[Report] SNOWFLAKE WORLD TOUR 2026 - TOKYO: Day 2 KEYNOTE #SWTTokyo26

This is Kawabata.

"SNOWFLAKE WORLD TOUR 2026 - TOKYO" was held from September 10, 2026 to September 11, 2026.

https://www.snowflake.com/ja/world-tour/tokyo/

This article is a report blog for the session "Day2 KEYNOTE."

Please check the "Day1 KEYNOTE" at the link below.

https://dev.classmethod.jp/articles/snowflake-world-tour-2026-tokyo-day1-keynote-swttokyo26/

※Some features are in preview or coming soon. Please check the latest official documentation for availability.
※Speaker remarks are summarized based on the author's notes. Portions of conversations in English are paraphrased by the author.

Speakers

• Sridhar Ramaswamy
  • Snowflake Chief Executive Officer (CEO)
• John Robertson
  • Snowflake APJ President and Chairman Executive Officer (VP, APJ Sales)
• Ryuji Ukita
  • Snowflake President and Executive Officer
• Munehisa Matsumoto
  • Sumitomo Mitsui Trust Asset Management Co., Ltd., Executive Officer
• Chihiro Ueno
  • CyberAgent, Inc., Executive Officer, Head of AI Operations Division
• Ken Takao
  • CyberAgent, Inc., Group IT Promotion Division, Company-wide Data Technology Bureau, Bureau Chief (Manager)
• Toshiyuki Takahashi
  • Snowflake Marketing Division, Product Marketing Manager and Evangelist

The event started with Kumataro's drum performance, just like Day1

It was a very lively rock atmosphere.

From Snowflake's Ukita

As was mentioned on Day1, there were references to the companies and partners presenting.

• Snowflake World Tour is held in more than 20 cities around the world. Tokyo is the only 2-day event in the world and is the largest event in the World Tour.

• Further expanding from last year, with more than 12,000 registrants, 144 sessions, more than 42 partners, and 202 speakers
• This event is centered on customers, not Snowflake. Day2 will feature a conversation with CEO Sridhar, as well as introductions to data and AI initiatives at Sumitomo Mitsui Trust Asset Management and CyberAgent
• Thanks to sponsoring partners and guidance on the Expo area. Hope everyone finds ideas useful for new technology, services, business, and personal projects
• For those who couldn't attend Day1, Evangelist Takahashi will recap Day1

Day1 Recap by Snowflake's Takahashi

This was a section reorganizing the content of the Day1 KEYNOTE as "The 4 Elements of the Agentic Enterprise." Please see the Day1 report above for details.

• The era of AI agents doing work has arrived, and the way we work and the nature of companies is about to change dramatically. Snowflake calls this the Agentic Enterprise
• The reason for the feedback "we tried agents but not everything works" comes down to one thing: data and systems are fragmented and siloed, putting them in a state where AI cannot be utilized
• Snowflake provides the 4 elements for overcoming this fragmentation on a single platform
  • Data: Govern with Horizon Catalog and add meaning to data with Horizon Context. Control access for both humans and agents with Trust Center
  • AI Models: Choose models based on performance and cost. Cortex AI Gateway automatically selects the appropriate model for each task and tracks agent activity
  • Applications: Connect to data from SAP and others without copying using zero-copy, and integrate with other services via MCP
  • Control Plane: The command center that organically connects the three elements. Snowflake CoWork for business users and Snowflake CoCo for builders
• Memorable words from Day1 customer sessions
  • Nissan Motor: "AI is an enabler of corporate transformation." As data became usable by AI, the mindset of business divisions changed
  • Fujifilm: "Governance is not a brake, but a guardrail that lets everyone confidently step on the accelerator"
  • Anthropic: "The biggest risk in AI is not using AI"
• Three closing points: organize your data assets, get to an AI-ready state, and use AI for business value rather than making AI use itself the goal

From Snowflake's John Robertson

• Head of APJ, based in Tokyo. This is his 35th year in Japan, having worked almost entirely in the IT industry for American companies
• American companies often don't understand Japan, and half of Snowflake Japan's job is communicating Japan's needs to headquarters
• The most important thing is "whether headquarters and the CEO understand Japan." Sridhar has been visiting Japan almost every year since his Google days, and it's reassuring to have a CEO who loves Japan

Conversation Between Snowflake's John Robertson and Sridhar Ramaswamy

Q. What do you love about Japan?

• The passion poured into everything. Not just event management, but even small things like hotel doorknobs show craftsmanship and care
• The way small coffee shop and restaurant owners take pride in their work. He tries to bring that same passion to all Snowflake employees in their contribution to products and customers

