MEDIUM Announces Sales Data Platform to Solve 'Data Deficiency' Bottleneck in Sales DX

MEDIUM Inc. (Headquarters: Shibuya-ku, Tokyo, CEO: Shotaro Seki), which promotes DX in sales strategy, announced today that it has defined the structural challenge in AI utilization within Japanese sales organizations as 'sales data deficiency' and will strengthen the provision of its 'STRIX' as a sales data platform for automatically accumulating and structuring primary negotiation information. ■ AI Adoption Accelerates in Japanese Companies. However, Sales Alone is Being Left Behind Generative AI adoption is rapidly progressing in Japanese companies. According to NRI's 'IT Utilization Status Survey 2025,' 57.7% of companies have already introduced generative AI, an increase of 24 points in just two years from 33.8% in 2023 (*1). Yano Research Institute predicts that adoption rates will reach 80-90% in the near future (*2).

Figure: Status of Introduction or Consideration of New Technologies Image Source: NRI (Nomura Research Institute) Conducts 'IT Utilization Status Survey (2025)' Targeting Japanese Companies, November 2025

However, this wave has not fully reached the sales frontlines.

According to HubSpot's 'Survey on Awareness and Actual Conditions of Sales in Japan 2025' (6th Edition), the utilization rate of generative AI in sales activities is 28.9%. Although this is an improvement from 21.1% in the previous year, there is still a gap of approximately 29 points compared to the overall corporate AI adoption rate of 57.7% (*3). Image Source: PR TIMES HubSpot Announces Results of 'Survey on Awareness and Actual Conditions of Sales in Japan 2025', February 2025

Why is Sales Alone Being Left Behind in AI Utilization? The Structure of 'Data Deficiency' For generative AI to support sales processes, it requires primary information to determine 'what to propose' and 'which deals are at risk.' However, actual sales environments suffer from problems such as 'data silos' and 'data abstraction,' which hinder generative AI utilization. ・【Data Silos】 Data accessible to AI is scattered (siloed) in the first place, creating a barrier to AI utilization, and limiting AI use to individuals.

・【Data Abstraction】 Information remaining in SFA tends to be abstract and subjective summaries, lacking the customer's verbatim statements (primary information). AI cannot be used for deep inference.

In many companies, negotiation records in SFA/CRM are limited to a few lines of summary. What concerns the customer, what they compare with competitors, and who within the company is the decision-maker – this information exists only in the memory of individual salespeople and is not structured anywhere.

Therefore, we believe that AI utilization in sales begins with 'building a data foundation,' which led us to strengthen the provision of STRIX as a 'Sales Data Platform.' ■ Overview of the Sales Data Platform Realized by 'STRIX'

The Sales Data Platform provided by STRIX is neither a mere minutes-taking tool nor just an SFA. By redefining traditional SFA/CRM tools, it eliminates data silo issues and provides an environment where generative AI can demonstrate its true value by analyzing and structuring all negotiation data using generative AI, rather than manual input by sales representatives, to build an accurate and granular 'usable database.' Feature 1: Accumulates 'All' Primary Negotiation Information

The biggest reason for the deficiency of sales data is not that it's 'not entered,' but that primary information disappears (is forgotten or lost) before it can be entered.

STRIX continuously acquires conversations as primary information simply by 'recording' negotiations, whether online or in-person. This allows for the preservation of conversation logs as verifiable 'facts' rather than reports dependent on memory or individual subjectivity. Feature 2: Structures Data with the Granularity Required for Sales Activities (Makes Data Usable by AI)

Even if conversation logs are saved, they are unstructured data that is difficult to search or analyze, and also difficult for AI to handle. STRIX automatically extracts elements required for sales activities from conversation content, such as hearing summaries, BANT, decision-making processes, concerns, agreements, and next actions, organizing them while maintaining context in a three-tier structure: Company/Deal (Opportunity)/Negotiation Log.

As a result, it creates a data model for sales that allows tracking 'how deals are progressing' and 'what the bottlenecks are' chronologically, rather than just single meeting minutes. Feature 3: Connects with SFA/CRM to Resolve Data Deficiency and Abstraction

In many organizations, SFA/CRM becomes a mere formality because data entry becomes an additional task for the field staff. STRIX aims to simultaneously reduce input burden and ensure data quality by linking structured data obtained from negotiations to SFA/CRM.

For the field staff, it provides an experience of 'easier reporting' and 'information retention without manual input.' For management, it brings them closer to a state where 'deal assessments are aligned' and 'materials for review and training are available.' STRIX does not replace SFA/CRM; instead, it serves as a data source for SFA/CRM to function effectively. Feature 4: Cross-Searches Accumulated Data to Create an Environment Where Generative AI Can 'Infer'

Once data is collected, structured, and accumulated with context, generative AI can finally play a role in advancing sales organization decision-making.

With STRIX, it becomes possible to perform tasks such as searching for similar deals, reusing past proposals, detecting signs of lost deals, extracting reasons for stagnation, and extracting winning patterns across the organization to update playbooks, all based on primary negotiation information. In other words, STRIX's Sales Data Platform is a foundation for creating a cycle where data is not just 'stored' but also 'used, learned, and continuously updated.' ■ STRIX, Which Builds an Environment Where Generative AI 'Demonstrates True Value,' Eliminates the Black Box of Negotiations, and Improves Win Rates Image: STRIX Agent (AI Agent Function)

When data is accumulated and structured by the Sales Data Platform, generative AI can go beyond simple text generation and achieve the following in the sales field: ・Predict the near future (foresee lost deals, create proposals closest to closing)

・Assetize past negotiation logs and finally succeed in knowledge management (cross-deploy successful patterns)

・AI agents can task and execute post-negotiation work (focus on proposal activities)

'Quickly creating meeting minutes or emails' does not directly lead to sales results. True value is demonstrated when primary negotiation information is continuously accumulated, structured within the context of the deal, and can be referenced across the organization – in other words, when there is a 'data foundation that AI can infer from.' ■ CEO Comment (Shotaro Seki, Representative Director, MEDIUM Inc.) The truly difficult aspect of implementing generative AI in sales is not model selection or prompt engineering, but the absence of sales data that AI can reference. Corporate AI adoption in Japan has exceeded 57%. However, in sales fields, primary negotiation information is not recorded, SFA only contains two-line summaries, and salespeople operate based on individual memory and intuition. No matter how excellent an AI is introduced, it cannot lead to results without the raw material for inference.

We are convinced that AI utilization in sales begins with 'building a data foundation,' not 'tool implementation,' and have redefined STRIX as a Sales Data Platform. We aim to make this the standard equipment for Japanese sales organizations, driving sales with 'organizational science' rather than 'individual intuition.' ■ About MEDIUM Inc. A startup specializing in AI product development and strategic consulting. It offers solutions that fuse the knowledge of its serial entrepreneur CEO with the real-world challenges of the sales field. Starting with the release of the new feature 'STRIX Agent' in 2026, it is strongly promoting stagnant Japanese sales DX.

STRIX Service Site: https://strixai.jp/ For those wishing to schedule a service introduction meeting: https://strixai.jp/contact * Please inquire via the form.

Survey Data Sources *1. NRI 'IT Utilization Status Survey 2025' (https://www.nri.com/jp/news/newsrelease/20251125_1.html) *2. Yano Research Institute 'Generative AI Usage Status Survey' (https://www.yano.co.jp/press/pdf/202407031.pdf) *3. HubSpot 'Survey on Awareness and Actual Conditions of Sales in Japan 2025' (https://prtimes.jp/main/html/rd/p/000000068.000037724.html)