ReleasePoint and the Art of Reading Between the Records

Consider a medical chart that contains the words “lung cancer.” A machine can find them. An insurance underwriter needs to know whose cancer it is. The patient’s? A parent’s? A diagnosis from years ago? ReleasePoint uses this example to explain a surprisingly durable problem: collecting medical information and making it useful are two different jobs.

That distinction gives this company its shape. ReleasePoint works in the space between healthcare providers who hold records and organizations that need to evaluate them. Its customers are trying to underwrite a policy, examine a disability claim, prepare litigation or study clinical information. Before any of that can happen, someone must assemble the evidence.

The expensive part was getting it right

In early 2013, the problem was painfully ordinary. According to a historical MicroSourcing case study, ReleasePoint had long-term contracts and rising service costs. Asking clients for higher prices risked reopening the work to competitive bids. Reliable revenue had become an uncomfortable constraint. The company needed to change how it produced its service.

It evaluated overseas providers, selected MicroSourcing and launched a 34-person Philippines operation in November 2013. The partner’s account says production cost per unit subsequently fell by more than 40%. ReleasePoint retained operational control rather than accepting a vendor’s standard operating framework. The interesting bargain was cheaper production with room to manage quality.

The company’s present-day offering still reflects that practical concern with the work underneath the software. Its expertise crosses retrieval operations, quality control, clinical data handling and integration. Someone has to notice the wrong document, pursue a missing page and keep a request from disappearing into a queue.

The chart arrives. The meaning does not.

MIDAS is ReleasePoint’s answer to the second job. It takes image-based medical records and moves them through optical character recognition, human curation, further analysis and structured delivery. Trained specialists validate and annotate extracted information. The result can feed another system through an API, giving downstream software something more usable than a picture of a page.

ReleasePoint calls the output “decision-grade data.” That is a product promise rather than an independent accuracy score. Its design recognizes that a correctly detected term may still be misunderstood. A date needs an event attached to it. A condition needs a patient and a context. Extracting the words leaves plenty of work unfinished.

A PDF is a product decision

The delivery options reveal another useful detail. Customers can use a web portal or a PDF abstract as well as structured data. A carrier may want better information without rebuilding its internal systems. A legal reviewer may prefer a document that works in familiar software. ReleasePoint offers several ways to consume the result.

DataLens addresses a narrower task: finding relevant passages. Teams create dictionaries of keywords and topics, then review matches with surrounding text. Clicking a match takes the reviewer to its location in the medical record image. Customers can provide records they already possess, so the review tool can be used separately from retrieval.

RP Insights adds AI-assisted summaries. Together, these offerings make ReleasePoint a supplier of managed information services with software attached to the work. Buyers engage its sales team to configure services and pricing. The commercial question is how much retrieval, review and integration a particular organization needs, rather than which consumer subscription to choose.

The checkbox that fooled the summary

In August 2026, ReleasePoint published an unusually useful example of a limitation. A medical form consisted largely of checkboxes and X-marks. AI struggled to distinguish checked items from crossed-out ones, and the summary understated the patient’s concerning presentation. The client could follow the summary back to the original document and resolve the discrepancy.

The first weakness in that example was interpreting the source form. A polished summary could not repair a mistaken reading underneath it. ReleasePoint also identifies poor scans, handwriting, conflicting documentation and ambiguous language as difficulties. For a buyer, the practical test is whether a finding can be checked against its evidence.

“We live and breathe cycle time.”Cheryl Zatopek · operations interview, February 2026

The lesson others can copy is specific: keep a route from an extracted assertion to the original document, and give someone responsibility for checking consequential ambiguity. This approach depends on usable records, appropriate access and reviewers who understand the decision. A summary alone cannot supply missing evidence or settle conflicting clinical accounts.

Keep the request moving

ReleasePoint applies similar thinking before the reading begins. Its EHR service can pivot to traditional retrieval when electronic records are unavailable or fail quality requirements. RPNet, its client portal, exposes request status, authorization problems and provider responses. The fallback and the status note address different ways a case can stall.

Partnerships place that work inside other people’s systems. MIB announced an agreement in 2021 to add ReleasePoint’s traditional retrieval to its medical-data workflow. Appian’s June 2024 announcement described requesting records and receiving reports within its underwriting interface; it described MIDAS integration as future work. Fewer handoffs are part of the value.

ReleasePoint shares the insurance retrieval market with alternatives such as ExamOne and Parameds. Its case rests on combining collection, contextual review and flexible delivery. The revealing measure is what happens when the tidy route fails: an empty electronic search, a confusing form, a missing authorization. Those are small moments. They determine whether anyone can make the next decision.