The healthcare context layer for AI assistants and agents
Connect Dmand AI and your assistant answers from one resolved map of US healthcare: every health system and IDN, hospital, medical group and site of care, with 7.4M+ HCPs and 11M+ executives placed where they work, buying signals, and verified emails and direct dials returned in the chat.
- Live in Claude
- 30 days free
- No credit card
Illustration: a question typed into an AI assistant falls as light onto a horizon of US healthcare. Where it lands, Regional Health resolves as an IDN with its hospitals, medical group and outpatient sites, and the Chief Information Officer at Metro Hospital appears with a verified email and a direct dial.
Customers
Teams selling into US healthcare use Dmand AI
- G2 review (opens in a new tab)
Circle Health · G2 review
I like Dmand AI’s extremely curated and comprehensive healthcare database, which you won’t find on any other platform with this level of detail.
Anupam D.Founder & CEO, StratAnalyze (for Circle Health) - G2 review (opens in a new tab)
NextDimension AI · G2 review
It’s like having a hyper-personalized mail writer that’s unbelievably efficient.
Simran K.Co-founder & CEO, NextDimension AI
Why a context layer
Without healthcare context, an AI assistant answers from press releases and switchboards
Public pages name a CIO from two years ago and stop at the hospital door. The context layer places every site under its health system and every person at the site where they work today.
Assistant alone
Public web pagesWho runs IT at each Regional Health hospital?
I found a 2024 press release naming Regional Health's CIO. I couldn't find a list of its hospitals or the IT leader at each one. You could try the main switchboard.
- Press release
- Hospital homepage
- Main switchboard
- No hospital list
- No site
- No verified contact
With Dmand AI
Resolved recordsWho runs IT at each Regional Health hospital?
- Regional Health, IDN
- Metro Hospital, Hospital
- Chief Information Officer
- Email verified
- Direct dial verified
- North Hospital, Hospital
- Chief Information Officer
- Email verified
- Direct dial verified
- Metro Hospital, Hospital
- Every hospital
- The leader at each site
- Verified contact
Assistant alone
Public web pagesWhich surgery centers does Regional Health own?
Regional Health's website lists two surgery centers. I couldn't confirm which others it owns or runs through joint ventures. A state license search might help.
- Hospital homepage
- News article
- State license search
- No ownership tree
- No joint ventures
- No verified contact
With Dmand AI
Resolved recordsWhich surgery centers does Regional Health own?
- Regional Health, IDN
- Metro Surgery Center, ASC
- ASC Administrator
- Email verified
- Direct dial verified
- North Surgery Center, ASC
- ASC Administrator
- Email verified
- Direct dial verified
- Metro Surgery Center, ASC
- Every surgery center
- Its parent organization
- Verified contact
Assistant alone
Public web pagesHow do I reach the Chief Nursing Officer at Metro Hospital?
Metro Hospital's leadership page lists a Chief Nursing Officer, but no email or direct line. The main number goes to the operator.
- Leadership page
- Professional profile
- Main switchboard
- No direct dial
- No verified email
- Title may be out of date
With Dmand AI
Resolved recordsHow do I reach the Chief Nursing Officer at Metro Hospital?
- Regional Health, IDN
- Metro Hospital, Hospital
- Chief Nursing Officer
- Email verified
- Direct dial verified
- Metro Hospital, Hospital
- Current title
- The site where they work
- Verified direct dial
Inside the context layer
The context layer stacks healthcare data, org context and verified contacts in three layers
Each layer adds what the one below cannot. The data says who exists, the context says how they connect, and the reach says how to contact them.
Illustration: three stacked layers lit by a beam of light. The data layer is a field of record columns. The context layer is an IDN wired to its hospitals and sites, with people tied to the sites where they work and a buying signal on one hospital. The reach layer is a set of people pins, one with a verified email and a direct dial. The layer chosen in the list is the one lit.
01 · Data layer
7.4M+ HCPs by NPI, 11M+ healthcare executives and 997K+ sites of care, from integrated delivery networks down to 235K+ solo practices.
Browse the healthcare database02 · Context layer
Every site sits under its parent organization and IDN, every physician is tied to the groups and sites where they practice, and 80+ buying signals say when to reach out.
How the org hierarchy is built03 · Reach layer
Contacts come from healthcare-specific waterfall enrichment and are verified before they reach the chat. Save the people to a list and start outreach in the same login.
How contact enrichment works
Where the data comes from
Dmand AI combines NPPES, CMS and PECOS records with its own healthcare data into one profile per HCP, site and health system
Public registries disagree on names, addresses and affiliations. Dmand AI matches them into one record, then adds what no registry holds: the org hierarchy, the people at each site, buying signals and verified contacts.
