Graphed LogoGraphed
Login

How to Create an Insurance Dashboard in Looker with AI

Cody Schneider•

Creating a comprehensive insurance dashboard can feel complicated, but it provides an essential, real-time view of your agency’s or company's health. By bringing all your key metrics into one place, you can move from reactive problem-solving to proactive strategy-building. This article will walk you through how to use a powerful tool like Looker, with help from AI, to build a dashboard that tracks everything from claims processing to policy renewals.

Why an Insurance Dashboard is Your Operational Command Center

The insurance industry is overloaded with data scattered across different systems - claims are in one system, policyholder information in a CRM, underwriting data in spreadsheets, and financial reports in another tool. Manually pulling this information together for a weekly report is slow, prone to errors, and means you’re always looking at outdated information. A well-designed dashboard solves this.

It acts as a single source of truth, giving you instant answers to critical questions:

  • Are our claims settlement times improving or getting worse?

  • Which policy types are the most profitable?

  • What is our current loss ratio, and how does it compare to last quarter?

  • How are our underwriters performing against their quote-to-bind targets?

Instead of spending Monday morning building reports, you can spend it analyzing trends and making decisions that move your business forward.

Key Metrics to Include on Your Insurance Dashboard

A great dashboard is focused. Cluttering it with dozens of charts creates confusion, not clarity. Start by focusing on the metrics that directly measure the health and efficiency of your core operations.

Claims Performance Metrics:

  • Claim Frequency: The number of claims filed per number of policies in force. A rising frequency could signal issues with underwriting or external factors.

  • Claim Severity: The average cost per claim. This helps you understand the financial impact of claims.

  • First Response Time: How quickly your team contacts a policyholder after a claim is filed. This is a huge driver of customer satisfaction.

  • Claim Settlement Time (or Cycle Time): The average time from when a claim is opened to when it is closed. Faster settlement times improve customer experience and reduce operational costs.

  • Loss Ratio: A critical profitability measure, calculated as (Incurred Losses + Adjustment Expenses) / Earned Premiums.

Underwriting & Sales Metrics:

  • Submission-to-Quote Ratio: The percentage of policy submissions that result in a quote. It measures the efficiency of your initial underwriting process.

  • Quote-to-Bind Ratio: The percentage of quotes that are converted into bound policies. This tells you how competitive your pricing and offerings are.

  • New Business Written Premium: The total premium generated from new policies sold within a specific period.

Policy & Customer Metrics:

  • Policies in Force (PIF): The total number of active insurance policies at a given point in time.

  • Policy Renewal Rate: The percentage of customers who renew their policies. A high renewal rate is a strong indicator of customer loyalty and satisfaction.

  • Customer Lifetime Value (CLV): The total net profit a business can expect to make from a single customer over the entire relationship.

Getting Started with Looker and Your Data

Looker (now part of Google Cloud) is a powerful business intelligence tool well-suited for the complex data in the insurance industry. Its main strength lies in its LookML modeling layer, which essentially acts as a central rulebook for your data. Once you define a metric like "Loss Ratio" in LookML, it will be calculated the exact same way across every report and dashboard, ensuring consistency and trust in your numbers.

Here’s the basic workflow for setting up your insurance dashboard in Looker:

  1. Connect Your Data Sources: First, you need to grant Looker access to where your data lives. This could be a cloud data warehouse (like BigQuery or Snowflake) that consolidates data from your claims processing system, a CRM like Salesforce, your accounting software, and any relevant spreadsheets.

  2. Model Your Data with LookML: This is the most technical part of the Looker setup. You’ll use Looker’s specialized language, LookML, to tell Looker how your database tables are related. You define logical objects, called "explores," and specify dimensions (like "Policy Type" or "Customer Region") and measures (like "Total Claims Paid" or "Average Premium"). This step ensures everyone in your company is speaking the same data language.

  3. Build Dashboards and Visualizations: With the LookML model in place, your team can start creating visualizations. Users can explore the data you’ve defined, mix and match dimensions and measures, filter results, and save their charts (called "Looks") to a dashboard for easy monitoring.

Traditionally, step two—modeling with LookML—is where most teams hit a bottleneck. It requires specialized knowledge and can take a data analyst days or weeks to get right. This is where AI can step in and significantly speed up the process.

