THE LATEST
● KEYMARK ACQUIRES ALL STAR SOFTWARE SYSTEMS · APRIL 2026● DOCUMENTS → DATA → DECISIONS

Enterprise / Company profileField notes № 001

KeyMark knows where the paperwork is buried

Before an AI system can answer a question, somebody has to make sense of the files. KeyMark has spent three decades doing that work - and its most revealing projects start with a very human mess.

Advertisement

The files had already been scanned. For 13 years, a county government had worked through the patient business of turning paper records into digital ones. Then a cyberattack corrupted the data on its server. Employees went back to hunting through physical files. The supposedly finished job was waiting for them again.

KeyMark’s account of the recovery describes a different route through the backlog. The county handed over its files; KeyMark used its CloudCapture service to separate, classify, verify and index them, then put the information into Hyland’s OnBase system. The reported result: 1.67 million document records processed in six months.

That is a useful place to meet KeyMark. Much of the technology industry sells the future. This company is frequently called into the room where the past is stacked in boxes. Its work concerns the information an organization already possesses but cannot reliably use.

THE RECOVERY JOB1.67M

document records processed in six months

Company-reported outcome for an unnamed county government.

The file room gets an AI budget

Founded in 1996 and headquartered in Liberty, South Carolina, KeyMark describes its business as enterprise content intelligence. Translated into office English, that means getting information out of documents, checking it, organizing it and delivering it to the person or system that needs it. Customers include agencies and businesses in government, insurance, healthcare, finance, manufacturing and education.

Its current pitch puts those old problems inside a newer one: enterprise AI. A model cannot make much use of an inaccessible scan or an invoice filed under the wrong account. KeyMark’s proposed sequence is to locate the content, make it usable, apply governance and then feed it into AI or operational workflows.

The distinction matters. Buying an AI platform does not settle who may read a file, whether its extracted numbers are correct or how long it should remain available. Those decisions live inside the business. KeyMark sells the work of connecting the software to them.

Its Vertesia offering extends that approach into AI applications and agents, including document enrichment and coordinated workflows. The attraction is continuity: information that has been captured and organized for employees can also become useful input for machines. The hard part remains making that information trustworthy.

FROM DOCUMENT TO DECISION
  1. 01FindLocate the files
  2. 02ReadExtract and classify
  3. 03CheckValidate and govern
  4. 04UseRoute to people + AI
A scanned page still has a journey ahead of it. This is the work between storing information and using it.

A business built between the bricks

Founder Jim Wanner began with automated data entry, including reading machine print and handwriting. Early business included Blue Cross Blue Shield of South Carolina. The opportunity widened as the company moved into document management and business processes, working with partners including Watermark and Hyland.

“We’re the mortar connecting the bricks.”

Cameron Boland, in Upstate Business Journal

That description is unusually helpful for a value-added reseller. The bricks are products that already exist. The commercial proposition is knowing how to fit them together: choose the tools, configure the workflow, connect the systems and help people operate the result.

KeyMark’s market position follows from that work. A buyer can purchase from a software manufacturer, hire another specialist integrator or ask its own IT team to build the connections. KeyMark’s argument is that document expertise and continuing services justify bringing in an intermediary. Its familiarity with Hyland and Tungsten Automation gives it a defined practice, rather than a claim to solve every technology problem.

Five KeyMark executives pose for a 2018 team portrait, with founder Jim Wanner in the center.
Five people, zero filing cabinets. KeyMark’s 2018 leadership portrait puts faces to the work behind the workflow. Photograph: Will Crooks / Upstate Business Journal.

Start with the invoice nobody wants to download

At Dot Foods, the opening experiment was agreeably small. In 2020, the foodservice redistribution business examined a recurring chore: employees visiting a supplier website to download invoices. Automating that task offered a test that could be repeated elsewhere if it worked.

Dot and KeyMark wrote a functional specification before proceeding. KeyMark then implemented Blue Prism development and production environments and helped Dot’s developers learn the platform. The first digital worker began downloading invoices; Dot subsequently built seven additional processes for other supplier sites.

Success changed the staffing conversation, too. Dot hired another developer and formed a committee to identify further back-office automation opportunities. In the case study, the company projected savings equivalent to 2.5 full-time roles by the end of 2021, following its start in late 2020. That was a forecast, not a reported layoff count.

The sequence is worth copying: define a repeated task precisely, equip the people who will maintain it, prove it and expand to similar work. A working invoice download is a more useful internal sales pitch than a grand announcement about transformation.