- People linked by org tree, not by email domain
- Site-level affiliations, including secondary practice locations
- Validated emails and direct dials from healthcare-specific sources
Public records
- NPPES NPI Registry
- PECOS enrollment
- CMS Provider of Services
- CMS Doctors and Clinicians
- Medicare cost reports
- CMS Open Payments
- Medicare Part D prescribers
- Medicare claims
- IRS Form 990 filings
- PubMed and ClinicalTrials.gov
Dmand AI
- Org hierarchy model
- Site-level affiliation mapping
- Healthcare waterfall enrichment
- Email and direct dial validation
- Relationship graph
- 80+ buying signals
- AI web research
Illustration: ten public records and seven Dmand AI sources run as lines of light into one point, which resolves into a single record for an illustrative cardiologist: the site of care where they practice, their medical group, hospital and health system, a buying signal, and a verified email and direct dial.
7.4M+
HCPs
One profile per NPI, with practice locations, affiliations and contacts.
997K+
Sites of care
Unique locations across nine segments, from hospitals to pharmacies.
181K+
Medical groups
Single-site and multi-site groups, tied to the physicians who practice there.
124K+
Parent organizations
With 11M+ affiliated contacts, from health systems to practice groups.
22K+
Hospital outpatient units
Outpatient departments, clinics and inpatient units modeled under 6.2K+ hospitals.
How it works
Dmand AI works inside Claude in four steps: find the accounts, map them, reach the buyers and start outreach
One conversation takes you from a region to a first email. Each step picks up the answer before it, so nothing is copied between tools.
Illustration of Dmand AI inside Claude, step 1: asked which Ohio health systems opened new outpatient sites this quarter, Claude searches Dmand AI, checks buying signals and lists Regional Health, Valley Health and Northern Health with the new site at each, then offers to map how Regional Health is organized. All names are placeholders.
01
Find the accounts worth calling
Ask for health systems, hospitals or groups by region, type or buying signal. You get the accounts and the reason to call now.
02
Map each account from IDN to site of care
Hospitals, medical groups and outpatient sites in one tree, with the HCPs and executives who work at each one.
03
Reach the people who decide
The buying committee at the right site, with verified work emails and direct dials.
04
Start outreach from the answer
Save the people to a Dmand AI list, draft the first email in Claude and run the sequence in the same login.
Who it’s for
Teams that sell to health systems, medical groups and physician groups each ask it a different question
One illustrative question per team, with placeholder accounts. The answer comes back resolved, placed and reachable.
Health IT and services
Buying committees by hospital, with the EHR in place and the signal that says when to call.
See Dmand AI for Health ITIllustration: asked which Regional Health hospitals still run an on-premise EHR and who leads IT at each one, Dmand AI returns the Chief Information Officer at Metro Hospital, the Chief Medical Information Officer at North Hospital and the IT Director at Community Hospital, each with a verified email or direct dial, and flags a new IT leader at Metro Hospital.
Healthtech and digital health
Groups sized by specialty headcount, with the person who signs.
See Dmand AI for digital healthIllustration: asked for Ohio physician groups with 20 or more cardiologists and the practice administrator at each, Dmand AI returns the practice administrators at Valley Heart Group and Lakeside Cardiology and the Director of Operations at Metro Physicians, each with a verified email or direct dial, and flags that Valley Heart Group is opening a new site.
Medical staffing and recruiting
Hiring signals by facility, with the leaders to call and candidates’ contacts.
See Dmand AI for medical staffingIllustration: asked which Arizona hospitals opened nurse leader roles this month and who runs nursing there, Dmand AI returns the Chief Nursing Officer and a Nurse Recruiter at Desert Regional Hospital and the Director of Nursing at Canyon Valley Hospital, each with a verified email or direct dial, and flags an ICU nurse manager role at Desert Regional Hospital.
Illustration: asked for a one-page brief on Regional Health, Dmand AI returns its leaders, including the Chief Executive Officer, a new Chief Operating Officer at Metro Hospital and the VP of Strategy, each with a verified email or direct dial, and flags the opening of Metro Imaging Center.
MedTech and medical supplies
Procedure volume by site, with the people above it who approve the purchase.
See Dmand AI for MedTechIllustration: asked to rank Florida surgery centers by orthopedic volume and name the supply chain lead at each parent, Dmand AI returns supply chain leaders at Coastal Health and Bayside Health and the Materials Manager at Gulf Orthopedic Center, each with a verified email or direct dial, and flags a new operating room at Coastal Surgery Center.
Pharma, biotech and CROs
Investigators, trial sites and medical affairs, one record each.
See Dmand AI for pharma and biotechIllustration: asked for oncology investigators with active trials at Northeast academic medical centers and their medical affairs contacts, Dmand AI returns a Principal Investigator and the Medical Affairs Director at Northeast Cancer Institute and a Sub-Investigator at Northeast University Hospital, each with a verified email or direct dial, and flags a new oncology trial site.
I like Dmand AI’s extremely curated and comprehensive healthcare database, which you won’t find on any other platform with this level of detail. The ability to filter by specialty, location, independent physician or not, and many others is incredibly useful.
