Practical guides, case studies, and insights into Salesforce consulting, Snowflake consulting, and AI-powered delivery. Written by the consultants who’ve been there, done that.
Drive AI Outcomes with Snowflake and Salesforce
June 18, 2026
Heather
Rhyne-Christensen
Jump to section:
Snowflake Summit made one thing very clear: the enterprise AI conversation is growing up.
We are moving past the “look what this chatbot can do” phase and into the much harder, much more valuable question: how do companies actually use AI to change the way work gets done?
Snowflake’s answer is not just more AI. It is AI with trusted data, governed context, open architecture, secure action, and fewer operational headaches. In other words, less AI theater, more business impact.
And with the Natoma acquisition in the mix, the message is clear: Snowflake is not just building smarter AI on top of enterprise data. It is building toward agents that can understand business context, connect into workflows, and actually help move work forward.
Here are the biggest takeaways.
Everyone wants AI that can answer business questions. The problem is that most AI does not really understand the business.
That is why Horizon Context matters. It is designed to give AI a fuller picture of the enterprise data estate — not just access to raw tables, but the semantics, lineage, metadata, and governance needed to make answers trustworthy.
That trust is where the business value shows up. When people believe the answer, they move faster. Analysts spend less time validating. Leaders spend less time waiting. Teams spend less time debating whose dashboard is right. The result is not just better AI; it is faster, more confident decision-making.
One of the most important Summit themes was that enterprise AI cannot live in a data platform bubble. The business runs across apps, systems, approvals, messages, tickets, customer records, planning tools, and a thousand “just checking in” Slack threads.
That is where the Natoma acquisition fits the bigger story. It reinforces Snowflake’s push toward agentic workflows that do not just summarize what happened, but help coordinate what happens next.
Because the real productivity gap is rarely the insight itself. It is everything after the insight: finding the right system, triggering the right workflow, notifying the right person, updating the right record, and making sure the action is governed and auditable.
Natoma gives the narrative more teeth. Snowflake is not merely saying, “Ask questions of your data.” It is moving toward, “Let governed agents help execute the work your data points to.” That matters for the business because the value of AI compounds when it shortens the distance between knowing and doing.
Snowflake already has “Snowpipe Streaming” as well as a “Kafka Connector” for it that supports ingesting streaming data from externally hosted Kafka services. Now, DataStream provides a Snowflake-managed Kafka service itself.
DataStream is Snowflake’s fully managed, Kafka-compatible streaming service, which is a fancy way of saying: streaming data can move into Snowflake without customers managing a separate maze of brokers, connectors, vendors, bills, and brittle pipelines.
That matters because every company says it wants real-time analytics and AI, but very few want to fund an engineering team whose full-time job is keeping the streaming architecture alive.
The Lighthouse and Iceberg announcements point to a bigger theme: Snowflake is leaning into interoperability and architectural choice.
For enterprise customers, this is not just a technical preference. It is a cost, control, and risk conversation. Data movement is expensive. Duplicate copies create governance headaches. Lock-in makes every platform decision feel heavier than it should.
Snowflake’s message is that customers should be able to work across engines and formats while keeping governance intact. That kind of openness gives companies more flexibility to modernize on their own terms. It also helps AI initiatives use a more complete view of the business without requiring every dataset to be copied, moved, and reconciled first.
The practical impact is simple: less duplication, lower pipeline and storage costs, and fewer architecture choices that feel irreversible.
Historically, governance has had a branding problem. It often shows up as the department of “not so fast.”
But in an AI world, governance becomes the thing that makes scale possible.
Snowflake’s Trust Center AI security package is aimed at exactly that shift. As agents start doing more than answering questions — taking actions, moving data, triggering workflows — businesses need guardrails, policies, audit trails, and protections built in from the start.
That changes the conversation with security and compliance teams. Instead of asking them to bless a black box, the business can show who did what, what the agent was allowed to do, what was blocked, and how activity is monitored. That turns governance into an accelerator, because leaders can expand AI usage without expanding uncertainty at the same rate.
Adaptive Compute is one of those announcements that sounds technical but lands squarely in the business pain zone.
No executive wakes up excited to discuss warehouse sizing. They care that analytics are fast, operations are efficient, and cloud bills do not create surprise meetings with finance.
Adaptive Compute is designed to remove more of the manual configuration and tuning required to run workloads efficiently. That means teams can focus less on managing infrastructure and more on delivering outcomes. Faster queries mean faster reporting cycles. Better utilization means less waste. Less operational overhead means technical teams get time back.