Q. What is the Agentic Enterprise? What impact is it having on customers?

• Agentic AI is "the industrialization of intelligence." Just as writing was invented not for poetry but for tax records, software is a way to express processes. Now software creation is cheap and fast, and processes can be expressed as skills in natural language
• This is especially important for Japan, where you can't scale by increasing population. You can achieve significant acceleration with the people you have
• Snowflake's internal practice: distributed Snowflake CoCo and Snowflake CoWork to the sales team
  • Everyday tasks like "what use cases are open for this customer?" and "how should I prepare for a meeting?" have changed
  • Sridhar put a photo of the guest list for a lunch invitation into Snowflake CoWork on his smartphone to look up who he had met before. Something that would have previously taken someone a week to research
• Governance is necessary precisely because everyone uses the same agents. Not everyone should be able to see real-time sales figures for all customers
• In the near future, everyone will have a "text box to ask questions," with the entire organization's wisdom and context available there. That is the beginning of the Agentic Enterprise era

QBR Preparation Done in 4 Minutes

• This morning, Sridhar asked "do you know the status of your top 30 customers?", and got an answer faster than Sridhar using Snowflake CoWork
• Could discuss in 4 minutes which customers are using which use cases and AI products, and where they are headed. Previously, this would have required a QBR and hours of preparation
• The entire cycle of asking a data engineer to create a chart, not understanding it, and scheduling a Zoom meeting disappears entirely
• Due to RBAC, you can only access data you are authorized to see as a GM. This is exactly the governance needed when deploying AI

Q. How do you view the Japanese market?

• A progressive country with one of the world's largest economies, enthusiastic about leveraging technology. Many global brands, and they have deep relationships with key players
• It's also a market that demands perfection. A CFO who says "the revenue numbers are roughly right" is not acceptable. Snowflake meets those demands
• You should expect more from Snowflake. Some customers who previously took more than 3 years for legacy migration are now doing it within 3 months. They'll reduce it further to 1/10, enabling complex migrations within one month
• Snowflake's own data team has transformed from an organization with 5 years of backlog to one leading transformation. Data teams can go from a slow cost center to the center of transforming their company with data
• What he tells every customer he meets: "If you fully adopt Snowflake CoCo and Snowflake CoWork, you'll be 5-10x faster, not 5-10% faster"
• John: Japan still has a lot of on-premises. It's an opportunity to migrate to the right cloud data platform and get ready for AI

Q. Flexibility and choice in AI and cloud. Any announcements for Japan?

• They've adopted open data formats like Iceberg and Polaris, and are available on all of AWS, Azure, and GCP. Data should not be locked in by anyone, including Snowflake
• You can also choose from OpenAI, Anthropic, and open source/open weight models
• Snowflake will become available on Google Cloud in Japan (Tokyo region) in FY27 Q4

[Press Release]
https://www.snowflake.com/ja/news/press-releases/snowflake-google-cloud-tokyo-region-japan/

Q. What's your advice on AI investment ROI and tokenomics?

• AI is a technology where you can demand real ROI. There are only three reasons to spend money in business: "earn more," "spend less," or "provide better service to customers"
• Inside Snowflake, they stopped spending on dashboards in exchange for deploying AI to the sales team, saving approximately $5 million. Spending on AI was considerably less than that
• A per-user quota can set "the maximum amount one person can spend per month." This preserves the benefits of the consumption model while preventing runaway spending by individual users
• "Earn more" metrics: proposed use cases per account executive up more than 40% year-over-year, deployed use cases up approximately 90%, new logo acquisition at one of the best quarters ever
• The cost governance skill in Snowflake CoCo (one of the top 10 skills) can reduce unnecessary usage. Cortex AI Gateway automatically routes to the lowest-cost model appropriate for the task
• Advice: Relentlessly demand positive ROI. AI is a means, and you don't need to burn tokens just to say "we're using AI"

Q. What's your advice on making data a competitive advantage?

• In a world where software and intelligence are no longer expensive, your alpha is your data, context, and operational knowledge of your business. Complex UIs and algorithms can be expressed as natural language skills
• The visual layer is generated on the fly and disappears. Inside Snowflake, SKO reports are created on the spot, uploaded to Snowflake CoWork, and shared
• What matters is data quality and the logic that decides what to do with that data. Companies that are progressing rapidly are those that have organized their data assets and identified and optimized their most important processes
• Large companies achieving dramatically higher growth rates doesn't happen without real magic. That magic is the ability for the people doing the actual work to drive the liberation of data assets and thorough process optimization

Presentation by Sumitomo Mitsui Trust Asset Management's Matsumoto

As one of the core companies of the Sumitomo Mitsui Trust Group, with approximately ¥111 trillion in assets under management and approximately 714 employees, this was a presentation about their efforts to consolidate business data and logic centered on Snowflake, and what happens as a result.