Where it runs
Dmand AI is a healthcare MCP server for Claude, with more AI agents in early access
Ask in Claude, work the list in the app or call the API. The answer is the same because the context layer is the same, and one account and one credit balance cover all three.Add Dmand AI to ClaudeRead the API docs
Illustration: in Claude with Dmand AI connected, the question "Who leads cardiology at Valley Health's hospitals?" returns the Chief of Cardiology at Valley Hospital, the Cardiology Service Line Director at North Valley Hospital and the Cardiology Practice Manager at Valley Heart Clinic, each with a verified email or direct dial.
In Claude, ask in plain English
Illustration: the same three cardiology leaders saved to the list "Valley Health cardiology" in the Dmand AI app, each with a verified contact, ready to start outreach.
In the Dmand AI app, work the list
Illustration: a REST API response for Valley Health that lists the same three people with their title, site and verified contact.
Through the REST API, build on it
Early access
The context layer is opening to ChatGPT, Gemini, Claude Code, Cursor, Codex and your own AI agents
Claude already runs on it. The next assistants open in batches, most requested first. Pick the ones your team uses and we’ll email you the day each one opens.
Illustration: the data, context and reach layers under an assistant layer. A beam of light rises from the data into the assistant layer, where Claude is live at the center and ChatGPT, Gemini, Claude Code, Cursor, Codex and your own agent wait in their sockets. Each assistant you pick lights up and links to Claude.
FAQ
Questions about the healthcare context layer
Can’t find your answer? Email hello@dmand.ai or book a demo with our team.
What is a context layer for healthcare AI?
A context layer is the verified data an AI assistant or agent reads before it answers, so it works from records instead of guessing from web pages. For healthcare it means every US health system, hospital, medical group, site of care, HCP and executive resolved into one org hierarchy, with affiliations, buying signals and verified contacts. Dmand AI serves it to people in its app and to assistants and agents over MCP and a REST API.
Is a context layer the same as a semantic layer?
They are related. A semantic layer defines what the fields in your own warehouse mean so tools query them the same way. A context layer brings outside knowledge your systems do not hold, here the US healthcare market resolved into records and relationships, and serves it to AI assistants and agents in a form they can reason over.
Does Dmand AI return verified emails and phone numbers inside Claude?
Yes. Ask Claude for the contacts and the Dmand AI connector returns verified work emails and direct dials in the chat, from the same healthcare waterfall enrichment the app uses. Searching is free and a credit is used only when a contact is found.
Which AI assistants and agents can use Dmand AI today?
Claude on the web and desktop through the Dmand AI MCP connector, plus the Dmand AI app and the REST API. ChatGPT, Gemini, Claude Code, Cursor, Codex and custom agents are in early access on the same account.
What is a healthcare MCP server?
An MCP server exposes a data source to AI assistants through the Model Context Protocol, so the assistant can search it and act on the results. Dmand AI is a healthcare MCP server: it gives an assistant search over US health systems, hospitals, medical groups, sites of care, HCPs and executives, the org hierarchy around each, and verified contacts.
What healthcare data does the context layer include?
1.6K+ health systems and integrated delivery networks (IDNs), 6.2K+ hospitals, 181K+ medical groups and physician groups, 124K+ parent organizations, 997K+ sites of care, 235K+ solo practices, 7.4M+ HCPs and 11M+ healthcare executives, each placed in a parent-child org hierarchy with physician-to-site, physician-to-group and executive-to-organization affiliations, plus 80+ buying signals and verified contact coverage.
Where does Dmand AI’s healthcare data come from?
Public registries and filings, including the NPPES NPI Registry, CMS Provider of Services, Doctors and Clinicians and cost reports, PECOS enrollment, CMS Open Payments, Medicare Part D and claims data and IRS Form 990 filings, research sources such as PubMed and ClinicalTrials.gov, and ongoing web research. Dmand AI resolves them into one record per person and organization. Contacts come from healthcare-specific enrichment sources and are verified before they are returned.
How is this different from searching NPPES or an NPI database directly?
NPPES returns one provider record per NPI as it was registered, often with an old address and no employer. Dmand AI starts from those records and adds where the physician practices today, the medical group, hospital and IDN above that site, the executives around them and a verified way to reach them, all searchable in plain English from your assistant.
Can my own AI agent use Dmand AI?
In early access, yes. Any MCP-compatible client or healthcare AI agent connects to the same Dmand AI account and credit balance that powers the Claude connector. Join the list and name the client you use.
What does the context layer cost?
The trial is 30 days free, self-serve, with no credit card. Paid plans start at $249 per user per month and include credits; searching from any surface is free and a credit is used only when a verified contact is returned.
The big picture
All of US healthcare, in context.
Every IDN, hospital, medical group and site of care in the country, with the people who work there, inside the assistant your team already opens.
- 30 days free
- Self-serve
- No credit card
Illustration: a dot map of the contiguous United States lies under one cone of light. Health systems glow with rings and rising light, arcs join each one to its nearest hospitals, and sites of care brighten across the country.
1.6K+
Health systems and IDNs
6.2K+
Hospitals
997K+
Sites of care
7.4M+
HCPs
11M+
Healthcare executives