In May 2023, ITOCHU Corporation (“ITOCHU”) launched a cross-company task force to promote the use of generative AI across all businesses. This included incorporating generative AI into FOODATA, a data analytics dashboard which the ITOCHU Food Company provides to customers. ITOCHU is using Azure OpenAI Service and Azure AI Studio to evolve its data analytics dashboard into a service that provides immediate recommendations by automatically creating evidence-based product proposals. The corporation has already forged more than 70 ideas for using generative AI and has begun developing a data integration platform to bring these ideas to fruition. Microsoft Fabric is a front-runner for this platform for its potential to leverage retrieval-augmented generation (RAG) in higher quality generative AI responses.
Creating a platform to analyze a massive range of disparate data and provide ideas
A rapidly growing number of companies are adopting Azure OpenAI Service to streamline operations, achieve higher productivity, and solve labor shortages by leveraging generative AI internally. Further, endeavors to deliver the value created by generative AI to customers and drive new business value are flourishing. A notable example is ITOCHU’s use of generative AI in FOODATA.
“ITOCHU’s Food Company services about 10,000 manufacturers and around 3,000 retailers and restaurants,” says Kento Tsukada, ITOCHU Food Company Product Manager. “We launched FOODATA in July 2021 to support customers’ businesses by providing information in addition to trade and logistics services.” The FOODATA dashboard aggregates and enables multidimensional analysis of consumer purchasing trends and surveys, social media discourse, quantitative data on palates, and other information for product planning and development.
According to Tsukada, “The aggregated data includes ITOCHU's own information and some from partner companies. FOODATA’s standout trait is its capacity to collect, use, and analyze a diverse array of disparate data.”
While FOODATA may be a revolutionary data analysis platform, Tsukada encountered two major issues after launch. The first was a growing demand for faster data processing and greater operability. The other was a need for output with greater value-adds, as well as a more efficient data analysis process.
“Half of product planning is data collection and analysis. This is the area in which FOODATA excels: our customers rate it highly and we want to make it even better,” explains Tsukada. “The remaining half involves generating ideas. There is still room for greater functionality here, but developing new functions is tricky because the area is closely related to human intuition and experience. Overcoming this challenge was our next major hurdle.”
Generative AI was a potential solution. Masato Hayashi participated in the FOODATA project as part of the ITOCHU DX Project Management Department. “Kento had been investigating the possibility of applying generative AI to FOODATA for a while,” he recalls.
A task force launched in May 2023 to explore the potential of company-wide use of generative AI to achieve the vision.
A task force to support AI use in FOODATA, automatically generating product proposals
“In May 2023, ITOCHU established a generative AI lab with the aim of the entire company using generative AI in a secure environment and a task force to solicit ideas for use from younger workers,” says the ITOCHU ICT Division’s Yusuke Tsujii, who led the task force. Since the lab had already opted to use Azure OpenAI Service, in June 2023, the company looked for a technology partner skilled in AI integration and chose Headwaters following a recommendation from Microsoft Japan. The first step was to study the use of generative AI.
Tsujii explains the reason for selecting Azure OpenAI Service as ITOCHU’s generative AI engine below.
“The launch of ChatGPT in November 2022 was a stunning emergence of AI with the potential for many novel uses. Microsoft was quick to strike a capital alliance with the owner of ChatGPT, OpenAI, to provide Azure OpenAI Service with stronger security and a multitude of functions.”
After completing a two-month study, ITOCHU started to formulate ideas on applying generative AI. We didn’t merely introduce generative AI—we viewed it as a service to fine-tune through use and decided Azure was a cloud platform that supported such use. Our generative AI lab had already adopted Azure OpenAI Service, so we decided to use it for our task force too Hayashi therefore proposed the use of generative AI in FOODATA. The task force set two themes for its first PoC, including generative AI for FOODATA.
Comprehensive discussions for the PoC began in November 2023 before kickoff in January of the following year. Tsukada explains his future vision for the services and the reason for these tests below.
“Our goal with generative AI is to transform FOODATA into a service that provides immediate product proposal drafts for food manufacturers by conducting comprehensive data analysis. On top of delivering greater efficiency by eliminating the need for analysis, we want to add new value in the form of idea generation.”
The first step to achieving this goal was to incorporate comprehensive analysis capabilities aligned with product planning and development processes. ITOCHU is using the current PoC to implement this functionality through a combination of Azure OpenAI Service and Azure AI Studio. This includes verifying that generative AI assembles the SQL information necessary for data analytics.
The goal of the second step is to automatically generate product proposals for food manufacturers based on this analysis. Upon entering theories or conditions into FOODATA, generative AI will analyze the data and generate multiple proposals incorporating project concepts such as product targets, usage scenarios, and benefits and over 20 product features like flavor, manufacturing methods, raw materials, prices, and sales channels. The only thing left for managers to do is evaluate the proposals and make decisions.
A company-wide generative AI platform, Microsoft Fabric as a foundational data preprocessing solution
“Headwaters supported everything from the learning phase through to development,” says Hayashi. He also points out that having such partners is one of the appeals of the Microsoft AI ecosystem. “The abundance of great partners like Headwaters is another superb aspect of Azure OpenAI Service. We can also anticipate rapid functional improvements leveraging Microsoft's capital.”
The PoC phase will end in March 2023, and test marketing for practical use will begin the following month. The task force has started a new project to translate IR information into English for the General Affairs Department. According to Tsujii, “These two projects are the task force’s first steps in promoting the use of generative AI. In the next phase, we’ll examine cross-company infrastructure rather than project-specific environments.”
The configuration diagram below shows the shared infrastructure. In addition to adopting Azure AI Studio as the AI platform for the entirety of ITOCHU, the company is using Microsoft Fabric to aggregate and cleanse data to make it suitable for AI use. Further, it will leverage data and generative AI through Microsoft Fabric’s notebooks.
“For higher quality generative AI responses, we must leverage relevant knowledge sources using RAG. To search these knowledge sources with greater accuracy and adequacy, pre- and post-processing of source data is crucial. Microsoft Fabric can be the de facto standard as a RAG-inclusive data platform for this pre- and post-processing,” says Headwaters’ Director and task force participant Masaya Nishimaki.
“Ultimately, we want to enable non-engineers to easily perform everything from pre- and post-processing of source data to the use of generative AI,” Nishimaki continues. “A preview version of Microsoft Fabric Copilot is available, which will make it even easier for us.”
Tsujii has already fielded many calls from sections within ITOCHU wanting to use generative AI, and workers across the company have devised over 70 ideas for specific applications.
“Some ideas for generative AI can be applied across ITOCHU, while others are specific to the Trading Company,” says Tsujii. “Over the coming one to two years, domain-specific generative AI usage will soar. The innovative ability of Microsoft Fabric to import data virtually from many sources will establish it as a key platform for supporting diverse applications of generative AI. We want to leverage this foundation to promote the use of generative AI and other advanced technologies over the long term.”
“Our goal with generative AI is to transform FOODATA into a service that provides immediate product proposal drafts for food manufacturers by conducting comprehensive data analysis.”
Kento Tsukada, Product Manager, Food Products Marketing & Retail Section No. 2, Food Products Marketing & Distribution Division, Food Company, ITOCHU Corporation
Follow Microsoft