Clektive Thinking: Snowflake CoWork – bringing agentic AI to every employee 

For the last few years, the promise of AI in business has largely centred around being able to ask questions and get answers. What were our sales last quarter? Which regions are underperforming? What is causing our delivery KPIs to fall behind? 

That is useful. But it is only one part of the picture. 

The next stage of enterprise AI is about moving beyond simply chatting with data and towards AI that can actively support the way people work. That is where Snowflake CoWork comes in. 

As part of the latest episode of Clektive Thinking, Chris and Elliott explored how Snowflake’s AI capabilities have evolved, what CoWork makes possible, and why businesses should be paying attention now. 

The evolution from Cortex to Snowflake Intelligence to CoWork 

Snowflake’s AI journey has moved quickly. 

Cortex introduced the ability to use large language models within Snowflake and start doing more with business data. Snowflake Intelligence then made it easier for users to ask questions of that data using natural language. 

CoWork takes this a step further. 

Rather than simply asking AI a question and receiving an answer, users can begin to automate tasks, create reusable outputs, connect different systems and build AI capabilities around specific roles and business processes. 

In simple terms, the shift is from AI that answers questions to AI that works alongside you. 

This matters because the value of AI does not come from having another chat interface. It comes from reducing the manual work involved in finding information, analysing it and turning it into something useful. 

Putting trusted data in the hands of more people 

One of the biggest opportunities with CoWork is making business data easier to access without relying on someone else to find, query or interpret it. 

Imagine an operations director wanting to understand which KPIs are currently in the red. Instead of logging into multiple systems, asking a data team to run a report or working through a dashboard, they can ask the question directly. 

The AI can query the underlying data and return an answer in seconds. The same applies to finance, marketing and other areas of the business.  

A finance leader could ask which regions are underperforming against budget, then drill further into the data to understand which business units or products are driving that result. An executive could ask for a summary of the quarter, pulling together information from multiple parts of the organisation. 

The important point is that this is not about giving everyone unrestricted access to everything. 

Different agents can be created for different departments, roles or use cases, giving people access to the information they need while maintaining control over what they can see. 

The foundation still matters 

This kind of experience can look deceptively simple. You ask a question. The AI returns an answer, chart or analysis. But the quality of the answer depends heavily on what sits underneath it. 

For AI to work reliably with business data, organisations need a clear and trusted foundation. That includes well-structured data and a semantic layer that gives the AI a shared understanding of what key business terms mean. 

What does the business mean by revenue? What counts as a KPI breach? How is gross margin calculated? 

Without clear definitions, you risk getting different answers depending on who asks the question or which tool they use. 

With the right semantic layer in place, the aim is consistency. The same underlying data and definitions should produce the same answer, whether someone is accessing it through CoWork, another Snowflake interface or an application built on top of the platform. 

That is what makes it possible to move faster without losing confidence in the information being used. 

From one off questions to automated work 

One of the most interesting developments within CoWork is the ability to move beyond individual conversations. 

Users can create automations that run work on a schedule and notify them when it is complete. 

For example, instead of manually checking a set of KPIs every Monday morning, a leader could set up an automated check that highlights anything requiring attention. A monthly board pack could pull together key information automatically. A recurring performance report could be refreshed without someone having to rebuild the analysis every time. 

This is where the conversation around AI starts to become much more practical. 

The question businesses should be asking now is what work are people repeatedly doing that AI could help us automate or simplify? 

Creating outputs that can be revisited and built upon 

CoWork also introduces the idea of artifacts. 

Rather than losing useful analysis within a chat history, users can save outputs such as charts, tables and reports, revisit them and continue building on them. 

If a piece of analysis becomes useful, it does not need to be recreated from scratch every time. It can be refreshed with the latest data, explored in more detail or shared with colleagues. This is particularly valuable for recurring business questions. 

A leadership team may want to review the same set of metrics every month. An operations team may regularly need to understand where performance is falling behind. A finance team may want to explore budget performance from different angles. 

Instead of repeatedly asking someone to build the same report, the work can become more reusable. 

Connecting AI to the wider business 

The opportunity extends beyond data held within Snowflake. 

Through capabilities such as MCP connectors, AI can connect with other tools and services used across the organisation. 

That opens up the possibility of bringing information and actions together rather than treating every platform as a separate destination. 

For business leaders, this could mean being able to pull together information from finance, marketing and operations in one place. For employees, it could mean using AI to access information across the systems they already rely on. 

The potential is significant, but it also makes governance more important. 

Businesses need to understand what information AI can access, what actions it can take and how employees are using it. This is one of the reasons enterprise platforms are becoming increasingly important in the AI conversation. 

Many employees are already using AI tools independently. The challenge for businesses is making sure that adoption is secure, governed and aligned with how the organisation wants to work. 

AI in the hands of users 

Another important shift is accessibility. 

AI capabilities are no longer limited to technical teams working within a data platform. With web, mobile and desktop experiences, the aim is to put trusted information closer to the people making decisions. 

That could mean answering a question during a meeting rather than going away and waiting for someone to pull the data. It could mean checking performance while travelling. It could mean having a conversation with your business data in the same way you would ask a colleague for information. 

The technology is becoming easier to use. But that does not mean implementation should be treated casually. 

Where should businesses start? 

The biggest mistake businesses can make is starting with the technology rather than the problem. 

CoWork and other AI tools can potentially be applied across hundreds of use cases. That does not mean every use case is worth pursuing. 

A better starting point is to look at where people are currently spending time. What reports take days or weeks to produce? Where are employees repeatedly searching across different systems for information? Which decisions are slowed down because someone needs to manually pull together the data? Where are teams doing work that could be automated, but where the right controls and data foundations are already in place? 

From there, businesses can start to identify use cases where AI could deliver measurable value. That value might come from saving time. It could come from faster decision making, reducing manual processes or giving more people access to the information they need. 

The important part is being able to quantify the outcome. 

The opportunity is huge, but the journey needs direction 

The shift towards AI agents and more autonomous ways of working is moving quickly. 

Tools such as Snowflake CoWork show what is becoming possible when AI, trusted data and business processes start to come together. But the technology itself is only part of the equation. 

Businesses still need to understand where AI can have the greatest impact, make sure the right governance is in place and build on a data foundation they can trust. 

For some organisations, the first step might be improving governance and putting the right guardrails around AI. For others, it could be identifying a time-consuming process that can be automated and proving the value. 

The right starting point will depend on where the business is today. What is clear, though, is that the conversation is moving on. The future of enterprise AI is not just about asking better questions. 

It is about creating systems that can help people find answers, carry out work and make better decisions with the information already available to them. 

The businesses that get the most value will be the ones that identify where AI can make a genuine difference, build the right foundations and give their people the confidence to use it. 

Want to understand where AI could have the biggest impact across your organisation? Clekt helps businesses build the data, governance and AI foundations needed to move from experimentation to practical, measurable outcomes. Get in touch today.