# CLA and Databricks Enhance Document Processing Efficiency with Lakebase Postgres

> Discover how CLA’s collaboration with Databricks revolutionizes document processing with a streamlined orchestration solution, dramatically reducing extraction times and enhancing operational efficiency.

**Source**: databricks.com | **Published**: 2026-07-22 | **Type**: case_study

## Key Facts

- CLA's solution cuts document extraction time from hours to minutes, enhancing operational efficiency.
- Lakebase Postgres enables autoscaling, reducing costs by optimizing compute resources during peak loads.
- Real-time dashboards provide visibility into task status and costs, improving decision-making for finance teams.
- Eliminating external infrastructure reduces integration complexity, positioning Databricks as a competitive advantage.
- Rate-limit-aware scheduling prevents task overload, ensuring consistent performance across high-volume workloads.

## Summary

\## Summary
CliftonLarsonAllen LLP (CLA), a professional services firm, faced challenges in auditing processes that required extensive document review and extraction. To address this, they implemented a Databricks-native solution using Lakebase Postgres, which reduced document extraction time from hours to minutes without sacrificing quality.

\## Background
CLA operates in the professional services industry and has a growing global presence. Before deploying the new solution, their auditing process was labor-intensive and time-consuming, relying on manual document review and extraction, which often took hours to complete.

\## Challenge
The primary challenge was to accelerate the document auditing process, which involved converting large volumes of documents, such as contracts and invoices, into structured data efficiently. The existing method was slow and cumbersome, leading to delays in processing.

\## Solution
CLA collaborated with the Databricks Forward Deployed Engineering team to develop a document processing application built entirely on Databricks. This application utilized Lakebase Postgres as the orchestration backbone, managing long-running tasks, retries, cost attribution, and real-time visibility. The architecture included a web application for user interaction, a task management system in Lakebase, and AI agents executing the parsing work.

\## Results
The implementation of the Lakebase-powered orchestration layer reduced document extraction time from hours to minutes. This significant improvement streamlined the auditing process and enhanced operational efficiency.

\## Key Insights
Organizations can benefit from integrating orchestration and observability into a single platform, reducing the complexity and overhead associated with managing multiple systems. Utilizing a unified architecture, like the one built on Databricks and Lakebase, can lead to substantial efficiency gains in processing workflows.

\## Customer Testimonial
No direct quote is available from the source material.

## Entities

- **Companies**: CliftonLarsonAllen LLP, Databricks
- **Products**: Lakebase Postgres, Databricks Apps, Lakeflow Jobs, MLflow, Unity Catalog Volumes
- **Technologies**: Postgres, Server-Sent Events, FastAPI
- **People**: Li Yu, Michelle JanneyCoyle, Jon Cormack, Yarri Bryn, Alec Sorensen, Darshana Nair

## Key Concepts

AI agent orchestration, document processing, task queue, real-time observability, cost attribution, rate-limit-aware throttling, concurrent task management, idempotent callbacks

## Definitions

- **Orchestration Layer**: The orchestration layer coordinates long-running tasks, manages retries, attributes costs, and provides real-time visibility in the system.
- **Rate-Limit-Aware Throttling**: A mechanism that limits the number of concurrent tasks based on the processing capabilities of AI model endpoints to prevent overload.
- **Idempotent Callbacks**: Callbacks that can be safely retried without causing duplicate processing or billing, ensuring consistent results.
- **Real-Time Observability**: The ability to monitor and track the status and performance of tasks in real-time without significant overhead.
- **Cost Attribution**: The process of assigning costs to specific tasks or agents based on their resource usage and processing time.

## Use Cases

- Document parsing for contracts and invoices
- Long-running agentic workloads management
- Real-time task status monitoring
- Cost tracking for AI processing tasks
- Concurrent task execution with priority management
- Automated retries for failed tasks

## Frequently Asked Questions

**What is Lakebase Postgres?**

Lakebase Postgres is an autoscaling Postgres database designed to manage orchestration, queueing, and observability for long-running tasks in a Databricks environment.

**How does the orchestration layer improve task management?**

The orchestration layer simplifies task management by coordinating retries, attributing costs, and providing real-time visibility, all within a unified platform without the need for external systems.

**What are the benefits of using Databricks for AI workloads?**

Databricks offers a fully integrated environment for managing AI workloads, allowing for scalable compute, real-time observability, and streamlined cost attribution, which enhances operational efficiency.

**How does rate-limit-aware throttling work?**

Rate-limit-aware throttling ensures that the orchestrator limits the number of concurrent tasks based on the processing capabilities of AI endpoints, preventing overload and ensuring smooth operation.

**What is the significance of idempotent callbacks?**

Idempotent callbacks are crucial for ensuring that tasks can be retried without the risk of duplicate processing or billing, which is essential for maintaining data integrity and cost accuracy.

## Links

- [Read on Welcome.AI](https://welcome.ai/content/cla-and-databricks-enhance-document-processing-efficiency-with-lakebase-postgres)
- [Original source](https://www.databricks.com/blog/simplify-ai-agent-orchestration-lakebase-postgres)
- [Databricks](https://welcome.ai/company/databricks): Featured company

---

Source: Welcome.AI | https://welcome.ai/content/cla-and-databricks-enhance-document-processing-efficiency-with-lakebase-postgres