【AISC Introduction】 The Three Teams That Make Up AISC
Introduction
Hello, I’m Asami Minei, a member of GA Technologies’ Advanced Innovation Strategy Center (AISC).
In this article, the third installment of the ”AISC Introduction" series, I will introduce the three teams that make up AISC.
Some of you may already be familiar with the "AISC Introduction" series, which is a project aimed at not only introducing AISC but also providing insights into the department’s atmosphere and showcasing examples of members’ skill development, told through the words of the members themselves.
In previous installments of the "AISC Introduction" series, we’ve discussed why we started the series and the development training we provide within the department. If you haven't read the series yet, I encourage you to take a look at the table of contents below.
AISC’s Role and the Three Teams
First, AISC’s mission is to “contribute to our business activities using advanced technologies”, leveraging data and technology to conduct research and development in the real estate and M&A fields.
In simple terms, it’s a department that develops technology aimed at automating and streamlining various tasks to the utmost extent. Our research spans a wide range of fields, including image recognition, speech recognition, natural language processing, matching algorithms, and data analysis. With these research and development efforts as our core, we work daily to contribute to the business through various initiatives.
We also have articles introducing projects that our members have been involved in, as well as prototypes that are publicly available, so feel free to check them out for reference.
As of October 2024, AISC consists of the following three teams, each with different functions.
Next, we’ll dive deeper into each team with the insights shared by the team members themselves.
DX Team Introduction
Featured Member: Shota Yamauchi
Shota Yamauchi graduated from the Graduate School of Social Informatics at Kyoto University. In graduate school, he conducted research on systems to encourage visits to medical institutions by potentially serious patients. He joined GA technologies as a new graduate and has since been engaged in customer analysis work to support sales.
- Could you explain the DX team and its role?
The DX team focuses on tasks where the ROI (Return on Investment) can be calculated with a certain degree of clarity. ROI refers to the balance between the effort invested and the return achieved. For example, if a system that took 10 hours to build results in a savings of 20 hours per month, we determine that the ROI is worthwhile for AISC. Given the nature of these tasks, we often work on improving the efficiency of various departments.
Currently, we are developing systems to optimize workflows in various departments, such as RAM, Spica, Renters Warehouse, and Partners.
- What are the main tasks of the DX team and how do you approach your work?
Our work begins with interviews with field staff. Specifically, we listen carefully to the workflow and the issues they face. Afterward, we engage in discussions with them to explore the best possible improvements. Once the proposed solution is agreed upon, we move on to building the actual system. At this stage, we aim to develop an MVP (Minimum Viable Product) as quickly as possible. The MVP is a product with only the essential functions, and we refer to it as a prototype. We rapidly create it, have it used in the field, and continuously improve it based on feedback. This fast-paced cycle is characteristic of our development flow.
- What is unique about the DX team?
There are two aspects that differentiate us from other teams: our approach to problem-solving and development.
Problem-Solving Approach
In problem-solving, it's common to first identify issues, calculate ROI, prioritize them, and then address the highest-priority issues. While this approach is generally correct, conducting a thorough identification process can be time-consuming. At AISC, we adjust the period for identifying issues based on how much improvement is needed in the business. If the business is not yet very efficient, we can complete the identification and prioritization in 30 minutes to an hour. Of course, this isn't a perfect process, so we continue to review and adjust prioritization as we make improvements. This allows us to move very quickly compared to traditional problem-solving methods.Development Approach
In a traditional waterfall development model, where requirements are defined strictly and then implemented accordingly, there's often a gap between defined requirements and functional requirements. That's why we adopt an agile development process. Specifically, we quickly develop an MVP, get it used in the field, and receive feedback. We then make flexible adjustments to the specifications as needed. We believe this flexibility is key to achieving the best results quickly.
- What does the DX team value?
What we value most is maintaining a balance between our own know-how and the ideas of the intended staff users. While field staff are the ones who understand their work best, if we rely solely on their input, we risk developing suboptimal solutions. Conversely, it’s not ideal to implement every suggestion without considering our expertise in technology. We believe that the best solutions are found by working closely with field staff and discussing issues together.
