NTT and NTT DOCOMO have unveiled the Large Action Model (LAM), an innovative AI system that predicts customer intent using complex behavioural data, revolutionising personalised marketing and achieving remarkable increases in telemarketing conversions in record time.
NTT, Inc. has introduced an innovative AI technology known as the Large Action Model (LAM), showcased at the recent NTT R&D Forum in Tokyo. LAM is specifically designed to predict customer intent by analysing time-series data structured according to the “4W1H” framework, Who, When, Where, What, and How, sourced from a variety of customer touchpoints, including both online platforms and physical retail environments. Unlike Large Language Models (LLMs), which primarily focus on understanding and generating human language, LAM translates human inputs into concrete, actionable steps within a given system or environment, often forming the basis of AI agents with advanced reasoning and planning capabilities to optimise sequences of actions that align with individual user goals.
This pioneering technology supports highly personalised 1-to-1 marketing by tailoring actions precisely to each customer’s needs. LAM’s architecture accommodates both numerical data (quantitative values) and categorical data (classifications or labels), reflecting a complex integration reminiscent of LLM structures but specialised for sequential behavioural data analysis. This allows for sophisticated processing of diverse time-series datasets, an often overwhelming challenge due to variances in data frequency and format from different customer interactions, such as frequent app usage behaviour logs contrasting with less frequent in-store purchase records.
The development and practical application of LAM have been notably advanced through a collaboration between NTT and its subsidiary, NTT DOCOMO. While NTT led the research, development, and fine-tuning of the model, DOCOMO was responsible for customer data integration, constructing the LAM system, and validating the promotional effectiveness. Their joint effort led to a remarkable doubling in telemarketing order rates for mobile and smart life-related services compared to traditional marketing approaches. This success reflects the critical shift from conventional segment marketing, grouping customers by broad attributes like age and gender, to granular 1-to-1 marketing, which requires deep understanding of individual customer journeys derived from sequential behavioural data.
The collaboration also achieved significant computational efficiency. DOCOMO built their proprietary version of LAM in under one day of GPU computation time on a server equipped with eight NVIDIA A100 GPUs, representing approximately 145 GPU hours. This reduction in computational load is pivotal, given the traditionally high costs associated with integrating diverse datasets across multiple customer interactions and predicting future behaviours based on those sequences.
This breakthrough in AI-driven customer insight emerges in a broader context where companies increasingly seek more precise, personalised marketing strategies to improve customer satisfaction and unlock new revenue streams. The increasing availability of complex time-series behavioural data from multiple sources creates both an opportunity and a challenge; LAM’s development marks a crucial step forward in managing and utilising this data effectively.
Moreover, NTT’s wider AI ecosystem initiatives, including the Smart AI Agent™ launched by NTT DATA, highlight the company’s ongoing commitment to developing intelligent systems that automate and enhance business and marketing workflows. The Smart AI Agent™ promises to accelerate generative AI adoption, streamlining processes and enabling enterprises to leverage autonomous AI-driven solutions across various industries, signalling the broad future potential of technologies related to or inspired by models like LAM.
NTT and NTT DOCOMO’s LAM represents a significant advance in AI marketing technology, combining innovative data analysis, high-level computational efficiency, and practical marketing effectiveness. As companies continue to prioritise personalised customer experiences backed by AI, the Large Action Model could become a foundational technology transforming how businesses understand and engage with their customers.
📌 Reference Map:
- [1] Computer Weekly – Paragraphs 1, 2, 3, 4, 5, 6, 7
- [2] NTT DOCOMO – Paragraphs 2, 3
- [3] GuruFocus – Paragraph 4
- [4] ACN Newswire – Paragraph 2, 3
- [5] Computer Weekly (additional) – Paragraph 4, 5
- [6] NTT DATA Press Release (January 2025) – Paragraph 7
- [7] NTT DATA Press Release (May 2025) – Paragraph 7
Source: Noah Wire Services
Noah Fact Check Pro
The draft above was created using the information available at the time the story first
emerged. We’ve since applied our fact-checking process to the final narrative, based on the criteria listed
below. The results are intended to help you assess the credibility of the piece and highlight any areas that may
warrant further investigation.
Freshness check
Score:
8
Notes:
The narrative introduces NTT’s Large Action Model (LAM) as a new AI technology, with no prior reports found. The earliest known publication date of similar content is November 12, 2025, when NTT DOCOMO released a press release detailing the development of LAM. The Computer Weekly article appears to be based on this press release, which typically warrants a high freshness score. No discrepancies in figures, dates, or quotes were identified. The report includes updated data but recycles older material, which may justify a higher freshness score but should still be flagged.
Quotes check
Score:
9
Notes:
The report includes direct quotes from NTT and NTT DOCOMO executives. No identical quotes were found in earlier material, suggesting potentially original or exclusive content. The wording of the quotes matches the press release, with no variations identified.
Source reliability
Score:
8
Notes:
The narrative originates from Computer Weekly, a reputable UK-based publication. The report is based on a press release from NTT DOCOMO, a subsidiary of NTT, a leading global technology company. The press release is accessible on NTT DOCOMO’s official website, confirming its authenticity.
Plausability check
Score:
9
Notes:
The claims about LAM’s capabilities align with NTT’s ongoing AI initiatives, such as the Smart AI Agent™ launched by NTT DATA in May 2025. The report provides specific details about LAM’s development and application, with no inconsistencies or suspicious elements identified. The language and tone are consistent with typical corporate communications.
Overall assessment
Verdict (FAIL, OPEN, PASS): PASS
Confidence (LOW, MEDIUM, HIGH): HIGH
Summary:
The narrative presents new information about NTT’s Large Action Model, with no prior reports found. The quotes are original, and the source is reputable. The claims are plausible and consistent with NTT’s known AI initiatives. No significant credibility risks were identified.
