Fully ramped: The Analyst Agent is now available in Pigment

Today, we’re lifting the lid on the newest member of Pigment users’ teams: the Analyst Agent.

So how does it work?

Imagine you’re a hard-working CFO. Q2 performance was solid, but there were some gaps in the guidance you gave Wall Street.

With your earnings call just days away, you and your team need a story for your investors - what happened, why, and how you’re going to get back on track.

Welcome to mission control

Rather than conducting the work yourself, you recognize that variance analysis is exactly the sort of routine, time-consuming task that you’re better off delegating. But not to a member of your team - to the Analyst Agent.

To do so, you craft your prompt - known in Pigment as a mission.

INTRODUCTION

Here, you provide the Analyst Agent with a brief overview of what it’s expected to do.

You are tasked with conducting a monthly variance analysis for the company. Your goal is to identify and explain key differences between actual and budgeted financial performance for the latest period.

DATA SCOPE

Next, you provide a list of the metrics (1) and dimensions (2) the agent should use in its analysis.

1. #️⃣Net Income, #️⃣Total Revenue, #️⃣Total Gross Margin, #️⃣OPEX.
2. ℹ️Version  = "Actual", "AOP / Budget". Do not include ℹ️Version = "Q2 Forecast" or ℹ️Version = "Q3 Forecast"

INSTRUCTIONS

Here you provide a step-by-step process for the agent to follow.

1. Calculate the absolute and percentage variance between actual and budgeted values for each .
2. Identify any variances greater than ±10% or $5,000 (whichever is higher).
3. For each significant variance, generate a brief explanation based on historical trends or typical patterns (if available).
4. Provide a summary table showing
   
   1. Metric
   2. Actual
   3. “AOP / Budget”
   4. Variance ($)
   5. Variance (%)
   6. Flag (Significant/Not Significant) (use emojis)
   7. Explanation
5. Summarize key findings
6. Do not include charts
7. Make recommendations about what focus on in future forecasts

OUTPUT FORMAT

This section tells the agent what you’d like the output to look like.

Deliver results in four sections, with double line bold section dividers:

1. "Conclusions"
2. "Detailed Variance Analysis"
3. "Summary Table"
4. "Recommendations for focus"

ASSUMPTIONS

Finally, provide any further details that the agent may find helpful.

• All amounts are in USD
• Time period is 🗓️Analysis Month.
• Use prior 6-month trends if needed for context

Once that's done, you're ready to go.

Teamwork in action

Clicking the ‘preview’ button will start the mission.

As you can see, the Analyst isn’t the only agent in town. Pigment employs a multi-agent framework, in which specialized agents handle different parts of the workload. 

This means that in this case, the Analyst works alongside another agent: the Supervisor. The Supervisor is a sort of ‘manager’, there to coordinate the work of other agents and maintain quality.

Together, the two work to produce the final report.

Real-world impact

While today will be the first time many get their hands on agents in Pigment, we have some success stories we’d like to highlight from a few innovators that’ve been using them in the private preview program.

Carta automates management reporting

Carta’s finance team have been using the Analyst Agent to produce weekly ARR summary reports for their leadership team: explaining sources of growth and flagging areas that require attention. 

In addition, they’re using the agent to perform deep dive analysis of specific departments and cost centers - for example, to dig into why marketing spend has fluctuated vs expectations.

They estimate they’ve saved:

2 hours per week, per team member on weekly ARR reporting

10-15 hours across the team, per month on monthly departmental deep dives

“Pigment's agentic AI capabilities make us feel like 'mini CFOs' within our departments. By taking away the often repetitive grunt work, our team is free to focus on strategy and driving impact.”

Montaek Kohli, Strategic Finance Manager

ClickUp makes dramatic time savings on BvA

With the Analyst, ClickUp has automated the first draft of monthly budget variance analysis reports. It helps them understand, even when at a glance there isn’t a variance vs. plan, whether there are underlying drivers which have changed. 

For instance, the overall costs of a team might be as planned - but the mix of expenses that got them there might be different. That type of insight is super valuable when it comes to forecasting and planning.

The team is saving 1-1.5 days per team member, per month.

“At ClickUp, AI isn’t a future ambition. It’s a strategic priority we’re acting on today to drive top line and bottom line results. That’s why Pigment’s Analyst Agent - and the company’s broader agentic AI roadmap - will play a central role in helping our finance team deliver on both.”

Zane Olfert, Strategic Finance Manager

Recognition

In a recent report, Dresner has recognized Pigment as a leader in Agentic AI and how finance and business leaders are unlocking the next generation of intelligent planning.

Read the report here.

We’re recruiting

The Analyst Agent is the first step in Pigment’s Agentic AI roadmap, and part of its broader vision to build a modern, intelligent planning platform for the enterprise.

The Analyst Agent will shortly be followed by: 

• The Modeler Agent, which autonomously builds and adapts planning models as conditions change
• The Planner Agent, which simulates scenarios and recommends strategic actions
• Use case specific agents, including a Consolidation Agent, which automates financial consolidation to unify reporting and actuals with FP&A

If you’re a customer and would like to use the Analyst Agent today, please speak to your CSM.

If you’d like to learn more about the Analyst Agent, join the upcoming product tour.