AI Scribe Transition Support Guide
The AI Scribe Transition Support Guide is a step-by-step resource designed to help clinicians transition to using an AI scribe. This guide outlines each stage of the transition process, highlighting key actions and considerations. The guide also includes useful resources for added support. Learn all you need to know before you adopt an AI scribe.
AI scribes are digital tools designed to automate the administrative tasks of documenting patient consultations. AI scribes use artificial intelligence to summarize or capture spoken conversations with consenting patients into electronic and clinically relevant medical notes for health care professionals.
AI scribes: benefits and challenges
- Gain a general understanding of AI scribes, their capabilities, functionality (AI Essentials) and challenges (When AI Gets It Wrong: Hallucinations in AI Scribes).
- Identify pain points in clinical documentation and administrative tasks.
- Gather data on time spent charting and its impact on provider well-being and patient care.
- Define clear objectives for AI scribe adoption (e.g., reduce documentation time by over 70%).
- Engage clinicians, administrative staff, and IT teams to understand workflow requirements.
Additional Resources:
- AI Scribe Demonstration
- OMD Educates – AI 101: Intro to AI and AI Scribe
- Can AI Boost Safety and Quality in Patient Care?
Who to involve in selecting an AI scribe
- Secure leadership buy-in by demonstrating the value of AI scribes.
- Identify key stakeholders, including physicians and other clinicians, administrative staff, and IT support.
- Establish a change champion or project lead to guide the transition.
Additional Resources:
How to choose an AI scribe
The Ontario AI Scribe Program’s Vendor of Record (VOR) list of AI scribe vendors has ensured that critical privacy, legal and security considerations have been built into the program. This included vetting vendors and their products for important privacy and security requirements and drafting contract terms that support your practice and functionality. A variety of stakeholders, including OntarioMD and Ontario Health, provided input to develop these requirements to align with the professional obligations of physicians.
Supply Ontario (SO) have pre-qualified AI scribe vendors who have met provincial VOR requirements for clinical functions, privacy, and data security. SO can provide assistance to clinicians with vendor selection through the Ontario AI Scribe Program. You can access the 2-step SO procurement process by first filling out the Expression of Interest form.
Some of the key advantages for clinicians who use the Ontario AI Scribe Program’s VOR:
Reduce Administrative Burden
- Selecting the AI scribe that is right for you and your practice can free up more time for patient care.
- Reduce the time you spend on documenting patient encounters by 70% or more, and save almost four hours a week on administrative tasks.
Compliance with Provincial Requirements
- A variety of stakeholders, including OntarioMD, provided their input to develop requirements that align with the medico-legal obligations of clinicians
- The vendors and their products were required to satisfy provincial requirements, including those related to clinical functions, privacy, security and other legal considerations.
Fair Pricing
- The program eliminates guesswork and relieves the burden of sourcing an AI scribe yourself, saving you both time and money.
Peace of Mind
- Vendors on the VOR list offer contracts that were evaluated by legal and privacy experts.
- You will also receive vendor management support from Supply Ontario throughout the journey of acquisition and use.
Streamlined Support
- Participation in the program provides you with support at every step of the process to adopt and use an AI scribe.
- Supply Ontario will help you select the right AI scribe based on your requirements.
- OntarioMD will answer your questions about AI scribes, help manage the changes to practice and workflows during adoption and provide ongoing advice to optimize your use.
Additional Resources:
- Choosing an AI Scribe: Evaluation Matrix
- OMD Educates: Beyond the Hype: Choosing an AI Scribe
- Legal and Privacy Considerations
- Preparing your practice for an AI scribe
- Select appropriate devices (e.g., exam room computer, handheld device, or both) and acquire necessary hardware (such as high-quality microphones).
- Review technical requirements, including supported platforms (web, iOS, Android) and operating systems (Windows, Mac).
- Assess internet requirements and redundancy solutions to ensure reliable connectivity during AI scribe use.
- Review privacy and security requirements, including relevant policies, standards, and regulations that may apply to AI tools in your clinical setting.
- Assess potential downtime scenarios and identify mitigation strategies or workaround protocols to ensure documentation continuity and minimize workflow disruption.
How to Manage Patient Consent
- Streamline collection of valid consent (OMD Patient Consent Toolkit):
- Online: Pre-visit forms and booking systems
- Online: Pre-visit forms and booking systems
- In-clinic: Update your privacy policies and/or signage to include notice of use of recording and retention of data for the use of AI in patient encounters.
- During visits: Have an information sheet in simple, understandable language available for your patients to review. This is not a replacement for having an informed discussion with your patients. If your patient has opted out of using an AI scribe, don’t use it.
Additional Resources:
Adopting an AI scribe means changes to your workflows and helping your staff and patients adapt to the new tool.
OMD is committed to helping you manage changes as smoothly as possible when implementing an AI scribe. Contact support@ontariomd.com at any time during your AI scribe implementation.
How to run a pilot implementation
- Implement the AI scribe on a limited scale (i.e., a pilot), allowing for real-time observation and support.
- Clinician offices may consider piloting with a limited number of patient visits, specific days, specific hours or types of appointments.