In business terms, this is Snowflake taking more of the complexity tax out of the system.
Snowflake Intelligence is becoming CoWork, and the positioning is important: this is not just for technical users. It is aimed at knowledge workers — analysts, operations teams, business users, and leaders who need answers and actions without writing SQL.
That is where the productivity story gets interesting. If a finance leader can investigate cash flow, a sales leader can explore pipeline risk, or an operations team can surface issues and trigger next steps without waiting in a queue, the business changes its operating rhythm.
CoWork matters because it brings AI closer to the people making daily decisions. It gives them a governed way to ask, reason, create, and act from one surface. Add Natoma’s workflow DNA to that picture, and the story becomes even more compelling: the future is not just AI that tells business users what is happening, but AI that helps them move the work to the next step.
The payoff is not simply “more self-service analytics.” It is shorter cycles between question, insight, decision, and action.
Cortex Code is becoming CoCo, Snowflake’s coding agent for builders. CoCo meets engineers where they work…with Snowsight, CLI, and Desktop interfaces. To each their own.
This is where the Summit story connects directly to delivery velocity. Most companies are trying to modernize platforms, migrate workloads, build pipelines, and operationalize AI at the same time. The backlog is enormous. The talent is precious. The timelines are aggressive.
CoCo is aimed at giving builders leverage. If teams can generate, refactor, test, and ship data workflows faster — while staying inside Snowflake’s governed environment — customers can accelerate modernization without turning every initiative into a custom engineering marathon.
That is the kind of productivity gain that shows up in reduced delivery cost, faster migration timelines, and more business initiatives making it out of the backlog.
The hosted MCP connectors and gateway are easy to underestimate, but they may be one of the clearest signs of where enterprise AI is headed.
The next phase is not just asking AI for an answer. It is asking AI to take the next step across the systems where work already happens — Salesforce, Slack, email, calendars, tickets, and more.
That matters because business value often leaks out in the handoff between insight and action. Someone sees a dashboard, writes a note, opens a ticket, sends a follow-up, schedules a meeting, updates a CRM field, and then hopes the loop closes.
MCP connectors, paired with the Natoma acquisition, point toward a more connected workflow layer where governed agents can help execute across tools with permissions and auditability. That means less swivel-chair work, fewer dropped balls, and a faster path from “we found something” to “we did something about it.”
The Atrium Lounge was a big hit at Summit, giving attendees a premium place to recharge, connect, and keep conversations moving just steps from Moscone at the W San Francisco.
Hendrick Automotive Group, National Debt Relief, Snowflake, and Atrium discussed how Snowflake helps teams connect data across systems to deliver more seamless, data-driven customer journeys.
Salesforce, Sony Interactive Entertainment, DraftKings, and Atrium joined the executive conversation on how enterprises are using Snowflake Cortex and CoCo (Cortex Code) to accelerate AI adoption, streamline data operations, and move from insight to action.
Together, the lounge and sessions reflected the bigger Summit theme: AI and data strategy become more valuable when they connect directly to real people, workflows, and business outcomes.
The biggest takeaway from Summit is not any single product name, although there were plenty of them.
The real story is that Snowflake is making enterprise AI operational.
That means AI needs trusted context. It needs live data. It needs open architecture. It needs security controls. It needs cost governance. It needs to meet business users and builders where they already work. And increasingly, it needs to act across the systems where business actually happens.
That is why Natoma matters in the broader Summit narrative. It helps signal that Snowflake’s agentic ambition is not limited to answering better questions inside the data cloud. It is about connecting trusted intelligence to enterprise workflows so teams can move from insight to execution faster, with governance built in.
Because the companies that win with AI will not be the ones with the flashiest demos. They will be the ones that can safely, reliably, and repeatedly turn data into decisions and decisions into action.
Snowflake’s Summit message was clear: the agentic enterprise is not built on prompts alone. It is built on trusted data, governed workflows, open systems, and teams that can move faster without breaking everything.
A little less magic trick. A lot more operating model.
Get practical advice on Snowflake, Salesforce, and AI-assisted delivery delivered to your inbox. No fluff, just insights from our team on what’s new and next.
4055 Valley Commons Drive
Bozeman, MT 59718
Airport Plaza, Plot No. 15, 1st Floor,
Tonk Road, Beside Cocoon Hospital,
Jaipur, Rajasthan, India 302016
© 2018-2026 Atrium. All Rights Reserved
Leave a Reply