Snowflake Finance Cuts Monthly Forecast From a Week to Hours With CoCo

From a weeklong process to a midmonth update

The sales finance team runs more than 30 recurring analyses each month for Snowflake's sales leaders. They cover the topline forecast, sales pipeline, commissions and business performance reporting. The forecast is the most involved, and it comes with a variance analysis explaining what has changed since the previous forecast.

Previously, building it meant reviewing pipeline deals, reading account notes in the CRM, comparing against historical trends and checking current field behavior. The work was split across several analysts, and because it took more than a week, it could only run once a month. With the topline forecast skill in CoCo, the same analysis now takes a few hours, according to Snowflake.

That has changed the cadence. The team now publishes a midmonth update so sales leaders can see where the quarter is heading while there is still time to act. Any analyst can also rerun the skill to test a scenario.

The GTM Finance profile and why it lowers cost

Early attempts with AI were inconsistent. The team says the AI would reference a related but incorrect data source or ask for clarification that an experienced analyst would not need, because it had to infer things the team knew but had never written down.

The fix was a CoCo profile called GTM Finance, built by Snowflake's strategic finance analytics team. A profile is standing context the AI reads before it answers, so nobody has to restate it in a prompt. According to Snowflake, it holds these layers:

• Certified data sources: 20+ sources certified by the analytics team.
• Terminology: definitions for bookings, pipeline, quota and other core metrics.
• Repetitive analyses: recurring monthly and quarterly work.
• Output formats: standard report and variance layouts.
• Escalation paths: where to go when a number looks wrong.

Snowflake calls the efficiency effect memory scaling: an AI agent gets cheaper per unit of work as it accumulates context about its environment. Across 44 finance users, those with a profile reached an answer in about 1.3 steps, compared with about 3 steps for those without one (Snowflake internal CoCo usage, February to May 2026). Each extra step adds about 0.05 credits, roughly 10 cents, so a five-step session costs about three and a half times a one-step session. Snowflake also says that at every step count in its data, a session with the profile cost less than one without it.

Analysts build 36 skills of their own

With the profile in place, analysts began building their own skills, which are reusable workflows the AI runs on request. Snowflake says building one requires no technical background. In two months the library grew from three starter skills to 36, and several Excel workbooks became shared skills. Three months after launch, 87% of first adopters were still using CoCo daily.

• Bookings BvA: replaced manual month-end variance analysis across data and finance. An analyst types "bookings bva", answers a few parameter prompts and gets variance across deal categories in minutes, including which opportunities slipped and why.
• Sales travel expense report: replaced hours of cross-team reporting and manual email drafting. It sets the reporting window itself, produces per-leader insights on budget variance, out-of-policy spend and top spenders, and drafts the email to each leader. Work that took hours across two teams now takes about 15 minutes.
• Topline forecast: replaced a weeklong multiperson process.
• Commission reconciliation: replaced an Excel-based manual workflow.

Keeping control and what it means

Letting analysts write their own logic raises a control question. Snowflake says three measures keep it in bounds. The analytics team certifies the source of truth through a technical source guide, so follow-up questions and ad hoc analysis draw on the same sources. The context layer is owned centrally, with analysts building on top of the profile while the analytics team defines data sources and terminology. And skills are shared and inspectable, which the company says is a stronger review position than a spreadsheet on one laptop.

The results are Snowflake's own, drawn from its internal use, and the post is written from the team's perspective. Still, the account points to a pattern for other companies: platform experts set the foundation, and a functional team builds its own workflows on it. The post says the approach applies beyond finance.