Updated:
Pieces Technologies
Dr. Ruben Amarasingham, a former Parkland Health & Hospital System physician-executive, launched Pieces Technologies in 2015 as a clinical-AI company built...
Pieces Technologies
Dr. Ruben Amarasingham, a former Parkland Health & Hospital System physician-executive, launched Pieces Technologies in 2015 as a clinical-AI company built inside a live hospital environment. The founding thesis was straightforward: most healthcare AI fails because it is trained on static claims data, not the messy, real-time reality of a safety-net hospital. Pieces embedded engineers and clinicians side by side at Parkland in Dallas, annotating real clinical notes to train models that predict patient deterioration, readmission risk, and optimal discharge timing. The company's platform ingests structured EHR data alongside unstructured physician notes, lab results, and social-work assessments. Pieces then surfaces risk scores and recommended interventions inside the clinician's existing Cerner or Epic workflow. Confirmed deployments span major Texas health systems including Parkland Health & Hospital System, UT Southwestern Medical Center, and Children's Health. The firm's predictive focus targets avoidable days, 30-day readmissions, and missed social-service referrals—measures directly tied to value-based reimbursement schedules. Pieces raised venture funding from Concord Health Partners and OSF Ventures. In January 2024, the firm announced an AI-powered utilization-management tool integrated with clinical documentation workflows, aiming to reduce payer-provider friction around inpatient authorization (per Healthcare IT News, January 2024). The company has grown its headcount modestly in the Dallas-Fort Worth area, maintaining a deliberate model of deploying clinical liaisons alongside each health-system rollout. Pieces differs structurally from clinical-AI peers because its foundational training data was built inside a safety-net institution rather than retrofitted from academic medical centers. That gives the models a built-in understanding of patients with unstable housing, food insecurity, and fragmented outpatient follow-up—precisely the populations where readmission penalties hit hardest. The firm does not sell a general-purpose large-language model; it sells condition-specific predictive modules with clear operational owners inside the hospital.
General information
Firm type
other
Year founded
2015
Location
Region
North America
Country
United States
City
Irving
Corporate office
Irving, TX, United States
Principals
Ruben Amarasingham
CEO & Founder
Sector focus
Frequently asked questions
Who runs product and clinical strategy at Pieces Technologies?
Dr. Ruben Amarasingham serves as CEO and Founder. He joined Pieces from an unusual perch—he was previously a practicing physician and executive at Parkland Health & Hospital System, where he also founded the Parkland Center for Clinical Innovation. This dual background means the product roadmap reflects operational pain points inside a safety-net hospital rather than theoretical use cases identified in academic AI labs.
How does Pieces source its training data, and why does that matter?
Pieces models are trained on clinically annotated inpatient data drawn from Parkland Health & Hospital System, one of the nation's busiest safety-net hospitals. The firm deployed embedded clinical data scientists who annotated real physician notes, lab values, and social-work assessments. That process created a dataset representing high-acuity, socially complex patients—the population where predictive models have historically been least accurate.
What does Pieces actually do inside a hospital's EHR?
Pieces integrates with Cerner and Epic to surface predictive risk scores within the clinical workflow. The platform generates flags for discharge readiness, 30-day readmission risk, and social needs that might delay safe discharge. Clinicians see these as structured alerts inside their existing charting environment, rather than needing a separate application. The firm's utilization-management module, announced in 2024, extends this logic to payer authorization by aligning clinical documentation with the evidence plans require.
Is Pieces a healthcare AI platform or a clinical-services vendor?
Pieces is a clinical-AI platform company structured to deliver predictive models as an embedded software layer. Deployment requires a health-system partnership that includes data sharing, clinical liaison staffing, and workflow mapping. It is not a pure SaaS vendor selling an out-of-the-box model; each deployment involves initial annotation work and integration with local clinical workflows.
Which health systems use Pieces today?
Publicly confirmed clients include Parkland Health & Hospital System, UT Southwestern Medical Center, and Children's Health—all in the Dallas-Fort Worth area. The firm's scaling strategy has prioritized deep local partnerships and multi-year co-development agreements rather than broad but shallow national rollouts. No public disclosures detail deployments beyond Texas.
How does Pieces handle the shift from fee-for-service to value-based reimbursement?
The firm's predictions align with measures that carry financial penalties or shared-savings upside: avoidable days (length-of-stay), 30-day readmission rates, and missed social-determinant referrals. By focusing on these specific operational metrics, Pieces builds the business case for health systems managing at-risk contracts or facing CMS readmission penalties. The utilization-management module targets payer denials, which have risen sharply as prior-authorization requirements expand.
Who invested in Pieces, and what does the funding structure look like?
Pieces raised venture funding from Concord Health Partners, a healthcare-focused growth equity firm, and OSF Ventures, the corporate venture arm of OSF HealthCare. The firm remains privately held, does not disclose revenue or valuation, and has not announced a Series C round or secondary transaction as of mid-2025. The investor profile suggests alignment with strategic health-system partners rather than generalist tech VCs.
Profile maintained by Altss using OSINT (open-source intelligence), regulatory filings, licensed data partners, and verified direct submissions. Read the methodology. Last updated: . Continuous refresh with full update cycles at least every 30 days.
Need institutional-grade insight on investors?
Altss delivers:
Prefer a guided tour?
We’ll walk you through: