It’s the start of a Tuesday day shift on a 26-bed med-surg unit, and the unit secretary tells you the IT educator is doing a 10-minute fly-by at 0900 to “introduce” two new AI tools the hospital is piloting. You haven’t seen the email. You don’t have time before the 0700 huddle to read it. The charge nurse hands you assignments — six patients, two of them admissions still on telemetry from the ED, one a same-day discharge with brittle diabetes, one altered with a possible UTI on contact precautions, one post-op day two on a PCA. You’ll be lucky if the shift settles by 1100.
This guide is the thing the IT educator should have sent you before the fly-by, written by someone who actually understands that your bedside reality is not the same as a vendor demo. Two nurse-specific AI tools shipped in April 2026 with mainstream pilots at Cleveland Clinic, Cincinnati Children’s, and OhioHealth. They are not the same tool. They do different work, at different points in the shift. By the end of this read, you’ll know which one to invoke when, what each one actually writes back to the chart, what they don’t do (and won’t ever do), and the three questions to take into the staff meeting next week so the rollout serves the floor instead of administration.
What’s Actually New in April 2026 (The 90-Second Snapshot)
Tool 1: Hippocratic AI Nurse Co-Pilot. Launched April 16, 2026 (PRNewswire; Hippocratic AI product page). It’s a voice AI that calls patients on the phone — not chats with you, not summarizes notes, but actually places a phone call to the patient or caregiver — to handle teaching tasks: admission education, ongoing patient education, caregiver engagement around discharge, and medication adherence. You trigger every call from inside the EHR; the AI does the talking; a structured summary and a full transcript come back into the chart afterward. The vendor claim is 1-4 hours of nursing returned per shift; that number is directional, not peer-reviewed yet.
Tool 2: Ambience Healthcare Chart Chat for Nursing. Launched April 1, 2026 (BusinessWire; Fierce Healthcare). It’s a conversational AI sidekick embedded in the EHR. You ask it plain-language questions about your patient — “summarize this for change-of-shift,” “why is bed 12 NPO,” “what’s the trend on the A1c, last three results” — and it answers from the actual chart with citations you can click back to. Cleveland Clinic is the lead pilot site. The pitch is shrinking the 15-minute “piecing together a complex patient’s history” task to seconds.
One sentence to remember the difference: Co-Pilot does the talking-to-the-patient work. Chart Chat does the understanding-the-chart work. Same shift; different bottlenecks.
Tool 1, Up Close: Hippocratic AI Nurse Co-Pilot
The use case Hippocratic’s product team designed around is the time you currently spend explaining unit routines to a newly admitted patient at 1900 while three other call lights are ringing. That work is the kind that has to happen — JCAHO and your hospital’s HCAHPS scores both depend on it — but it’s the work that gets shaved when staffing is tight or the unit is hot.
Here’s how the Co-Pilot fits into a real shift.
The four workflows the bedside nurse triggers.
- Admission Education. A new patient arrives on the unit. After your initial assessment and vitals, you trigger the Co-Pilot for admission teaching. The AI calls the patient’s room phone (or cell, if you’ve confirmed they prefer that), walks them through the unit routine — call light use, fall precautions, pain scale, when meals come, when the team rounds — and does it consistently every single time. You stay free for assessments, med passes, and the patient who actually needs you in the room.
- Patient Education. A patient on a complex regimen needs scripted teaching during the stay — disease education, prep for an upcoming procedure, what to expect from a discharge plan that’s being built. You trigger the Co-Pilot. It calls. You see the call’s progress (ringing → answered → completed) on your screen. You read the structured note that drops in afterward and decide what (if anything) needs reinforcement at the bedside.
- Caregiver Engagement. This is the workflow that earns its keep most consistently for a discharge nurse on a complex case. A post-stroke patient is going home with their daughter as primary caregiver. The Co-Pilot calls the daughter — you can’t physically reach her right now because she’s at work — and walks her through swallowing precautions, the BP-monitoring schedule, when to call the doctor, what counts as a 911 call. The note it writes back into the chart confirms what was covered and flags any teach-back issues, so when you call her later you start from a known baseline instead of starting from scratch.
- Medication Adherence / Discharge Prep. Before a complex discharge — long MAR, new insulin regimen, new heart-failure cocktail — you trigger the Co-Pilot to walk through the meds, the dosing schedule, the side effects, and what trends to watch for at home. You verify the teach-back at the bedside before discharge. The Co-Pilot has done the volume work; you’ve done the clinical-judgment piece.
How the trigger actually works.
You launch it from inside the EHR. The Hippocratic press materials say nurse triggers it from the chart, no app download required. The vendors haven’t published Epic-specific or Cerner-specific screenshots yet, but at the launch sites — Cleveland Clinic and Cincinnati Children’s are large Epic shops, and OhioHealth runs Epic in most facilities — the realistic pattern is a button or side panel on the patient’s Storyboard or summary that opens a small embedded panel:
- Confirm patient identity (name, MRN, room/bed)
- Pick the workflow (Admission, Patient Education, Caregiver, or Med Adherence)
- Confirm the call target (room phone, patient’s cell, or caregiver number from demographics)
- Set the language preference if your tenant has multiple enabled
- Hit Start; monitor status (ringing, answered, completed)
- Read the note when it drops; sign off if your unit’s documentation policy requires it
That’s the whole interaction on your end. The Co-Pilot does the rest, on the phone, in real time. The note arrives as a separate documentation artifact (something like “Nurse Co-Pilot Education Note”) that you can scan during change-of-shift report and reference when calling the provider.
