# Healthcare case study | CloudFiles for Salesforce > CloudFiles brings intelligent automation to medical assessment approvals through advanced extraction, smart rule-based completion checks, and AI-driven paragraph response ratings. *Case study* ## How a healthcare services firm reduced document approval time by 50% using CloudFiles' Document AI CloudFiles brings intelligent automation to medical assessment approvals through advanced extraction, smart rule-based completion checks, and AI-driven paragraph response ratings. Tags: Healthcare, Document AI, Salesforce - **Industry:** Healthcare, Patient assessment - **Headquarters:** United States - **Products used:** Document AI - **External storage:** SharePoint ### Overview A healthcare organization specializing in patient assessments streamlined its document approval process using CloudFiles' Document AI. With extensive forms containing structured and unstructured data, the team automated data extraction, rule-based validation, and paragraph-level grading – cutting approval time in half while maintaining high accuracy and compliance standards. ### Business challenges - **Manual processing at scale** — forms included extensive patient data, condition reports, and assessment notes, each requiring manual extraction into Salesforce. - **Conditional completion rules** — many fields depended on prior answers, so assessors had to remember and apply complex logic throughout the review. - **Subjective paragraph inputs** — open-ended responses varied widely in quality, so assessors had to benchmark them against internal guidelines, making approvals time-consuming. ### Solution - **AI-powered extraction** — Document AI automatically captured all data types, from checkboxes to long-form text, and updated them into Salesforce custom objects. - **Smart approval workflows** — rule-based flows and prompting validated form completion, adapting to conditional logic with ease. - **AI response rating** — paragraph responses were graded against benchmark examples, helping assessors fast-track evaluations. ### Business challenges, in depth With growing volumes of medical forms, manual review became inefficient and inconsistent. Each form required data extraction, rule-based validation, and evaluation of subjective responses, making approvals slow and error-prone without automation. - **High-volume manual extraction** — "Each assessment form captured detailed patient information – from symptoms to response notes. Processing it manually was a major time drain." Forms contained dozens of structured and unstructured fields that had to be extracted into Salesforce; repetitive data entry slowed down approvals and left room for inconsistency. - **Complex rule-based completion checks** — "Some questions were optional, others conditional – keeping track manually was challenging and error-prone." Even basic validation required applying branching logic based on prior answers, which human assessors had to remember and enforce consistently. - **Subjective paragraph responses** — "One patient writes a full page, another a single line – evaluating quality and completeness wasn't straightforward." Open-ended answers varied widely in depth and clarity, and assessors needed to compare responses against internal benchmarks, adding time and variability to approvals. ### How CloudFiles solved it CloudFiles streamlined the approval process by automating form extraction, applying smart validation logic, and rating paragraph responses using AI, integrated directly into Salesforce. - **Comprehensive AI-based form extraction** — CloudFiles automated the extraction of all form elements – names, dates, checkboxes, and paragraph responses – directly into Salesforce custom objects, ensuring instant readiness for validation workflows, no risk of skipped fields, and support for both typed and scanned inputs. - **Dynamic flow-based approval logic** — Using Salesforce flows and chain-of-thought prompting, CloudFiles enabled smart approval workflows that could skip, validate, or flag responses based on real-time logic, handling conditional dependencies with precision. - **Level 1 AI rating for paragraph responses** — CloudFiles introduced an AI-generated classification layer that scored open-ended responses against benchmark examples, providing a quick reference rating (e.g., low, medium, high detail) that helped assessors prioritize which forms needed attention. ### Rollout & secondary impact The organization adopted a structured, 3-phase rollout designed to deliver incremental value while allowing teams to adapt smoothly – from idea to ROI in just 8 weeks. - **Phase 1 (2 weeks)** — Set up AI-based data extraction. All key fields, including names, checkboxes, dates, and paragraph responses, were automatically mapped and updated into Salesforce with no changes to current workflows. - **Phase 2 (2 weeks)** — Introduced dynamic completion logic using Salesforce flows and custom rule sets, with CloudFiles' team providing hands-on support to build, test, and fine-tune approval logic based on real case scenarios. - **Phase 3 (4 weeks)** — Rolled out AI-based paragraph grading, training models using benchmark responses and internal guidelines; assessors received brief training to interpret the rating system, leading to immediate productivity gains and measurable ROI within two months. ### Impact & results Integrating CloudFiles' Document AI into the verification workflow achieved faster turnaround, deeper reporting capabilities, and higher data quality. - **50% faster document approvals** — Manual review and approval previously took ~60 minutes per form; after CloudFiles, automation reduced it to ~30 minutes by handling extraction, rule validation, and initial response checks. Across 400 forms per month, that's roughly 200 hours saved out of 400 total monthly review hours. - **25% fully auto-approved forms** — One in four forms cleared all AI-driven checks with no human intervention. - **Enhanced compliance confidence** — Every decision step was logged, traceable, and auditable, ensuring transparency and audit-readiness. - **Improved accuracy & focus** — Reduced human error and ensured all responses were validated consistently, so assessors could focus only on edge cases and high-risk reviews. ## Book a demo Join 1,000+ teams using CloudFiles to streamline workflows, cutting document chaos by 75% while saving hours every week. [Book a demo](https://meet.cloudfiles.io/salesforce?utm_content=sf-case-studies-healthcare-services-firm-footer)