Dear Mikiya, NTT DATA is a leading Japanese multinational in the IT and AI sectors, serving a
diverse global customer base. I’d like to explore with you some of the key elements that
characterize NTT DATA’s approach to developing products and solutions, particularly those
involving advanced mathematical models and machine learning. But before we dive into
specifics, could you please introduce yourself?
Thank you. I’m Mikiya Tanizawa, a deputy manager at NTT DATA Group Corporation, where I
work closely with multidisciplinary teams across our global organization to deliver innovative
technologies to our customers to create co-R&D.
Could you share some examples of model-based AI solutions or products that NOR is currently
developing or has already introduced?
Certainly. There are numerous areas of focus, each influenced by regional trends. A common
global priority is revenue and pricing optimization, which has gained significant attention, as
highlighted in Gartner reports. Additionally, large language models (LLMs) have become a
pivotal technology, with applications across multiple sectors. At NTT DATA, we’re interested in
integrating these technologies into other tools like math-optimization solvers. These solvers
power everything from supply chain efficiency to pricing strategies, ensuring optimization at
scale.
That’s intriguing. Can you elaborate on the key elements of your pricing optimization
solutions?
Absolutely. Pricing is a strategic lever for retailers—it defines their market position, shapes
customer relationships, and influences competitiveness. At the same time, pricing is a highly
operational challenge. Retailers must update prices for thousands of products while
monitoring competitors and managing dynamic promotions.
To address this, we’ve developed pricing decision-making technologies that bridge strategic
and operational needs. It enables top managers to set high-level strategies while allowing
day-to-day operations to implement those strategies consistently. For instance, a retailer
might prioritize customer acquisition for specific products while maintaining profitability in
others.
Our technology employs predictive optimization, where forecasting is crucial. It’s also an
“explainable optimization tool,” emphasizing transparency. Users often find machine learning
and mathematical optimization models opaque, but we ensure the scenarios and proposed
solutions are clear and actionable.
Imagine managing thousands of products, each with unique costs, customer behaviors,
substitutes, and competing offerings. Add the dynamic nature of pricing in e-commerce, and
the complexity becomes clear. Our models help navigate this complexity while ensuring
coherence with the retailer’s strategy.
It sounds sophisticated. What challenges do retailers face when selecting and adopting these
kinds of solutions?
One major challenge is comparing solutions in the market. Many deep-tech aspects—like how
forecasts are generated or elasticity is modeled—are not immediately apparent but greatly
influence the quality and usability of the solution. Additionally, the rollout time and the
learning curve for users to master these tools can vary significantly.
Retailers also face unique internal hurdles, such as varying data quality or organizational
cultures. That’s why, during our sales process, we engage a consulting team to work closely
with customers. We help them define priorities, analyze their data, and ensure the solution
aligns with their specific needs and characteristics.
You mentioned that NTT DATA also has a math-optimization solver and is developing an LLM-
based solution. Could you elaborate on these?
Certainly. Our solver, Nuorium Optimizer, is a key component of our pricing and revenue
optimization tools. It’s a robust engine developed by NTT DATA Mathematical Systems Inc., which
has a lot of data scientists and engineers. Nuorium Optimizer delivers the computational power
behind our optimization models. Utilizing the solver, by controlling order and production
volumes, we were able to maximize profits while reducing food waste in Japan. In addition, we
are integrating the solver into Syntphony Pricing Management, which is developed by NTT DATA
Italia and has several commercial performances.
Regarding LLMs, NTT DATA launched Tsuzumi through the Microsoft Azure AI Models-as-a-
Service (MaaS) offering, which is a Large Language Model (LLM) with robust capabilities in Japanese and English. We’re actively developing applications that enhance user experiences. For
example, within our pricing solutions, we’ve proposed a great solution such that LLMs can act as
intelligent assistants, helping users configure optimization parameters or quickly extract data
insights. This adds a layer of usability and efficiency to our tools.
NTT DATA operates on a global scale, merging technologies and expertise from diverse teams.
How has that experience been for you?
It’s both challenging and incredibly rewarding. Combining complex technologies with the
insights and expertise of multicultural teams requires careful collaboration. But when it all
comes together in a successful solution, the sense of achievement is unparalleled. In one word,
it’s exciting.