- Multi-clinician offices may also choose to select a small group of clinicians to participate, or rotate the pilot across clinicians or departments.
- Define metrics that help you define success. You can also review the OMD AI Scribe Pilot when designing your evaluation framework.
- Clinician offices may choose to measure success based on individual indicators such as reduced documentation time, workflow improvements, improved provider or patient satisfaction, and better work-life balance.
- Multi-clinician offices might also choose to focus on team collaboration and scalability metrics.
- Monitor technical performance and gather lessons learned.
- Clinician offices should monitor the AI scribe’s performance, focusing on accuracy, reliability, speed and integration with the various workflows. Reflect on any issues or areas for improvement.
- Multi-clinician offices can also track performance across clinicians and gather feedback to identify trends, improve scalability and address technical challenges.
Get training and support
- Obtain training for clinicians and support staff on AI scribe use, troubleshooting, and workflow integration.
- Leverage vendor training resources (i.e. user guides, FAQs, and live or virtual demonstrations).
- Leverage external supports such as OMD Practice Hub, OMD Advisors and OMD Peer Leaders (support@ontariomd.com).
- Encourage clinicians to engage in simulated encounters before full implementation.
How to evaluate a Pilot Implementation
- Assess outcomes against your predefined success metrics. You can review the AI Scribe Evaluation Final Report Pilot as a reference.
- Evaluate clinicians’ experience and confidence, patient impact and workflow integration.
- Validate note quality, accuracy, consistency, progression in error reduction.
- Identify technical or workflow adjustments that may be needed.
- Clinician offices should review the AI scribe for what is working and what is not. (i.e., system improvements, reporting issues, provider/patient satisfaction).
- Multi-clinician offices may also establish a feedback loop with structured questions. Analyze and share feedback insights with stakeholders and team(s).
Optimize and Expand AI Scribe Use
- Update training, documentation or workflows as needed.
- Make technical or workflow adjustments that are needed.
- Develop a rollout strategy for broader implementation based on pilot learnings.
- Clinician offices may increase usage across more patient visits, clinic hours, or visit types. Multi-clinician offices may also increase usage by rolling it out to additional clinicians, roles, or sites.
- Continue to gather ongoing feedback and make iterative improvements.
Establish practice standards for efficient and effective use of AI scribes
- Create clear policies and procedures to guide responsible AI scribe use, including:
- Patient consent (see ‘How to Manage Patient Consent’ in the Learn section for guidance).
- Privacy and security concerns (see Privacy for key considerations).
- Be mindful of risks such as being overheard or inadvertently recording sensitive information.
- Design workflows that uphold privacy and security standards and comply with regulatory requirements.
- Develop downtime protocols for periods when your AI scribe is unavailable, including:
- Contingency workflows
- Communication plans
- Manual documentation protocols to manage:
- Delays
- Loss of notes
- Freezing or failed saves where the note cannot be recovered
As you become a better user of your AI scribe, you’ll want to take advantage of tips, tricks and best practices from experts and other users. OMD has assembled a growing list of information and resources to help you. We are also only an email or call away at support@ontariomd.com or 1-866-744-8668 whenever you need help.
Standardize and Streamline your Practice for AI Scribe Use
- AI scribes enter information as a free text chart note, which may not be recognized in a useful way by your EMR. Based on personal practices and preferences, important elements of the note may need to be manually entered into other sections of the EMR. This will improve the accuracy and reliability of EMR tools such as searches, reminders and decision support tools. Here are some examples of when this may apply:
- CPP: Current Medications, Problem Lists, Past Medical and Surgical Histories, Family History, Immunizations, Risk Factors and Allergies.
- Vitals: It should be possible for the AI scribes to document vitals in a format that is recognized by the EMR. If this is not happening, it is recommended that the AI scribe template could be modified to achieve this.
- Lab results, Imaging and other Diagnostic tests.
- Use templates that work well for you (content, format, style, etc.). AI scribes use templates to generate different types of chart notes from the transcripts of the conversation. These chart notes can include SOAP notes, consults, referral letters and reports etc. These templates can be shared and customized. To find a template that will work best for you, try these options:
- Use standard templates that come with the AI scribe
- Find a template in a shared community
- Customize an existing template
- Create a new template
Vendors can provide support on how to find and customize templates.
Routinely review for accuracy
Regularly audit scribe-generated documentation for accuracy.
Note review and editing:
- Ensure the quality, accuracy and completeness of the chart note. All notes are reviewed by the provider to eliminate errors, omissions, biases and hallucinations etc. (refer to bias and hallucinations).
- Review AI-generated notes as soon as possible after each visit.
Edit within the AI scribe platform.
- AI scribe platforms have AI-powered tools that will assist with editing.
- You can dictate additional information, which will be added into the chart note.
- The AI scribe can learn from your edits to automatically generate customized notes for you.
Edit within the EMR.
- While you are editing, you have access to the patient chart, which can help inform the chart note
- You can enter information into sections of the patient chart as you are editing the note
- You can use EMR formatting (bold, highlight, italicize)
- Beware of AI hallucination. AI scribes may introduce content that no one said, misrepresent key details, or rearrange the order of clinical events. Always review and edit AI-generated notes.
