How to Create a Work in Progress Report in Looker with AI

A Work in Progress (WIP) report is one of the most powerful tools for keeping projects on budget and on schedule. But building one in a sophisticated business intelligence tool like Looker has traditionally been a job for data analysts. This guide will walk you through how to use Looker's AI capabilities to create a detailed WIP report, even if you’ve never written a line of code.

Graphed

Build AI Agents for Marketing

Build virtual employees that run your go to market. Connect your data sources, deploy autonomous agents, and grow your company.

Build Your First Agent

What Is a Work in Progress (WIP) Report, Anyway?

Think of a WIP report as a real-time snapshot of any project, initiative, or job that’s currently underway. It’s not just a to-do list, it’s a financial and operational overview that helps teams understand progress, identify bottlenecks, and manage resources effectively. Without this view, you're essentially flying blind, letting gut feelings guide decisions on staffing, budgets, and timelines.

WIP reports aren't just for one industry. They're essential everywhere:

• Marketing Agencies: Tracking the status of client campaigns, including creative development, media spend vs. budget, and milestone completion.
• Software Development: Monitoring sprints, tracking feature development progress against the roadmap, and managing engineering resources.
• Manufacturing: Following a job from raw materials to a finished product, accounting for labor hours, machine time, and material costs at each stage.
• Consulting Firms: Keeping tabs on client projects, tracking billable hours logged against the project scope, and ensuring deadlines are met.

Key Components of a Standard WIP Report

While the specifics vary, most effective WIP reports track a few core metrics. Your goal is to combine operational data (what’s getting done) with financial data (what it’s costing). A great report typically includes:

• Project/Job Identifier: The name or ID of the project (e.g., "Q4 Website Redesign," "Job #542 Pendants").
• Current Status: The current stage of the project (e.g., "In Development," "Awaiting Client Feedback," "Blocked").
• Assigned Owner: The person or team responsible for the project's completion.
• Percentage Complete: An estimate of how far along the project is, often based on tasks completed, milestones reached, or hours logged.
• Start and Due Dates: The planned timeline for the work.
• Budgeted Costs: The total amount allocated for the project. This can be broken down into labor, materials, software, etc.
• Actual Costs to Date: The total expenses incurred on the project so far.
• Variance: The difference between the budgeted cost for the current progress and the actual costs incurred. This immediately flags if you're over or under budget.

AI Agents for Marketing

Crash Course

Graphed

Free PDF · the crash course

AI Agents for Marketing Crash Course

Learn how to deploy AI marketing agents across your go-to-market — the best tools, prompts, and workflows to turn your data into autonomous execution without writing code.

Download the PDF

The Traditional Way: Building Reports in Looker

Looker is an incredibly powerful platform. Top-tier companies use it to create a single source of truth for their data. The magic behind Looker is its modeling layer, LookML, which allows data teams to define business logic, metrics, and data relationships centrally. Once defined, anyone in the company can explore the data and build reports without writing SQL queries.

However, getting to that point has traditionally required deep technical expertise. Building a WIP report the “old way” involved several complex steps:

1. Data Connection and Modeling: A data engineer connects data sources (like your project management tool, accounting software, and CRM) to a data warehouse. Then, a LookML developer writes code to define the dimensions (like 'Project Name' or 'Task Status') and measures (like 'Total Cost' or 'Hours Logged'). They also have to define the "joins" that tell Looker how tables relate to each other - for example, how a project in Asana connects to its corresponding budget in QuickBooks.
2. Creating an "Explore": Once the LookML model is built, the developer creates an Explore. This is the interactive environment where business users can drag and drop fields from the left-hand panel to build their own reports.
3. Manual Report Building: Using the Explore, a user would manually select the fields they need for the WIP report, apply filters (like "Project Status = Active"), and choose the right visualization format (like a table or chart).
4. Refining and Saving: The user would then tweak the report, maybe adding a pivot for a different view or a custom calculation, before saving it as a "Look" or adding it to a dashboard.

This process is robust and ensures data consistency, but it has a massive bottleneck: it relies entirely on a technical user. If a project manager wants to add "Overdue Tasks" to their WIP report, they can't just do it themselves. They have to file a ticket with the data team, wait for a developer to update the LookML model, and then get the report updated. This process can take days or even weeks.

The Modern Approach: Create Your WIP Report with AI

Recent advances in AI, like the introduction of Duet AI in Looker Studio, are changing the game. Now, instead of navigating complex menus or waiting on a developer, you can use natural language prompts to generate, refine, and analyze your reports. It turns report building from a technical task into a simple conversation.

Here’s how you can approach building your WIP report using this new method.