The computer is allowed to be unsure

SC State Housing offers a closer look at what happens inside document automation. Its system classified incoming material, extracted a loan number and checked that number against a database before indexing the documents in OnBase. Material the software could not identify confidently went to an operator for review.

That last step is easy to overlook in a product demonstration. A doubtful document still needs an owner. Keeping a review queue makes the uncertainty visible instead of passing a questionable value into the next system.

In the agency’s case study, IT director Mark Phipps reported that preparation once taking hours could be completed in roughly 30 to 40 minutes. He also reported that overall loan processing time had been halved. These are two different measures: document preparation is one stage of a larger process.

The agency reassigned two employees, while another retired, reducing the loan origination department by three full-time staff. The benefit described was redistribution of capacity as well as speed. Better access to records also meant employees could retrieve loan information while speaking with a customer.

SC STATE HOUSING · REPORTED RESULT
Previous loan processing time 100
After implementation 50

Relative index: previous processing time = 100. The case study reports a reduction by half.

What arrives in the box - and on the bill

KeyMark’s offering spans several kinds of purchase. CloudCapture is managed processing: a combination of AI and human validation for work such as invoices, mailroom documents and data conversion. It lets a customer hand over a processing function, including temporary volume surges, instead of operating every part internally.

Forms InMotion tackles paperwork before it becomes a capture problem. The software works with existing PDFs, collects signatures and attachments, routes forms to participants and copies shared information across fields in a form package. Anyone who has entered the same address repeatedly can appreciate the ambition without a diagram.

FileHold adds owned document and records management software, covering storage, retrieval and the record lifecycle. Alongside those products sit partner platforms, implementation, integration, hosting, administration and support. The business therefore combines software with the services required to keep it useful.

Cost should be assessed across that whole engagement. KeyMark invites buyers to request a quote for outsourced invoice processing. For a buyer comparing proposals, the sensible calculation includes software, configuration, connections to existing systems, migration, training and continuing operation. A license price alone cannot describe the work.

There is also the cost of leaving a process as it is. KeyMark’s account of AnMed Medical Center describes more than 6,000 invoices manually processed each month before a customized capture system improved indexing and retrieval. Time spent locating an invoice belongs in the calculation alongside time spent entering it.

The economic test is practical: will the removed work and reduced rework justify the implementation and running costs? Time released is capacity; it becomes a financial saving only when an organization can use or reduce that capacity. Buyers should keep those two outcomes separate.

New owners of software, new leadership

KeyMark founder Jim Wanner in a 2018 portrait
The founder steps back from the CEO desk. Jim Wanner, pictured in 2018, became chairman in June 2025. Photograph: Will Crooks / Upstate Business Journal.

KeyMark has been changing its own business while helping customers change theirs. Its acquisition of FileHold took effect on December 31, 2024. Wanner’s explanation was specific: the purchase filled an offering gap and opened a previously unserved market segment.

A strategic investment from Source Capital followed in 2025. On June 26 that year, Wanner moved from CEO to chairman, with a continuing focus on acquisitions and expansion. Ed McQuiston, a longtime Hyland executive, took over as chief executive.

In April 2026, KeyMark acquired All Star Software Systems, another Hyland and Tungsten specialist. The stated plan preserved familiar customer teams while adding the larger organization’s support and delivery resources. Taken together, the moves suggest a company assembling more software ownership and implementation capacity around its established partner relationships.

There is a less corporate side to the place. An employee account describes a lunchtime Nintendo Switch habit that grew into the KeyMark Gaming Society, with a tournament, logo and shirts. Other groups cooked together or played Dungeons & Dragons after hours. McQuiston’s biography adds a different extracurricular pursuit: DJing.

Bring your awkward documents

The strongest buying conversation would begin with troublesome examples. KeyMark’s discussion of explanation-of-benefits documents describes why older capture methods struggled: inconsistent insurer layouts, complicated information and multi-line structures. A tidy sample cannot demonstrate how a system will handle that variety.

For a prospective customer, the implication is to test exceptions, measure correction work and decide who handles an ambiguous result. Low volumes, constantly changing processes or work that depends heavily on individual judgment may offer less scope for an elaborate automation project. The process needs to justify the machinery around it.

KeyMark’s appeal rests on an observable problem: organizations have information they struggle to put to work. Its useful promise is that an invoice can be found, a loan file can reach an officer sooner and a supplier download can stop occupying someone’s morning. Enterprise AI gives that work a new audience. The paperwork was already there.

Keep following the paper trail