What it’s cleared for, what it isn’t.
Public materials describe it as built for inpatient nurses and inpatient education workflows. Cincinnati Children’s being a launch partner suggests pediatric configurations exist alongside adult. Nobody at Hippocratic has publicly claimed ED, ambulatory, or home-health coverage for the Co-Pilot product line yet — that’s important to know if your unit role bleeds into ED or rapid-response work.
Languages: Hippocratic’s broader voice AI platform is multilingual, but the April press doesn’t enumerate a specific list (English / Spanish / Vietnamese / etc.) for Co-Pilot at launch. Ask your IT/CNIO contact what’s enabled in your tenant. If you’re on a unit with significant non-English-primary patients, this is a make-or-break configuration question.
Tool 2, Up Close: Ambience Healthcare Chart Chat for Nursing
The use case Ambience’s product team designed around is the thing every charge nurse and float nurse knows: you pick up a patient at 0700 with twelve days of accumulated chart, six different consult notes, three medication adjustments overnight, a complicated history of present illness, and you have eight minutes before bedside report ends to make sense of it. The honest current workflow is “skim the H&P, scan the most recent two progress notes, look at the MAR, hope you didn’t miss something.” Chart Chat is built to make those eight minutes more useful.
What Chart Chat does in the EHR.
It lives as a panel inside the EHR — likely a side pane tied to the active patient’s chart. You open it. You type or dictate a question. It searches the patient’s chart in real time and answers, with citations. Click any citation, you go straight to the source — the progress note, the order, the lab result, the policy. The advertised source set per Ambience’s own product page and the HIT Consultant launch coverage is: physician progress notes, other clinical documentation, hospital policies, orders, and recent labs.
The questions worth asking it (real examples from Ambience’s launch materials).
- “Summarize this patient for my change-of-shift report.” — Generates a short SBAR-style summary from notes, consults, labs, and orders. Use during the last fifteen minutes of your shift before report.
- “What’s this patient’s current oxygen requirement and how has it changed in the last 24 hours?” — Pulls vitals, respiratory notes, and orders. Useful before calling the provider about a sat trending down.
- “Why is this patient NPO?” — Finds the GI note or procedure order explaining the rationale and cites it. Saves you the page-through.
- “What are the current parameters for holding the beta blocker?” — Reads the order set and policy, surfaces hold criteria with the citation.
- “What were the last three A1c results, and what’s the trend?” — Pulls from labs, comments on improvement or worsening with sources.
- “Explain this diagnosis in simple terms” / “What’s the plan for this patient’s heart failure?” — Especially useful for new grads, float nurses, or anyone covering a population they don’t typically work with.
What’s in the answer that matters.
Two things distinguish Chart Chat from “just paste the note into ChatGPT,” which the floor has been doing informally for a year. First, it pulls from your patient’s actual chart, not a generic body of medical knowledge. Second, every answer includes citations you can verify. Per Fierce Healthcare’s coverage, Ambience signals when there’s ambiguity or insufficient data instead of fabricating — that “I’m not sure” output is the difference between a tool that helps you and a tool that gets you in front of risk management.
The pilot data, and the gap in it.
Cleveland Clinic is the lead pilot. Public quantitative data is thin — what’s been published is qualitative (“nurses report saving meaningful time on chart review”) and adjacent (“Mercy Health has reported ~2 hours of charting saved per 12-hour shift with adjacent ambient AI”). Hard nurse-specific time-savings data on Chart Chat will come in the next two to four months as the Cleveland Clinic pilot publishes its first-quarter results. Until then, treat the time-savings number as a vendor estimate, not a peer-reviewed outcome.
Using the Two Tools Together: A Real Shift
The cleanest mental model is to think of Co-Pilot and Chart Chat as covering the two opposite ends of the shift’s chart-and-communication burden.
| Time | Task | Tool that helps | What you still own |
|---|---|---|---|
| 0700-0730 | Pick up six patients, prep for bedside report | Chart Chat: “Summarize patient in 12B for change-of-shift” | The clinical synthesis, the questions you ask the off-going nurse |
| 0900 | New admission from ED | Co-Pilot Admission Education: call room phone, walk through unit routine | The hands-on assessment, the IV start, the safety checks |
| 1100 | Vitals trending: bed 7’s O2 sat down to 91 | Chart Chat: “What’s the O2 trend last 24 hours” | The clinical decision, the call to the provider, the documentation of your response |
| 1300 | Complex discharge prep on bed 9 | Co-Pilot Med Adherence: call patient/spouse, walk through new med regimen | Teach-back at the bedside, sign-off, “any questions?” before they leave |
| 1500 | New transfer to your unit, dense chart | Chart Chat: “Why was this patient transferred? What’s the active treatment plan?” | Your introduction, your assessment, your update to the charge nurse |
| 1830 | End-of-shift charting | Neither — this is yours | The narrative note, the medication reconciliation, the SBAR for the next nurse |
The pattern: Co-Pilot replaces the patient phone-call work you’d otherwise be squeezing into pockets of time. Chart Chat replaces the chart-skimming work you’d otherwise be doing standing up at the COW (computer-on-wheels). Neither replaces clinical judgment, hands-on assessment, or the nurse-to-nurse handoff that the system depends on.
