How Snowflake and CoCo Turned Game Art Planning Into a Conversation

When Game Art Planning Becomes a Monster

Every game art team knows the drill. The scope grows, the deadlines tighten, and the spreadsheet that was supposed to keep track of every asset, every variant, every texture set—it starts to creak. You add tabs. You patch formulas. You bolt on new logic to handle the latest art direction change. The model still works, but it's a Frankenstein. It's fragile, impossible to audit, and it eats hours of your week.

That's exactly where the financial planning team at Snowflake found themselves. Their long-range plan had grown to cover 40 entities, each with over 100 cost centers, plus hundreds of expense categories. They needed that granularity because the plan wasn't just for finance anymore. Tax needed legal entity breakdowns. Treasury wanted a cash view. HR needed headcount assumptions. Executives wanted to see the trade-offs between growth, margins, and free cash flow.

Sound familiar? For a game art department, swap 'tax' for 'character pipeline' and 'cash view' for 'render budget.' The core problem is the same: one model, many stakeholders, each with a different lens.

From Spreadsheet to Platform

Snowflake's answer was to rebuild their long-term planning model on Snowflake itself, with Streamlit as the UI layer. They called it Snowplan. It's not a dashboard—it's a planning platform. Analysts update assumptions in an editable Streamlit interface, the changes write straight back to Snowflake, and the model recalculates instantly. No more file versioning hell. No more broken formulas. The source of truth lives in one governed place.

For game art, imagine the same shift. Instead of a master asset list in a shared drive, you have a living database. Every asset, every variant, every LOD is a row. Art leads tweak priorities in a web app, and the downstream impact—on render time, on memory, on team capacity—updates in real time.

Why Putting the Model Where the Data Lives Changes Everything

The key decision was to build the model where the data already lived. For Snowflake, that meant the model sat next to their actuals, their governed data models, their permissions. No manual data pulls. No reconciliation. The model was embedded in the same environment as the finance data, logic, and history.

That brings three game-changing advantages:

• Scale: A ten-year forecast with entities, cost centers, expense categories, headcount, revenue, and balance sheets is a heavy load. Snowflake eats that for breakfast. For game art, that's thousands of assets, each with multiple states, across a multi-year production cycle.
• Governance: Role-based access and row-level security mean the art director sees the whole pipeline, but a junior artist only sees their own tasks. No more exporting different versions for different stakeholders.
• Reusability: Once the architecture is in place, it can support multiple workflows. Snowplan now handles headcount planning, equity modeling, treasury forecasting, hedging, legal entity projections, and M&A scenarios. For game art, the same platform could manage asset tracking, outsourcing, outsourcing budgets, and even narrative planning.

CoCo Makes Scenario Planning a Conversation

Streamlit made Snowplan scalable and usable. Snowflake CoCo made it conversational. Before CoCo, users still had to know where to click, which assumption to tweak, and how to interpret downstream effects. With CoCo, you just ask.

Instead of hunting through tabs, you can say, 'Compare the latest forecast to the one we showed the board last quarter and summarize the key drivers.' Or, 'What changed between this version and the last? What's the net impact on margins?' CoCo turns what used to be a manual, spreadsheet-comparison exercise into a dialogue.

In game art, imagine asking, 'What if we delay the third boss fight to next quarter? Show me the impact on texture budget and team allocation.' CoCo can pull from the data, run the scenario, and give you a clear answer—without anyone having to rebuild the model.

A Real Example: Planning for a Tax Change

Snowflake's team tested this with a potential tax change. In the old world, they'd start with a meeting. Define the problem, pull the data, build assumptions, update the model, review outputs, run sensitivity tables, and then decide who else to involve.

With CoCo, it became fluid. They asked CoCo to summarize the potential tax change. Then they asked it to create a new forecast version assuming the change passes. That immediately sparked the kind of back-and-forth a finance team would have in a meeting: Is this tax passed on to customers or absorbed as margin? What portion can realistically be passed on? Which sales are affected? What's the impact on revenue, gross margin, operating margin, and free cash flow?

Because the analysis sits on Snowflake tables, CoCo could identify which sales were affected, output the financial impact, and highlight the key metrics behind the numbers. It even generated sensitivity tables showing how operating margin would dilute under different pass-through rates.

Most importantly, CoCo flagged risks. It noted that the first-order model didn't include the extra indirect costs of compliance—the people needed to support filings, maintain datasets, or handle new reporting obligations. That's exactly the kind of pushback a good financial partner should give before a scenario becomes a conclusion.

And then CoCo went further. It drafted an email to the tax team, summarizing the analysis, key assumptions, open questions, and decision points. The system wasn't just spitting out numbers—it was helping to structure the problem, identify the right people, and move the process forward.

Why This Matters for Game Art Teams

Game art teams face the same kind of real-time strategic questions. What if we cut the open-world map in half? What if we add a co-op mode? What if we outsource more of the environment art to free up internal capacity? These aren't hypotheticals—they're the questions that come up in every production review.

Traditional tools—the giant asset lists, the static schedules—aren't built for that level of iteration. They're built to produce a plan, not to support an ongoing conversation. By building a planning platform on Snowflake with Streamlit, and adding CoCo, Snowflake made planning scalable, governed, and conversational.

The result: finance teams spend less time updating models and reconciling versions, and more time challenging assumptions, aligning executives, and shaping strategy. For game art, the same shift means spending less time on admin and more time on creative direction.

Trust Is the Foundation

None of this works if the numbers aren't trustworthy. CoCo doesn't generate predictions out of thin air. It interacts with the same governed data, assumptions, and logic that power Snowplan. Every scenario is versioned, every change is reviewable, and access control follows the app's role model.

This is critical. You're not asking leaders to trust a black box. You're using AI to operate a well-governed platform where data, logic, permissions, and outputs are visible, explainable, and auditable. That's what makes AI viable for enterprise finance—and for game art production.

The Path Forward

Snowplan started as a solution to one problem: long-term planning. But because it was built as a platform, it's now reusable across multiple workflows. The same foundation supports headcount planning, equity modeling, treasury forecasting, and M&A scenarios.

For game art, the path is clear:

• Step one: Move your asset data to a central, governed platform.
• Step two: Build an intuitive interface for your team.
• Step three: Add AI to make planning conversational.

The real ROI isn't about making your team more technical. It's about giving them more time for judgment. Long-term planning isn't just about maintaining a giant spreadsheet. It's about understanding where your game is going, where to invest, how to balance ambition with feasibility, and which risks are emerging.

Snowflake and Streamlit gave their finance team a platform that scales, governs, and connects to live data. CoCo made it faster, more interactive, and more strategic. Game art teams can do the same.

The result? Less time updating asset lists and tweaking assumptions. More time sitting with your leads and directors, iterating on the creative vision that actually shapes the game.