Company-wide Technology Utilization Structure

• Committed at the company level to "building a lean company that can operate with half the members"
• Each department identifies all tasks and prioritizes them, creating scenarios. Approximately 650 tasks across the company have been turned into Skills
• Select promotion members from each department and proceed with self-driven, in-house development. Members who deliver results are properly evaluated
• PDCA is automated rather than managed in Excel. It runs by combining three elements: the complete task list, created Snowflake objects, and usage logs

Approach and Skill Creation

• ① Organize data → ② Business owners use the data to execute tasks and verbalize their perspectives, decision criteria, and outputs → ③ With one phrase "turn what we just did into a skill," knowledge becomes a skill
• In 2025, the stage was "LLMs write code for you," but with the advent of agents, natural language can be run directly, significantly lowering the barrier to entry for in-house development

• A real skill example: a skill for creating stock commentary. "What to do," "when to use it," and "how to do it" are all defined in natural language, and with one phrase like "create an investment stock introduction comment for CyberAgent," the output is produced

• An episode illustrating the impact of skills
  • When Anthropic announced a business-specific plugin in February 2026, the stock prices of Thomson Reuters in the legal domain and IBM in COBOL analysis fell
  • What happened was "the public release of a mechanism for loading knowledge in natural language onto general-purpose models," and that alone had a non-negligible impact on the market

Activity Results

• Of 650 tasks, 120 projects have been initiated. Approximately 20% by count, with approximately 35% impact on work by effort
• Approximately 4,000 tables onboarded, 200 skills created (more in development and testing)

• Snowflake CoCo utilization rate went from 10% to 60% in April. Usage per person is 7x compared to March

Governance: Rogue Tasks Disappear

• "Are you sure business controls are okay?" is a common question, but the answer is the opposite. Rather than rogue EUC increasing, rogue tasks disappear
• Traditional EUC was centered on local Excel files and code that couldn't be controlled, but if done on Snowflake, everything is observable. Since the substance of the task is a skill (natural language), it also serves directly as a manual

• Introduction of a report auto-generated by Snowflake CoCo
  • Evaluates task coverage for a given department. Lists trends in active members, individual usage volumes, and created objects
  • Evaluates each task by criteria such as "established," "functional but not established," "temporarily spreading," or "data only," showing where problems lie: functionality, utilization, or data
  • Also outputs team-by-team and individual status and improvement suggestions. If you specify "summarize for presentation," it instantly produces a presentation-ready version
• The reason EUC control was difficult was not because creation was free and easy, but because what was created could not be observed

Commitment to Business Outcomes

• Questions like "what results can AI produce?" and "will it really increase revenue?" are fundamentally backwards
• The right approach is to set goals like "we want to double our sales touchpoints" or "we want to halve the time to market for products" and then think together about whether AI can achieve them. An offensive stance of agreeing on KPIs with business impact with frontline staff and using AI to achieve them is necessary
• Connecting the medium-term management plan, the complete task list, objects, and usage logs, and understanding feasibility toward goal achievement forms the baseline for discussion

Reducing the Cost of Problem Discovery and Loop Engineering

• As business data accumulates, the tasks that can be covered increase, and as rules and manuals accumulate, requirements become fully defined. Natural language query logs become a wish list of "what we want to do next"

• What comes next is loop engineering: a mechanism for systems to autonomously and recursively continue improving. What's needed is problem discovery capability and problem-solving capability
• Three technical elements for running it safely and reliably
  • Microservices architecture: prevents auto-generated items from affecting other components
  • Harness engineering: supports safe and reliable operation
  • Business semantics engineering: Snowflake's comment function, declaration of data responsibility scope, behavioral standards through skills, ontology through knowledge graphs and Cortex Sense

• A favorite quote is Netflix's Reed Hastings: "Lead with context, not control." In the LLM era, if you can provide LLMs with the right information, the organization becomes smarter

Q. How do you run the loop for problem discovery?

• Running two loops to improve efficiency per credit
  • For individuals: usage information is used to automatically send daily emails saying "here's how you could do better," and the effect is traced
  • For the overall system: automatically propose necessary hooks, controls, and context to provide to objects, and implement them across the entire system
• Loops come in large, medium, and small; it's best to start with small loops and work up to larger ones. All implemented in natural language

