Case Studies
Eight real engagements. Every number below is measured, not projected. The systems behind them are running in production today.
An energy client's internal reporting group estimated 18 business days to build 8 financial risk management reports. Our AI-assisted development pipeline delivered all 8, including full documentation with annotated walkthroughs and screenshots, plus server deployment, in under 2 hours.
A growing outpatient network acquires clinics faster than legacy records can follow. We extract each acquisition's legacy EMR (a dozen-plus distinct systems so far), standardize it, and load it into a cloud warehouse behind a single records application, so clinicians see the complete patient history no matter where it originated. One recent ledger conversion put 299,509 rows into production over a single weekend. Verification is not a person checking a spreadsheet: an automated nightly audit compares every acquisition's row counts and date ranges against the live application views and pages us on any regression, and a recovery program reattached 97.9% of documents that older migrations had orphaned.
An independent game designer wanted proof his combat math held up before printing anything. We built a simulation harness that plays his game thousands of times with scripted opponents and reports faction win rates with confidence intervals, so a real imbalance is distinguishable from noise. The batteries ranked all 11 factions, identified rules that could never actually trigger in the digital build, and caught a defect that silently hung about 1% of games. He now tests balance changes himself in a browser sandbox that re-simulates his edits in seconds.
A revenue-cycle operations team planned staffing in a large hand-maintained workbook. Working from their own written model, we delivered an interactive console that forecasts call volume in 96 daily intervals, converts it to workload with handle time, applies planned and unplanned shrinkage, packs 9-hour shifts around lunches and breaks, and shows over- and under-coverage at a glance, with Erlang C service-level math built in. A what-if panel projects twelve months of staffing under changed volume and handle-time assumptions. The first working version shipped the day after the spec landed.
Recorded meetings become searchable transcripts with named speakers and structured summaries, with audio processed entirely on local hardware. Transcription fans out across a small distributed compute grid; speaker identification matches voices against enrolled voiceprints and is match-only by design, so it names a voice it knows and stays silent about one it doesn't. The pipeline runs three times a day, unattended, and flags its own failures instead of waiting for a human to notice.
EHR extraction, conversion, and transformation across systems that were never designed to communicate. SQL Server, Snowflake, and custom transformation pipelines bridging clinical data, financial reporting, and operational analytics, built on thirteen years of running this class of problem at CIO scale.
Custom AI pipelines for a major energy credit client, applied to credit identification, processing accuracy, and revenue recovery across the portfolio. Every automated determination passes through human review gates: the AI accelerates the work, the client's experts stay accountable for it.
The discipline we sell is the discipline we run on. Our own delivery infrastructure keeps 32 scheduled jobs under heartbeat watch: every job reports success on an expected cadence, silence is treated as failure, and a 15-minute watchdog sweep pages a phone through a deduplicated alert gateway. On one engagement, an autonomous agent reads the client's feedback channel and ships approved changes to a live product on a 30-minute cycle. Every outbound message is clearly labeled as coming from the bot, and nothing reaches the client without human sign-off.
Client names are withheld throughout. Engagements described here span healthcare, energy, operations, and independent product development; figures are actuals from delivered work.
Everything we deploy, from AI agents to infrastructure.
The tools and environment behind AI-accelerated consulting.
Purpose-built development environment for rapid prototyping and deployment of AI systems, data pipelines, and analytics solutions.
Real-time monitoring, AI agent management, and multi-system development across containerized infrastructure.
Thirteen years as CIO managing technology for 25+ hospital/physician systems with $3B+ in accounts receivable. Every tool, process, and instinct behind Revelliance was forged here.
2014. PFS Group headquarters. Thirteen years as CIO of this company built the expertise that Revelliance was founded on.
2015. New office space under construction. Managing $3B+ in hospital A/R required serious infrastructure.
2015. The finished operations center. 25+ hospital/physician systems, hundreds of workstations, all running.
The team that managed patient financial services technology across 25+ hospital/physician systems nationwide.
Building culture and relationships that carry forward into everything we do today.
On-site work, because some solutions require showing up in person, not just shipping a dashboard.
Book a free AI strategy call and find out what's possible when consulting meets intelligence.
Book an AI Strategy Call