searching for content. After the testing phase, we will integrate the language model into the database.
  • In the final step, we will configure the models so that their behaviour is a representation of the selected target groups, i.e. the AI Persons, and plan ways to train them (fine tuning).

Be one of the first to seize the opportunity with AI

The AI-powered Research System is a new and experimental approach to integrating artificial intelligence into the research process. It is important to realise that this is only the beginning of the possibilities of using AI to capture, manage and distribute knowledge within an organisation. Imagine an AI persona speaking to you in the voice of a real respondent, with a digital persona and memory of recent conversations.

This is all possible now, but before we get too far ahead of ourselves, it’s worth building the basic system we’ve described in this article. Remember, implementing an AI-powered Research System is a process that needs to be fine-tuned to your organisation’s needs, strategic goals, and data structure, but it has numerous benefits that make research more accessible and useful. With properly implemented elements such as a research strategy, knowledge repository, AI assistant and AI persona, you will move your organisation closer to a place where decisions are made faster and more informed.

Unlock the future of research with AI. Let's talk!

Michał Madura
Senior Business Design Consultant

    Do you wish to receive the latest information related to the topics of business and innovation design, as well as information about Edisonda's activities, projects, and offers?

    Please select the channel through which we can contact you (consent is voluntary):

    Information provided in the form will be used only in order to reach back to you. Contents of the correspondence might be archived. More details could be found in our privacy policy.