How Health Systems Can Turn Long-Term Planning into a Living Conversation

The Planning Problem Nobody Talks About

Long-term planning is the kind of work that quietly eats a finance team's calendar. At Snowflake, the 10-year forecasting exercise involves more than 40 legal entities, each with over 100 cost centers, and hundreds of spending categories. That granularity isn't a luxury—it's a requirement. Tax teams need to see the split between goods and services, the legal entities involved, and the jurisdictions that matter. Treasury wants a cash view. Workforce planning needs headcount assumptions. Executives want to understand the trade-offs between growth, margin, investment, and free cash flow.

Health informatics teams face a similar pressure. A hospital network's long-range financial model has to serve clinical departments, supply chain, compliance, and strategic leadership all at once. The default response, in any industry, is to build a giant Excel workbook. It works—until it doesn't. Tabs pile up. Formulas get patched. New logic layers on old logic. The model becomes more valuable and more fragile at the same time.

From Spreadsheet Monster to a Real Platform

Snowflake's finance team hit that wall and decided to rebuild. They moved their long-term planning model onto Snowflake, using Streamlit as the interface layer. The result was Snowplan, an internal application that lets analysts and executives interact directly with forecast outputs. The goal wasn't to build a dashboard. It was to build a planning platform—one that felt familiar to finance users but had Snowflake's governance, scale, and compute underneath.

In Snowplan, analysts update assumptions through an editable Streamlit interface. Those changes write back to Snowflake, the model recalculates, and the new output appears immediately. No corrupted formulas. No version-control guessing games. The model is connected to actual data sources, so actuals flow in automatically. Assumptions are versioned. Scenarios can be compared. Different roles see different levels of detail—analysts get fine-grained input pages, managers get visibility into logic changes, and executives see the consolidated P&L and free cash flow.

Why Building on the Data Platform Changes Everything

The key decision was to put the model where the data already lived. Because Snowplan runs on Snowflake, it connects directly to governed data models and raw sources. That means no manual updates, no reconciliation of offline pulls. The model sits in the same environment as the finance data, permissions, and audit history.

This pays off in a few ways. First, scale. A forecast that spans ten years, multiple entities, cost centers, spending categories, headcount, revenue, balance sheet, and free cash flow generates a lot of data—exactly the load Snowflake handles well. Second, governance. Role-based access and row-level security mean each user sees only what they should. An executive doesn't need the same interface as an analyst, and analysts don't have to export separate versions for every stakeholder. Third, extensibility. The same platform can support workforce planning, equity modeling, treasury cash forecasting, hedging, legal entity forecasts, COGS planning, and M&A scenario analysis.

CoCo Turns Planning into a Conversation

Streamlit made Snowplan usable. Snowflake CoCo made it conversational. Before CoCo, a user had to know which page to visit, which assumption to tweak, and how to interpret downstream effects. CoCo changes that. Instead of clicking through screens, you ask a question in plain English.

You can ask CoCo to compare two versions of a forecast and summarize the key drivers. You can ask what changed between the version you showed the board last year and the one you're preparing now. You can ask about the net impact, the drivers of margin expansion or dilution, and which assumptions are most sensitive. This is powerful in executive planning because the real question isn't “Can you get me the latest numbers?” It's “What changed, why, and what does that mean for our narrative?” CoCo compresses that analysis into a dialogue.

A Real Example: Scenario Planning for a Tax Change

Consider a potential tax change. In the past, this would start with a meeting. The tax team would define the affected sales, pull data, build assumptions, update the model, review outputs, and create sensitivity tables. With CoCo, the process flows more smoothly. You can ask CoCo to summarize the potential tax change. Then you ask it to create a new forecast version that assumes the change becomes law.

CoCo can identify which sales would be affected, output the financial impact, and show the key metrics behind the numbers. It can also create sensitivity tables showing how operating margin would dilute under different pass-through rates. Just as importantly, it flags risks—for example, the first-order model may not include the indirect costs of compliance or new reporting obligations. That's the kind of caution a good finance partner raises before anyone treats a scenario as a conclusion. CoCo can even draft an email to the tax team summarizing the analysis, key assumptions, and open questions.

Why Trust Is the Real Foundation

For finance teams, conversational planning only works if the numbers are trustworthy. That's why architecture matters. CoCo isn't generating numbers in a vacuum. It's interacting with the same governed data, assumptions, and logic that power Snowplan. Every scenario is versioned, every modification is reviewable, and access control follows the application's role model. Analysts and executives can compare before and after, understand the changes, and roll back if needed.

This is a critical distinction. You're not asking leaders to trust a black box. You're using AI to operate a well-governed planning platform where data, business logic, permissions, and outputs are visible, explainable, and auditable. That's what makes AI viable in enterprise finance—and by extension, in health informatics.

From a Planning Tool to a Strategic Platform

Snowplan has already outgrown its original use case. Because the model lives on Snowflake, the architecture is reusable. The same foundation now supports multiple planning workflows: headcount planning, equity modeling, treasury cash forecasting, hedging, legal entity forecasting, COGS planning, and M&A scenario modeling. Each new workflow reuses the same governance backbone, connects to the relevant data sources, and exposes a user-friendly Streamlit interface. With CoCo, each workflow becomes queryable and adjustable through natural language.

I suspect more finance teams will follow a similar path: first, move the model to where the data lives. Second, build an intuitive app layer for users. Third, use AI to make planning conversational. For health systems, this could mean moving long-range financial models out of spreadsheets and into a governed platform that connects to clinical and operational data. It could mean answering questions like, “What if we open a new clinic?” or “What if supply costs rise 5%?” in real time, with defensible numbers.

The Real ROI Is Time for Judgment

The real ROI of Snowplan isn't that finance teams become more technical. It's that they get time back for judgment. Long-term planning should not be about maintaining a massive workbook. It should help teams understand where the business is going, help leaders decide where to invest, how to balance growth and profitability, what risks are emerging, and which trade-offs matter most.

Snowflake and Streamlit gave the team a platform that scales, is governed, and connects to real-time data. CoCo made it faster and more interactive. The result is that finance teams spend less time updating models and reconciling versions, and more time with executives iterating on long-term strategy. For FP&A teams, that's the real value of AI in planning. It doesn't replace the finance function. It removes the manual labor that slows it down, so finance can focus on what it's actually supposed to do—challenge assumptions, align leadership, and shape the future.