Using AI to Accelerate Dashboard Creation in Looker

While Looker has powerful capabilities, building your first dashboard isn’t as simple as clicking a few buttons. An insurance agency owner or a marketing manager can't just jump in and start writing LookML. However, by using AI as a "coding assistant," you can bridge that gap.

Think of it as having a junior data analyst you can ask for help anytime. You can use large language models (like ChatGPT, Claude, or Gemini) to translate plain-English requests into the specific LookML or SQL code you need.

A Step-by-Step Guide with an AI Assist

Let's walk through building a part of our dashboard, demonstrating how you might use AI along the way.

1. Define The Goal With Plain English

Instead of staring at a blank screen, you can start by asking the AI for a roadmap. This helps you structure your thinking before you even open Looker.

Your prompt to an AI chatbot could be:

"I am building an insurance dashboard in Looker. My data sources include a 'claims' table and a 'policies' table. The 'claims' table has columns like 'claim_id', 'policy_id', 'date_opened', 'date_closed', and 'amount_paid'. The 'policies' table has 'policy_id', 'policy_type', 'premium_amount', and 'start_date'. What are the most important calculations I should create in LookML?"

The AI can help you brainstorm and give you a list of measures to build, such as total claim count, average claim size, claim settlement time, and loss ratio.

2. Generate The LookML Code Snippets

Now, you can ask the AI to write the actual code for one of those metrics. This turns hours of searching through documentation into a 30-second task.

Your follow-up prompt:

"Great. Write the LookML code for a measure called 'average_settlement_time'. It should calculate the average number of days between 'date_opened' and 'date_closed' from the 'claims' table."

The AI will likely output a LookML snippet that looks something like this:

You can then copy this code directly into your LookML model file. You can repeat this process for every metric you need, asking the AI to write measures for claim count, total premium, policy count, and more. This dramatically lowers the barrier to entry, letting you focus on what you want to measure rather than how to code it.

3. Iterate and Refine Your Visuals

Once you have a few core metrics defined in your model, you can build visualizations in the Looker interface. Let's say you built a chart showing "Average Settlement Time by Policy Type," and it looks too high for "Commercial Auto" policies. Now you want to dig deeper.

You can ask the AI to help you figure out how to drill down further.

Your prompt:

"In my dashboard, I want to allow users to filter the 'average_settlement_time' metric by the claims adjuster assigned to the claim. My 'claims' table has a column called 'adjuster_name'. How would I add this as a filter in LookML?"

The AI can guide you on creating a LookML filter, enabling you to build interactive dashboards where users can slice and dice the data to answer their own follow-up questions right on the spot—without needing to ask an analyst for a new report.

Best Practices for Your Finished Dashboard

Once your dashboard is built, keeping it valuable requires a little more than just data. Follow these principles to make sure it's actually used.

  • Tailor Views for Different Roles: Your CEO needs a high-level overview of profitability and growth. A claims manager needs a detailed view of their team's workload and cycle times. Create separate dashboards or sections for each key role in your organization.

  • Prioritize Simplicity: A good dashboard communicates information at a glance. Place the most important KPIs (like overall Loss Ratio or New Business Premium) at the top in large, clear numbers. Use visual aids like color-coding (red for bad, green for good) to draw attention to performance.

  • Tell a Story with Your Data: Don't just show numbers, show trends. Use line charts to display metrics over time. For example, a chart showing the "Quote-to-Bind Ratio" month-over-month tells a much more compelling story than a single number.

  • Incorporate Clear Drill-Down Paths: Your dashboard should be the starting point for exploration, not the end. Looker excels at this. Configure your dashboard so that when a user clicks on a number—like "150 new claims this week"—they are taken to a detailed report showing the specifics of those 150 claims.

Final Thoughts

Building a high-impact insurance dashboard in Looker provides the clarity you need to run your business effectively. By unifying data from claims, underwriting, and customer systems, you can spot trends earlier, optimize processes, and ultimately improve profitability. While traditionally a technical task, using AI as an assistant can dramatically simplify the creation of LookML models and help you get to insights much faster.

For many teams, even with AI assistance, the learning curve and setup for traditional BI tools like Looker can still be a significant hurdle. At Graphed, we created a platform that removes these barriers entirely. Instead of struggling with LookML, you can securely connect your data sources in just a few clicks. From there, you just ask questions in plain English, like "Show me a chart of our loss ratio by month for the past year" or "Create a dashboard comparing an underwriter's submission-to-quote and quote-to-bind ratios," and Graphed builds the interactive, live-updating dashboards for you instantly.