Snowflake Summit Day 3: Data Readiness and Ecosystem Power

Conversations on day three of the Snowflake Summit shifted toward what it takes to make data truly AI-ready for the era of intelligent agents. 

My highlights from the day break down into two core areas: how leading retail brands are partnering with Snowflake to improve customer engagement and how the broader Snowflake ecosystem is building the integrated workflows to make enterprise AI a reality. 

From personalisation to the agentic era

Paul Winsor, Head of Retail EMEA at Snowflake, has worked in the retail industry for 40 years, having begun his career with UK supermarket Sainsburys. 

“2026 has really changed in terms of retailers talking much more about the agentic era,” he told Technology Magazine. 

“Every time we get in front of the customer today that is leveraging Snowflake, they want to understand how they can increase the understanding of their customers to really attract them and engage with them.”

He pointed to Sainsbury’s as a great example of a Snowflake customer using their data to do just this. 

“Sainsbury’s has 18 million people using its Nectar loyalty card scheme. It is so hyper-personalised now that every Friday, if you’re part of the scheme, you get 10 offer sent to you that are personalised – price personalisation or points. It’s the price one that’s very interesting because it’s rewarding you based on your previous shopping habits.”

Sainsburys runs more than 260 million personalised price and points promotions to these customers every week, and it is only able to do that by bringing all its customer data together into one platform.

Paul also mentioned how sportswear retailer Asics is using personalising experiences for its customers using data managed by Snowflake. 

“Asics has a Runkeeper app that holds all kinds of information about your running performance, taken from data collected from your runs,” he said.

The app tracks data taken from built-in GPS on phones and GPS watches like Garmin, Fitbit and Apple Watch to monitor location, pace and distance.

“ASICS brought all of that data together in Snowflake and now uses that for personalised content for its customers,” Paul said. “This conscious engagement [with customers] is really important.”

But personalisation is the first step. The real shift happening right now is the move toward live, agentic intelligence to manage product launches in real time. 

For example, Samsung is using Snowflake CoWork – the conversational assistant that sits on top of company data so employees can ‘talk’ to their data – to monitor its Samsung Galaxy S26 series.

Snowflake EVP of Product Christian Kleinerman invited Jung Suh, Head of Digital Commerce at Samsung, onstage during his keynote to talk about this. She explained how the company is using intelligent agents to understand the instant customer response to the product. 

“When we launch a flagship, we are simultaneously tracking massive streams of data. The real challenge has been philosophy.

“By the time a traditional analytics team surfaces wide, conversion is down in a region and they come up with a reason why we already missed the window of action. We already lost the attention of our target customer audience and we need to move with the launch, not behind it.”

AI changes that for Samsung. 

“We built what we call the Shoppers Inside Action agent. I can take the data and ask the agent to compare the launch performance of Galaxy 26, compared against the last model. It doesn’t just load it on number. It plans a set of steps and reconcile the signals and gives me a synthesised answer. Work that took my team hours now takes seconds and those seconds matter when you’re trying to troubleshoot an issue.”

Inside the Snowflake Ecosystem

Walking the floor, one message was clear: enterprise AI has moved past pure experimentation. 

Organisations are no longer asking if they should deploy AI, but how to do so safely, quickly and with clear business outcomes. 

The answer, according to some of Snowflake’s closest partners, lies in a deeply integrated ecosystem built around Snowflake Cortex, turning raw data into agentic workflows.

At the foundational layer sits the monumental, multi-billion-dollar alliance between Snowflake and AWS. 

Mona Chadha, Director of Strategic Partnerships at AWS, pointed out that this is a deeply symbiotic relationship that has already driven US$7 billion in joint contract value. 

Snowflake’s AI architecture is built directly on Amazon Bedrock, using AWS’ high-performance Graviton compute to power its native AI capabilities.

Meanwhile, Sigma aims to remove technical friction for the enterprise by meeting the traditional business user exactly where they already work. Rather than forcing non-technical teams to learn complex code, Sigma acts as an intuitive interface sitting directly on top of the data warehouse. 

Shawn Namdar, Senior Director of Partner Engineering, and Luke Stanke, Marketing Leader, summarised the dynamic: “Snowflake provides the building blocks – Sigma is the easy way to work with those blocks, always keeping within Snowflake.”

This native integration ensures strict data governance and security boundaries are never compromised.

For teams looking to build more complex data pipelines and predictive applications, Dataiku provides the critical visual layout. The company’s recently announced Dataiku Cobuild for Snowflake allows joint customers – like Air Canada, which uses the joint tech stack for advanced marketing modelling – to easily orchestrate AI success. 

Dataiku is highlighting at the event how the pressure on companies to deliver real business outcomes is immense, emphasising that successful enterprises are the ones working backwards from the business problem rather than just chasing the latest tech.

Finally, pulling these capabilities into specific industry verticals are systems integrators like Tredence. 

Rather than chasing entirely autonomous, human-free systems, Tredence advocates for a balanced architecture of humans in, out and for the loop. 

Data context is pulled directly from Snowflake to train highly specialised AI personas.