Notable and Epic: APIs, RPA and What Bi-Directional Really Means

In layers, and the layer depends on the workflow. For patient facing intake, registration and scheduling, Notable typically reads the schedule and demographics through Epic's interfaces, runs the patient interaction on its own text, web or voice channel, and writes results back through APIs or HL7 v2 where Epic exposes the right write path. For workflows where Epic does not expose a write path, or where the site has not approved one, Notable's assistants log into Hyperspace with a service account and enter the data through the screens a person would use.

The API and HL7 paths are the ones your integration team already understands, and the questions in our integration questions guide apply directly. The RPA path is different. It requires a bot user with Epic security points, it is sensitive to screen and template changes at each Epic upgrade, and its audit trail shows the bot user, not the human who approved the action. Those are manageable, and many health systems run RPA on Epic today, but they need an owner and a policy.

Not published: which method Notable uses for each of your target workflows on your Epic release. Ask for a workflow by workflow integration map before contract, with the method, the interfaces, the security points and the upgrade sensitivity for each.

The point of Notable is that staff should see less, not more. Intake answers arrive as completed registration and history data. Insurance is verified before the visit. Referrals are worked and status updated. Care gaps are identified and outreach sent. Documentation content is populated for clinicians to review. Voice agents answer inbound calls and book, cancel and confirm appointments. Catholic Health reports 64 percent containment on voice calls and a 360,000 USD annual saving, per the vendor's blog; those are customer stated figures from one site.

For the buyer, the interesting question is which of these writes are structured. A registration field is structured; the wrong value in it propagates to claims. Note content is unstructured; a clinician reads it before signing. The patient intake agent use case and the patient scheduling use case set out what to measure for each, and our agent versus copilot versus automation explainer is useful for deciding which workflows should act without a human in the loop and which should not.

• The Showroom vendor client and each API or HL7 interface, raised through your integration request process, per workflow.
• For any RPA workflow, a bot user with defined security points, a credential management policy, and an owner who tests it after every Epic upgrade.
• Security review and the business associate agreement, covering data retention, subprocessors and any model training. The HIPAA and AI page lists what to insist on.
• Workqueue and template design, so staff see Notable's results where they already work.
• Patient communication consent and TCPA review for text and voice outreach.
• A test plan per workflow, and a change control slot for each go live.

Because Notable spans many workflows, the temptation is to approve a platform once and add workflows quietly. Resist it. Each workflow is a separate interface, a separate risk and a separate governance decision. Our governance and compliance service designs the intake process for that.

The first is upgrade fragility on RPA workflows. When Epic changes a screen, a bot that types into it can break silently or, worse, type into the wrong field. Notable manages this at scale and its customers report smooth experiences, but you should know which of your workflows are exposed and who watches them after each upgrade.

The second is scope creep. A platform that can automate almost any administrative workflow invites a long backlog, and the ROI case for workflow twelve is rarely as good as for workflow one. Our agent pilot metrics guide covers how to keep the measurement honest.

The third is price transparency. Notable prices on request, typically as a platform fee plus per workflow or per volume components, and publishes nothing. Our Notable pricing page sets out what buyers report. For the voice and intake pieces, the patient intake software comparison lists the alternatives.

• For each workflow in scope, which method is used on our Epic release: API, HL7 v2 or RPA, and which interfaces and security points does it need?
• For RPA workflows, how are bot credentials managed, what happens after an Epic upgrade, and how quickly has a breakage been detected and fixed at a comparable site?
• Which writes are structured and which are reviewed by a person before they land in Epic?
• Where is patient data processed and stored, for how long, and is any of it used to train shared models? Put it in the BAA.
• What is the platform fee, the per workflow or per volume price, the minimum term and the exit terms?
• Which reference customers run our Epic release, and can we speak to their integration lead without the vendor present?

Our vendor HIPAA question list covers the compliance side in more depth.

Notable is a capable platform with real Epic customers, and the decision is rarely whether it works but which workflows to give it and by which method. An independent review builds the workflow by workflow integration map with your Epic team, separates the API workflows from the RPA ones, prices the result against point solutions for intake, scheduling and voice, and sets up the governance so that workflow twelve gets the same scrutiny as workflow one. We take no vendor commissions.

Our vendor selection service covers the platform decision and our deployment roadmap sequences the workflows. Book a call with your Epic release and your top five administrative pain points.