Step 1: Make Sure Your Data Is Connected

AI can only analyze data it has access to. Before you start prompting, you need to ensure your key data is flowing into Looker. To build a comprehensive WIP report, you’ll likely need data from a few different systems:

• Project Management Tool: Jira, Asana, Monday.com, Trello. This is your source for project status, tasks, owners, and deadlines.
• Financial Software: QuickBooks, Xero, Stripe. This provides the budget and actual cost data.
• Time Tracking Software: Harvest, Toggl. This gives you exact labor hours for cost calculations.
• CRM: Salesforce, HubSpot. This can provide context for client-facing projects, linking work back to specific deals or customer accounts.

Getting this initial setup done might still require some help from your IT or data team, but once the data is connected, the exploration and reporting become far more accessible.

Graphed

Build AI Agents for Marketing

Build virtual employees that run your go to market. Connect your data sources, deploy autonomous agents, and grow your company.

Build Your First Agent

Step 2: Prompt Your Report into Existence

This is where the magic happens. Instead of manually dragging and dropping fields, you simply describe the report you want in plain English. The AI interpreter translates your request into the necessary query and builds the report for you.

Start with a simple, high-level prompt. Let's say you're a marketing director managing several campaigns.

Your first prompt might be:

Show me all active marketing campaigns

The AI will likely generate a simple table with campaign names and perhaps their statuses. Now, you can start layering in more detail with follow-up prompts.

Refine your prompt to be more specific:

Create a WIP report for active marketing campaigns showing campaign name, project owner, percent complete, and the due date. For percent complete, use the number of completed tasks divided by the total number of tasks from our Asana data.

The AI now understands it needs to create a calculated metric and will generate a much more useful report. You can keep going to blend in financial data:

Add two columns to the WIP report: "Budget" from our QuickBooks data and "Actual Spend to Date". Also, add a third column for "Budget Variance" that subtracts actual spend from the budget.

In just a few conversational steps, you’ve built a sophisticated, multi-source WIP report that used to take an analyst hours or days to construct.

Step 3: Visualize and Refine with Follow-up Prompts

Raw data tables are useful, but visuals tell a story. The conversation with AI doesn’t stop after the initial data is pulled. You can continue prompting to shape the output into the perfect format.

Let's say your report is a big table. You can ask for a better way to see the information:

To tweak the visualization:

Change this report into a bar chart showing the percent complete for each campaign. Put the campaign name on the y-axis and the percentage on the x-axis.

To quickly find problems:

In the main table, highlight any campaign where the budget variance is negative in red.

To make the report interactive:

Add a filter to this dashboard so I can view WIP reports by project owner.

This iterative process allows you to explore the data freely. Every question naturally leads to another, and with AI, you can pursue that curiosity immediately instead of getting stuck on a technical hurdle. You’re no longer just building a static report, you're having a conversation with your data.

AI Agents for Marketing

Crash Course

Graphed

Free PDF · the crash course

AI Agents for Marketing Crash Course

Learn how to deploy AI marketing agents across your go-to-market — the best tools, prompts, and workflows to turn your data into autonomous execution without writing code.

Download the PDF

Best Practices for Effective AI-Powered Reporting

Using AI in Looker radically simplifies things, but a few core principles will help you get the most accurate and insightful results.

• Start with a Clear Question: Before you write a prompt, know what you’re trying to answer. Are you checking if projects are on budget? Identifying a bogged-down team? A clear goal leads to a clear and useful report.
• The Cleaner Your Data, the Better the Output: AI cannot fix messy data. If your team isn’t consistent with updating task statuses in Asana or logging hours in Harvest, the AI-generated report will reflect that inaccuracy. Getting clean reports starts with good data hygiene at the source.
• Iterate, Don't Expect Perfection on the First Try: Your first prompt is rarely your last. The power of conversational AI is its ability to build on previous requests. Start broad, then narrow your focus with follow-up questions to drill down into the insights you need.
• Review and Verify: AI is exceptionally good at retrieving and calculating data, but you're still the expert. After the AI generates your report, take a moment to sanity-check the numbers. Does the "Total Cost" for a project seem reasonable? The goal is to let AI handle the heavy lifting, not the critical thinking.

Final Thoughts

Learning how to effectively leverage AI to build reports in platforms like Looker turns a once-intimidating task into an accessible, creative process. Instead of getting stuck on the technical 'how,' you can now focus completely on the 'what' and 'why' behind your data, allowing anyone on your team to answer critical business questions about their work in progress.

Looker's AI functionality points toward a future where data is conversational, but navigating enterprise-level BI tools can still present a steep learning curve. This frustration is exactly why we built Graphed . We wanted to make it dead simple for anyone to connect all their commercial data - from project management platforms like Asana to sales and marketing tools like Shopify and Salesforce - and generate dashboards in a few seconds using only plain English. With our tool, you can create a real-time, cross-platform report that gives you a complete view of your business without ever setting up a data pipeline or learning a complex interface.