- What challenges does the DX team want to take on?
Our goal is to create a world where humans and programs work in harmony. We envision a future where programs handle the tasks they excel at, freeing people to focus on the work that only humans can do. By achieving this vision, we aim to relieve team members from tedious, repetitive tasks and allow them to focus on more specialized work.
- What kind of person is best suited for the DX team?
Several important characteristics are essential:
Balanced with Technology
Our primary goal is to solve business challenges, not to use the latest technology. Therefore, we seek people who view both new and old technologies as tools and can select the best tool for the problem at hand.Not a Perfectionist
It often takes more effort to raise a system from 90 points to 100 points than it does to build the system from 0 to 90 points. Being flexible enough to aim for 90 points, rather than 100, is crucial.Effective Communicator
Tasks that only involve oneself can be advanced quickly with effort. However, when collaboration with others is required, early communication is essential. Otherwise, progress can slow down. In the DX team, we frequently conduct interviews with field staff, and it’s critical to schedule these in a way that doesn’t disrupt their work.
DS Team Introduction
Featured Member: Yoshiyasu Shirado
Graduated from Tohoku University with a degree in Economics. Joined GA technologies in 2017 and engaged in web application development. Leveraging his engineering experience, he worked as a Product Manager (PdM) in launching multiple products and also created business KPIs as a data analyst. Since 2022, he has been part of the AISC.
- Can you explain the role of the DS team?
The DS team’s main role is the "formalization of tacit knowledge."
In not only the real estate industry but also many analog industries, the quality of work is often maintained by tacit knowledge (knowledge that is difficult to verbalize, often based on personal experience or expertise).
Since its founding, the GA group has been actively advancing the use of technology, leading to the accumulation of consistent data related to the acquisition, sale, and management of real estate. By utilizing this data, we aim to enhance the reproducibility of judgments made by professionals in the field, improving customer satisfaction and experience. Additionally, we promote research and data utilization that serve as barriers to entry for other companies from a data science perspective.
We focus on both short-term and mid-to-long-term issues that need to be solved.
- Can you tell us about the work style of the DS team members?
Each member of the DS team is responsible for 2 to 5 topics.
The topics vary: some are handled individually, some are considered by the team, and some are conducted as joint research with university professors. The content also spans all phases related to data utilization, from data acquisition to analysis and application/development.
- What is unique about the DS team?
As mentioned earlier, the DS team handles both short-term and mid-to-long-term themes, covering all phases of data utilization.
This creates a team where each member has different strengths, which is particularly noticeable in our team within AISC. Some members specialize in various aspects of data science, others have strong development skills, and people like me have a strong understanding of the business side. Each of us promotes our own themes.
- What is something the DS team values?
We prioritize showing respect to the professionals on the ground. Rather than solely focusing on refining the data, the DS team adopts an approach where we quickly introduce our work to the field and collect feedback for improvement. This means we pay attention to how we can provide information in a way that is easy to use and ensure there are no misunderstandings in our explanations.
Sometimes, there’s a gap between the data and the intuition of people in the field. In such cases, we first question the data and, if it still doesn’t explain the gap, we dig deeply into both sides to understand why the discrepancy occurred.
- What challenges do you want to take on as the DS team going forward?
In addition to continuously contributing to the business through data science, we plan to further strengthen research activities that contribute to the development of talented members and raise the GA Group’s presence.
- What type of person do you think is suited for the DS team?
Not just for the DS team, but for AISC as a whole, the ideal person is:
Someone who can think deeply.
Someone who can move toward achieving goals without being fixated on just the means.
