Cherish Grants cut their intake process from 11 days to under an hour.

Cherish Grants received a high volume of applications, each requiring a human to open it, read it, pull out the relevant data, assess basic eligibility, and route it to the right reviewer. For a small, mission-driven team, this intake work was eating into the time they should have been spending on actual grant decisions. Processing times stretched close to 11 days per application, a bottleneck that frustrated applicants and limited how many grants Cherish could realistically manage.

The goal wasn't to remove humans from the process, it was to remove them from the parts that didn't need them. We mapped the intake workflow and identified that the vast majority of the work was structured and rule-based: read the application, extract defined fields, check against eligibility criteria, route to the right person. Only the edge cases genuinely needed human judgment. We designed a pipeline that handles the standard path end-to-end and surfaces only the exceptions for the team to review.

An intelligent intake pipeline triggered on submission. It reads each incoming application, extracts structured data from unstructured content, scores eligibility against Cherish's defined criteria, flags edge cases with a confidence indicator, and routes each application to the appropriate reviewer, with a pre-built summary attached so reviewers arrive at the decision point, not the data extraction point. Every step is logged, timestamped, and auditable. The pipeline handles multiple application formats and integrates with Cherish's existing systems without requiring them to change how applicants submit.

Processing time dropped from nearly 11 days to under an hour per application. Standard cases move through the pipeline without any manual handling. The Cherish team now spends their time on the decisions that require them, not the data entry that doesn't. Capacity increased significantly without adding headcount, and the full audit trail means nothing falls through the cracks.