Q. How have development speed, cost, and people's roles changed with skill utilization?

• If asked to "build a search function," you would normally build a search function, but now that agents are available, that's not enough. You need to understand the motivation behind wanting to search and how the data will be used after searching, and can then entrust the whole flow to an agent
• Introducing thin-sliced functionality can become a bottleneck for overall throughput. This shift in mindset is the biggest change that has come with agentic AI
• People's roles are becoming more business-oriented, and committing to business value is more important

Session by CyberAgent's Ueno and Takao

History of AI Utilization and the AI Operations Division

• With a vision of "creating a company that represents the 21st century," the three core businesses are internet advertising, media (ABEMA), and gaming

• AI Lab research organization launched in 2016 (currently over 100 members). Has provided products such as advertising effectiveness prediction and creative generation
• With the arrival of ChatGPT in 2022, they recognized that "AI is not just for a few engineers," and established the AI Operations Division in 2023. In 2025, they also established an AI-driven development promotion office for development organizations
• Steps for company-wide rollout

• 2023: Held a contest to collect ideas for applying AI to tasks so that employees in each business unit could take ownership, along with reskilling on security and governance
• A one-day event held once a year where each business unit brings their cases and boasts about them

• What they've been putting the most effort into recently is the AI Rankings. Modeled on sumo rankings, business leaders are ranked by their degree of AI utilization. It's an initiative where everyone thinks about how to connect AI not just to efficiency but to business value

• Characteristics of the AI Operations Division
  • A company-wide function not tied to any specific business. Approximately 20 dedicated members centered on developers and product managers
  • Business leaders from each business unit serve concurrently, enabling unified perspective and messaging on AI across the company
  • Not a company that decides things centrally; values the freedom and self-responsibility of each business unit
• Snowflake has been used in internal systems for advertising delivery data for 5 years. Easy to add management information, etc., and has good compatibility with security reliability and machine learning features

• Working with the AI-driven development promotion office, group IT promotion, and technical leadership organizations to build an environment where everyone across the company can use AI

Q. What are the results of AI utilization?

• Not just engineers, but nearly 4,000 people company-wide are using tools like ChatGPT and Claude
• In development organizations, development output has approximately doubled. Significant reduction also seen in the domain of product managers and planners, such as writing specifications
• In sales departments, there are examples of individuals being able to handle more customers, leading to increased revenue. Cases directly tied to business outcomes are increasing internally

Data Platform and MCP Integration

• Responsible for group IT and data. An environment exists where everyone can use ChatGPT and Claude, and everyone is making good use of AI
• Data from all corners of the company is in Snowflake, and they are enabling secure access to data using MCP
• Use cases
  • Corporate information search
  • Safety confirmation system: instantly collect information during disasters and check the status of their own department
  • Cosmetics business: machine learning on store data, contract agents, sentiment analysis
• No matter how high-performance a model is, if data is siloed it can't access the right data and AI will make mistakes too. The fundamental principle of not letting data become siloed is important
• Infrastructure uses AWS and GCP in roughly equal proportions. Looking forward to the GCP Tokyo region announced this time, and wants to advance integration with data on Google Cloud

Q. What is your medium- to long-term vision?

• Aiming to be a leading company in the AI era
• In development organizations, they are aiming for complete automation of all company development processes by 2028
• In talent development, they have announced new careers: "Business Lead Engineer" and "AIX Designer." Job definitions and evaluation/education systems are being established
• Existing businesses will optimize operations and output through collaboration with AI agents, and provide new value through new businesses unique to AI

Closing

• We are in the midst of the largest structural change of this generation in enterprise technology
• With the Snowflake AI Data Cloud, you can leverage a governed foundation, connected systems, agents, and a control plane to turn autonomous AI into the ultimate competitive advantage
• AI business application has entered the phase of not "whether it can be done" but "how to do it"

Impressions

I attended the SNOWFLAKE WORLD TOUR 2026 - TOKYO Day2 KEYNOTE.

The presentation by Sumitomo Mitsui Trust Asset Management strongly drove home the importance of not asking "what results can AI produce?" or "will revenue really increase?", but instead setting goals like "we want to double our sales touchpoints" or "we want to halve the time to market for products" as the right question, and then thinking together about whether AI can achieve them.

Although the major Japan event SNOWFLAKE WORLD TOUR has concluded, our use of data and AI continues without interruption. I hope the insights gained at this event can be taken back and applied within organizations to contribute to business development.
We hope to be able to support you through information sharing and assistance as well.

I hope this article is helpful in some way!