R&D Team Introduction
Featured Member: Aaron Bramson
Graduated with a Ph.D. in Philosophy and Political Science from the University of Michigan in 2012 after completing an M.S. in Mathematics in 2004. He worked as an MBA instructor at the Rotman School of Management at the University of Toronto from 2010 to 2011. Before joining GA, he worked at the RIKEN Brain Science Institute in Japan, Sandia Laboratories in the US, the Department of General Economics of Ghent University in Belgium, and numerous consulting projects worldwide under his own company, Complexity Research Corporation. His research focuses on the many aspects of complexity science: theory, application, model construction, and analytical methodology. At AISC, he leads research and development projects integrating transportation networks, geospatial data, simulated flows, scoring/matching, and estimation.
- Can you describe the R&D team and its role?
Our team is an organization within AISC that focuses on mid-to-long-term contributions. We aim to conduct research projects that often integrate data science, DX, and other departments. The current focus is on scoring buildings and regions, which requires a broad range of tasks including floorplan evaluation, analyzing building facilities, discovering endogenous neighborhoods, creating scores to capture levels of lifestyle support, modeling multimodal transportation networks, and many other research tasks. Results of this research also need to be integrated into DS and DX projects such as rent prediction and streamlined workflows, and sometimes be developed into their own products.
- What is the main work of the R&D team and how do you approach it?
R&D projects are ones where the solution cannot be achieved by using an existing package or implementing some published algorithm. It is necessary to first understand what relevant technology, mathematics, and ideas are out there and evaluate their usefulness. Then we must create a research plan to develop a new capability to solve a problem or in expectation of its future value. We often receive support from other colleagues in collecting and processing data, backend support, app interfaces, etc., but most of the research is done by specialists and their team.
- What does the R&D team value?
Since we handle mid-to-long-term themes, there are often unclear paths to business contribution. To ensure that we don’t lose direction, it’s important to regularly communicate with various members in the field to understand their needs.
When solving existing problems, the needs are clear, but the technology needed is not. In this case we need to work efficiently and create partial solutions even though there are no deadlines imposed. Being able to retain focus and produce results in ambiguous environments is critical here. Business needs are constantly changing too, so it is important to be flexible and be willing to abandon promising ideas when they are no longer relevant.
When we are building capabilities that don't currently exist, we can't make progress by just fulfilling requirements because the need for them hasn't been realized. Therefore, I emphasize self-motivated actions based on reasons rather than rules. It also means that we can't be sure any single idea will produce worthwhile results; we need to be creative and explore many ideas.
- What challenges does the R&D team want to tackle in the future?
For me personally, I aim to improve our walkability scores by including factors like inclines, sidewalks, and road widths, to reflect mobility differences. I also plan to complete bus and train networks for the whole country and conduct new analyses on accessibility scores and demand estimation that include bus and bicycle transportation modes.
More broadly, we aim to shift the current paradigm of property searching that puts all the information burden on the customer. People often don't know what they really want, and/or don't know what home features complement and facilitate their desired lifestyle. Asking somebody who has only lived in their suburban parent's home and student housing to find a new home by filtering properties by attributes and arbitrary geographic areas is actually strange. Instead, we should inquire about their place of work, their hobbies, their lifestyle preferences, then use our data and models to tell them which properties are good matches. This way of thinking even extends to investment properties: in this case we are matching properties' financial characteristics to a customer's life plan rather than asking them what kind of property they want.
- What kind of person is a good fit for the R&D team?
For the projects that I primarily work on, someone interested in geography, geospatial data, and network data analysis. More generally, the ideal candidate can understand existing methods and go beyond them to create better indicators and scores. You need to manage your time independently and understand how individual tasks connect to the main goals. However, it’s also essential to ask questions when uncertain about the task’s purpose or direction to ensure smooth progress on the project.
Although we use AI approaches where appropriate, a good candidate should be curious about a wider range of methodology, and have the technical background to understand and expand on them. So, self-motivated, independent, curious, and creative thinkers are good for the R&D team.
In Conclusion
Thank you for reading this article to the end. I hope this team introduction has sparked some interest in AISC.
In the next installment of the "AISC Introduction" series, Mr. Shirato, who appeared in this team introduction, will share insights into the development training conducted at AISC. Be sure to give it a read!
We will continue to deliver more insights about working and activities at AISC through this note, so keep enjoying our updates!