- Be cautious about AI bias. If AI training data includes historical or systemic bias, these patterns and assumptions can appear in documentation, i.e., descriptions of dermatological symptoms may be inaccurate or vague for patients with darker skin tones if the model was trained on lighter-skin datasets. Review AI-generated notes for subtle assumptions that don’t align with your clinical judgment or knowledge of the patient – especially in sensitive or high-risk documentation.
Review physician feedback monthly and engage your vendor for updates on model accuracy.
Additional Resources:
Refresh learning and support services regularly
- Conduct regular refresher sessions to strengthen user skills and reinforce updated workflows.
- Review updates on new features and AI enhancements.
- Access hands-on workshops, real-time transition support, and tailored training materials that are available.
- Develop a user feedback loop for users to report issues, suggest improvements to the vendor, and share common issues and solutions with fellow users where appropriate.
- Establish and maintain a comprehensive troubleshooting and escalation protocol, including up-to-date contact information for all key support resources (e.g., vendor technical support, OntarioMD Advisor at support@ontariomd.com).
- Maintain a resource hub with ‘how to’ guides, release notes, known issues, and quick tips.
Foster a culture of collaboration and feedback
- Promote collaboration and feedback by seeking input from patients, team members, peers or networks – and sharing your own experiences to support others.
- Celebrate innovation and efficiency by nominating your practice or team for the OMA Award and Recognition Program.
- Share success stories with OntarioMD Marketing & Communications team to communicate the benefits of workflow improvements and efficiencies to and inspire broader learning.
- Seek individualized support as needed.
- Identify and share early wins with your team, colleagues, or network, e.g., more enjoyment of encounters, less time spent at home doing notes, or positive patient feedback.
- Recognize and reward early adopters who embrace the technology by highlighting contributions within your team, practice or network.
- Participate in provincial forums like OntarioMD’s Communities of Practice to connect with others, learn from shared experiences and contribute insights.
Make Iterative Improvements
- Understand your AI scribe’s limitations and inaccuracies and share your learnings with colleagues.
- Keep apprised of new technology and enhancements that become available.
- Follow and track your vendor’s updates and change logs.
- Check regularly for advice from the profession for updates on AI guidance. (CPSO, OMA, CMPA, etc.).
- Leverage collated and themed feedback insights to prioritize improvements.
Additional Resources:
Consider these features and functionality for an AI scribe according to your practice needs:
| Category | Function | Description |
| Recording/Transcription | Real-Time Transcription | Providing comprehensive, near-instant transcription (within 30 seconds) of the patient-physician conversation, capturing key details in real-time for immediate use. |
| Pause and Resume Transcription | Allowing the clinician to pause and resume transcription without data loss, supporting workflow interruptions and enhancing flexibility. | |
| Background Noise Filtering | Filtering out background noise and irrelevant content, ensuring only essential information is captured in clinical notes. | |
| Speaker Differentiation | Differentiates between multiple speakers (e.g., clinicians, patients, caregivers) and manages conversations involving multiple speakers in both transcription and notes. | |
| Notes | Detailed Medical Data Capture | Automatically generating detailed medical notes covering the full clinical information of the patient visit, formatted as structured notes, such as SOAP notes. |
| High-Quality, CPSO-Compliant Notes | Ensures all generated documentation aligns with CPSO standards, producing high-quality, compliant medical records. | |
| Multiple Note Templates | Offering various note templates (e.g., SOAP, referral, consult, summary) to suit different documentation needs and preferences. | |
| Customizable Note Templates | Allows clinicians to create or modify note templates to match specific practice requirements or documentation preferences. | |
| Flexible Template Selection | Enables switching between note templates mid-session without losing transcribed data, allowing for seamless adaptation during the visit. | |
| Direct Editing Options | Supports direct editing of generated notes within the platform via dictation or keyboard, giving control over final content. | |
| Language | Accent, Slang, and Acronym Recognition | Accurately transcribes diverse accents, slang, and clinical acronyms common in health care, capturing nuanced language used in clinical interactions. |
| English or French Language Support | Supports transcription and note generation in English or French, ensuring functionality in English or French-speaking healthcare settings and facilitating bilingual interactions. | |
| Multilingual Support | Supports clinician-patient consultations in additional languages with note generation in languages other than English and French. | |
| Patient | Patient-Facing Visit Summaries | Generates patient-friendly visit summaries and instructions, enhancing patient understanding and engagement in their care. |
| Readability Adjustment for Patients | Tailors content to accommodate different patient reading grade levels. | |
| Integration | Integration with EMRs | Allows for flexible transfer of notes to Electronic Medical Record (EMR) systems, such as through copy-paste functionality. |
| Quality Assurance | Continuous Accuracy Monitoring | Regularly assesses and improves transcription accuracy and reliability, adapting to new language patterns and technological advancements. |
| Personalization | Learning User Preferences | Learns and retains physician preferences, editing styles, frequently used terms, and templates to deliver a personalized experience over time. |
