ibl.ai Agentic AI Blog

Insights on building and deploying agentic AI systems. Our blog covers AI agent architectures, LLM infrastructure, MCP servers, enterprise deployment strategies, and real-world implementation guides. Whether you are a developer building AI agents, a CTO evaluating agentic platforms, or a technical leader driving AI adoption, you will find practical guidance here.

Topics We Cover

Featured Research and Reports

We analyze key research from leading institutions and labs including Google DeepMind, Anthropic, OpenAI, Meta AI, McKinsey, and the World Economic Forum. Our content includes detailed analysis of reports on AI agents, foundation models, and enterprise AI strategy.

For Technical Leaders

CTOs, engineering leads, and AI architects turn to our blog for guidance on agent orchestration, model evaluation, infrastructure planning, and building production-ready AI systems. We provide frameworks for responsible AI deployment that balance capability with safety and reliability.

Back to Blog

Samsung Medical Center Is Testing Heidi's AI Scribe On-Premise. What Should a Hospital Learn?

Jaione AmigotNovember 22, 2026
Premium

Samsung Medical Center signed an MOU with Heidi on November 12-13, 2026 to test an on-premise version of Heidi's AI scribe inside the hospital's own infrastructure, linked to its EMR. It is a proof of concept, not a live deployment, and its three test criteria are a usable template for any hospital.

The Short Answer

Samsung Medical Center signed an MOU with Heidi to test an on-premise version of Heidi's AI scribe inside the hospital's own infrastructure, linked to its EMR. It is a proof of concept judged on technical performance, physician acceptance and documentation quality, not a live rollout. The lesson is the deployment shape: with ibl.ai you own all the code and the data, inside your own walls.

Social posts described this as clinical AI "running entirely on hospital GPUs" with zero data leaving the building. The published reports say something narrower, and more useful.

What did Samsung Medical Center and Heidi actually agree in November 2026?

Samsung Medical Center (SMC) and Heidi, a clinical AI company founded in Melbourne in 2019, signed a memorandum of understanding, reported by Korea Biomedical Review on November 13, 2026. SMC President Park Seung-woo and Heidi CEO Thomas Kelly signed.

The MOU has four parts, per the same report and TNGlobal's November 16 coverage:

WorkstreamWhat the reports say
Speech-to-text integrationExplore integrating Heidi's speech-to-text into SMC's DARWIN platform
EMR linkConnect that integration to SMC's electronic medical record system
On-premise proof of conceptHeidi deploys its on-premise version inside SMC's infrastructure
StandardsA Korean clinical documentation standard, a Korean HL7 FHIR implementation guide and a "Heidi Korea" reference architecture

It is Heidi's first partnership with a Korean hospital. If the proof of concept succeeds, SMC could become a reference site for other Korean hospitals.

Is Samsung Medical Center's clinical AI already live on hospital GPUs?

No source we found says so. Every report describes a proof of concept, with no published timeline, clinician count or go-live date.

None of the coverage mentions GPUs or any hardware. The only infrastructure detail is that the on-premise version runs inside SMC's infrastructure, linked to the EMR.

The reports also do not state that no patient data leaves the hospital. That is a reasonable reading of an on-premise deployment, but it is an inference, and a hospital board should hear it as one.

The cost framing that travelled with the story needs the same care. Google Cloud's HLTH-week survey of 250+ healthcare executives found 30% have scaled agentic AI, not 9 in 10. We worked through those numbers in why 40% of healthcare leaders pause AI over unpredictable costs.

Why would a smart hospital test an AI scribe on-premise instead of in the cloud?

Heidi sells both. BioPharma APAC reports that hospitals can choose on-premise or cloud based on their infrastructure and data security needs. SMC chose to test the version that runs inside its own walls.

SMC is not a hospital that lacks cloud options. Korea Biomedical Review notes it holds HIMSS Stage 7 in four areas: IT infrastructure, EMR, digital imaging and data analytics. It ranked 4th globally in Newsweek's World's Best Smart Hospitals 2027.

An AI scribe hears the whole consultation. Heidi says it supports more than 3 million patient consultations per week across 190 countries and 110 languages. Audio of that kind is among the most sensitive data a hospital produces.

When that audio is transcribed inside the hospital, the transcript, the draft note and the EMR write stay under the hospital's own access controls and audit logs.

Is Samsung Medical Center the only Korean hospital keeping clinical AI inside its own network?

No. Seoul's Asan Medical Center announced a private AI knowledge search on May 18, 2026, per Asia Business Daily.

Staff ask natural-language questions and get answers grounded in internal manuals: sepsis protocols, antibiotic guidelines, infectious disease reporting rules.

It runs on a network fully separated from the public internet, uses only hospital servers and data, and was built by Asan's own IT staff.

Asan plans a sandboxed external search for current medical literature, where only anonymized questions that cannot identify a patient would leave the network.

Two of Seoul's largest hospitals, two different workloads, one pattern: the model goes to the data, not the other way round.

How should a hospital judge an on-premise AI scribe proof of concept?

Start with SMC's three criteria. They are a clean acceptance test for any ambient documentation pilot, wherever the model runs.