What These Tools Do NOT Do
This is the part nobody emails you. It matters more than the marketing.
Neither tool replaces a bedside assessment. A patient’s color, their respiratory effort, their level of distress, their ability to teach-back — that’s your work, and the Co-Pilot’s call cannot substitute. The Co-Pilot’s note will tell you the patient verbalized understanding; you still verify at the bedside, especially for older adults, post-op patients, or anyone with cognitive concerns.
Chart Chat does not catch what isn’t in the chart. The patient’s daughter mentioning yesterday that the mother has been more confused at home for the past two weeks — that’s almost certainly not in the structured chart fields, and the AI won’t surface it unless someone documented it. Your nursing assessment, your conversation with the family, your clinical pattern recognition — all still required.
Neither tool is cleared for emergent decision-making. If a patient is decompensating, the workflow is the same as it has always been: rapid response, charge nurse, provider, escalation. The Co-Pilot is not a triage line. Chart Chat is not a code blue tool.
Neither tool changes scope of practice or liability. If the Co-Pilot’s education call missed a critical discharge instruction and the patient bounces back in 48 hours, the documentation says you triggered the call, you reviewed the note, you signed off. The duty of care didn’t transfer to the AI when you clicked Start.
The vendors haven’t yet published peer-reviewed nurse-specific outcomes. The pilot data exists; the publications take time. Treat the “1-4 hours per shift” and “saves 15 minutes on chart review” numbers as directional vendor claims until your hospital’s own pilot report drops. Most facilities will publish those internally within the first two quarters of rollout.
Three Questions to Take to Your Staff Meeting
If your hospital is rolling out either tool — and if you’re at Cleveland Clinic, OhioHealth, or Cincinnati Children’s, you already are — these are the three questions to walk into the next staff meeting with.
Question 1: “Who owns the Co-Pilot’s note when something goes wrong?” If a Co-Pilot call missed a discharge instruction or said something inaccurate, what’s the workflow for amending the note, who notifies risk management, and how does it show up in our quality reviews? Don’t assume the answer is the same as a paper teaching-record handoff. Get it documented.
Question 2: “What’s the language coverage on our tenant?” If your unit has significant non-English-primary patients, you need to know which languages the Co-Pilot is configured for in your specific deployment. The vendor capability and your tenant’s configuration are not the same thing. If Spanish isn’t enabled and 30% of your patients are Spanish-primary, the Co-Pilot is a partial-coverage tool, and you need to plan accordingly.
Question 3: “What’s our escalation path when Chart Chat says ‘insufficient data’ but the chart has the answer?” This will happen. The AI will sometimes miss a citation that exists. The escalation path — who do you flag it to, how does the model improve over time, are nurse-flagged misses actually being incorporated — is the difference between a tool that improves with floor feedback and a black box you tolerate.
The Tools Are Not Replacing You
Worth saying directly. Voice AI calling patients with admission education does not replace the nurse who walks into the room thirty minutes later, looks at the patient’s color and breathing pattern, asks how they’re really feeling, and catches the early-pneumonia presentation that wasn’t in the H&P. Conversational AI in the chart does not replace the nurse who notices the patient’s daughter looking exhausted in the family chair and quietly asks whether anyone has spoken with palliative care. The work that AI is automating is the volume work. The work it isn’t automating — and structurally can’t — is the work that makes nursing nursing.
What’s changing is what your hour at the bedside is for. Less time on the seven-minute admission-orientation script. More time on the assessment that changes the plan of care. Less time triangulating the chart. More time on the patient. The role isn’t shrinking. It’s getting more clinical, faster.
Bottom Line
Two new tools, both narrowly defined. Co-Pilot owns the patient-phone-call education work. Chart Chat owns the chart-synthesis-in-the-EHR work. Both have launched at large pilot sites in April 2026 — Cincinnati Children’s, OhioHealth, Cleveland Clinic, with Cleveland Clinic doubling for both — and both should reach broader deployment in the back half of 2026.
If your hospital is in the rollout, the practical posture for week one is: pick one workflow you genuinely lose time to (probably admission teaching or chart-skimming for change-of-shift report), trigger the relevant tool on a single patient who’s stable enough to be a low-risk first try, read the output critically, and decide for yourself whether the time-savings claim survives contact with reality on your unit. Then take the three questions above to your staff meeting. Your floor’s experience over the next ninety days is what the next iteration will be built on.