SMC criterionWhat to measure in your own pilot
Technical performanceTranscription accuracy in your languages and specialties, latency to a draft note, and whether the EMR write succeeds without manual re-entry
Physician acceptanceShare of piloting clinicians still using it after the pilot weeks, and edits needed per note
Documentation qualityClinician or coder review of a sample of notes against your documentation standard

Add two questions SMC's criteria leave implicit. Who holds the transcripts and drafts after the pilot ends? And can you change the speech or language model underneath without rebuilding the EMR integration?

SMC is answering the integration question with standards: a Korean HL7 FHIR implementation guide means the EMR link is written to a published interface, not to one vendor's private one.

What does ibl.ai add for a hospital that wants clinical AI on its own servers?

ibl.ai is an agentic AI operating system a health system deploys on its own infrastructure, with full source code.

You own all the code and the data, and the platform runs model-agnostic across any LLM, so the language or reasoning model can change without the agents changing.

A hospital can use it for the layer around documentation as well as the note itself: agents grounded in internal protocols, like Asan's, and agents that act in clinical systems, which raises the control questions we covered in CVS cleared a backlog with a computer-use agent: who controls it?.

For voice workloads specifically, we mapped where PHI travels on each hop in self-hosted voice AI agents for hospitals.

There is no per-seat pricing, so a pilot that grows from 20 clinicians to 2,000 does not multiply a licence line. ibl.ai is family-owned and operated from New York, NY, and its engineers build alongside the hospital's own team.

Sources: MOU details from Korea Biomedical Review, November 13, 2026 and TNGlobal, November 16, 2026; deployment options from BioPharma APAC; Asan Medical Center from Asia Business Daily, May 18, 2026; survey figures from Healthcare Dive, November 16, 2026. Heidi's consultation, country and language figures are the company's own. No report we found mentions GPUs, a go-live date or a statement that no patient data leaves SMC.

Why does owning the AI stack matter?

ibl.ai is the agentic AI platform where you own all the code and the data. You self-host the entire stack inside your own perimeter, run it model-agnostic across any LLM and switch anytime, and pay by usage with no per-seat pricing β€” so you can deploy anywhere: your cloud, on-premise, GovCloud, or fully air-gapped.

  • You own all the code and the data

    Full source code under a perpetual license, running on your infrastructure. Not API access to someone else's platform β€” the stack itself is yours.

  • Model-agnostic

    Run any LLM β€” Claude, GPT, Gemini, Llama, Command, or your own fine-tune β€” and switch providers without rewriting the platform.

  • No per-seat pricing

    Usage-based billing against a budget cap you set. Cost tracks what your organization actually uses, not how many people you employ.

  • Deploy anywhere

    Your cloud, your VPC, on-premise, GovCloud, or a fully air-gapped network with no outbound connectivity.

1.6M+ users across 400+ organizations run the platform this way, including NVIDIA, MIT, and Syracuse University.

ibl.ai is family-owned and operated from New York, NY β€” a U.S.-headquartered, domestically-owned long-term partner, not a vendor that sells licenses and moves on.

See the ibl.ai AI Operating System in Action

Discover how leading universities and organizations are transforming education with the ibl.ai AI Operating System. Explore real-world implementations from Harvard, MIT, Stanford, and users from 400+ institutions worldwide.

View Case Studies
Work with our team

Pilots, deployment, and full ownership

Every engagement is scoped to the work: you integrate ibl.ai with your own data, deploy it on your own infrastructure, and the engineering hours scale with the scope β€” so the price tracks the scope, not your headcount. Most customers keep us on afterwards for platform updates, new agents and integrations, and ongoing services.

Start here

Pilot

from $15K

fixed scope Β· fixed timeline

A time-boxed proof of value on your real data β€” not a slide deck.

Best for: Teams that want to see ibl.ai working before committing.

  • Deployed on your infrastructure or our cloud
  • 1–2 production agents wired to a slice of your data
  • One integration (LMS / SIS / SSO / data source)
  • Weekly working sessions with our engineers
  • Pilot fee credits toward a full engagement
Scope a pilot
Most common

Integration & Deployment

$25K – $80K

one-time fee Β· scoped to the work

Full deployment integrated with your data and systems. Engineering hours scale with scope.

Best for: Organizations rolling ibl.ai out across a department, campus, or business unit.

  • Platform deployed in your VPC, on-prem, or air-gapped
  • Integrated with your data + identity (SSO / SAML)
  • Multiple custom agents built to your workflows
  • Engineering hours proportional to scope
  • You own the data Β· run any LLM you choose
Plan a deployment
Full ownership

Codebase Transfer + Custom AI Engineering

Custom quote

perpetual license Β· you own the stack

We transfer the full source code. You own and self-host the entire platform β€” outright.

Best for: Organizations and enterprises that benefit from perpetual ownership and sovereignty.

  • Complete source-code transfer + perpetual license
  • Dedicated AI engineering team on your roadmap
  • Custom agents, models, and integrations to spec
  • Air-gapped capable Β· zero vendor lock-in
  • Family-owned, New York–based long-term partner
Talk about ownership
You own the code and data Run any LLM β€” Claude, GPT, Gemini, Llama Family-owned & operated from New York, NY