The session already contains the story. Your SOAP note should not start from an empty page.
AI SOAP notes for SLPs, built from the session—not from your memory.
RelyCare turns speech therapy session audio into a review-ready Subjective, Objective, Assessment, and Plan draft. Instead of reducing the visit to a generic summary, it keeps audio-supported targets, first attempts, retries, self-corrections, cueing, errors, representative stimuli, and supporting transcript quotes available for clinician verification. You still edit and approve the final record; RelyCare gives that work a more grounded starting point.
Initial /s/ production — Word level
Draft for verification: 8/17 accurate first attempts (47%); 7 successful cued retries; 2 self-corrections. Representative transcript evidence attached.
Final /s/ clusters — Phrase level
Draft for verification: 12/15 supported first attempts (80%). One /sp/ production was self-corrected after an auditory model.
The Core Promise
Review an evidence-linked clinical draft instead of rebuilding the session from a blank page after hours.
Before
Blank page staring contest
You finish the session, try to remember what happened, and start typing from zero at 8pm while your brain is already checking out.
Review
A grounded starting point
RelyCare presents a structured draft with target data and representative evidence available for inspection. You correct the record, add clinical judgment, and decide when it is ready.
Control
The final note stays yours
The software does not sign, submit, or clinically validate the note for you. The treating clinician remains responsible for the final documentation and its use.
How It Works
The workflow should feel lighter with every step.
Step 01
Record with the required consent
Use RelyCare during an eligible single-patient session under your clinic's approved privacy, consent, and operating workflow. The recording becomes the source used to create a speaker-labeled transcript; it does not give the system access to visual observations or facts that were never spoken.
Step 02
Inspect the transcript and evidence trail
Review how audibly demonstrated or verbally reported details became atomic observations, draft target metrics, representative quotes, and SOAP sections. Check speaker attribution, scoring basis, original attempts, cued retries, self-corrections, cue levels, and any uncertainty before relying on the draft.
Step 03
Edit, approve, and carry context forward
Correct omissions or errors, add the clinical information the audio could not capture, and approve the final record only when it meets your standards. Finalized notes can feed the patient timeline for longitudinal review. Raw session audio is deleted after finalization, not immediately after transcription.
Subjective
Spoken context stays connected to the clinical draft.
RelyCare can organize caregiver reports, patient statements, and contextual changes that are verbally reported during the recording. It does not claim to see demeanor, eye gaze, posture, physiology, or other visual behavior from audio. The clinician decides which spoken context is relevant, corrects attribution when needed, and adds observations that occurred outside the recording before approving the Subjective section.
Caregiver Report
Mother reported consistent home practice this week. No concerns raised at session onset.
Patient Statement
Patient verbally reported practicing the assigned words after dinner and asked to begin with the matching game.
Objective
Supported trial data, cueing, and scoring context—organized for review.
The system identifies audio-supported trials and drafts target metrics for clinician verification while keeping original scored opportunities separate from cued retries. An illustrative output might read, "Initial /s/ at word level: 8/17 accurate first attempts (47%); 7 successful cued retries; 2 self-corrections," with the scoring basis, cue level, representative stimuli, and transcript evidence available for inspection. RelyCare does not promise that every opportunity will be detectable in every recording.
Initial /s/ production — Word level
Draft for verification: 8/17 accurate first attempts (47%); 7 successful cued retries; 2 self-corrections. Representative transcript evidence attached.
Final /s/ clusters — Phrase level
Draft for verification: 12/15 supported first attempts (80%). One /sp/ production was self-corrected after an auditory model.
Assessment
A synthesis grounded in documented performance, not invented causes.
The draft can summarize patterns supported by the current record, such as stronger performance with a specified cue, consistent first-attempt errors on a named target, or successful self-correction after feedback. It should not infer fatigue, attention, sensory state, emotional state, or readiness for advancement unless the audio explicitly supports that information. The clinician interprets what the pattern means and decides what belongs in the Assessment.
The audio-supported draft shows stronger /s/ performance after verbal or model cues than on scored first attempts. The treating clinician should interpret whether this reflects current cue dependence.
Two self-corrections were documented after feedback. No visual behavior, attention state, fatigue, or causal explanation is inferred from the audio.
Plan
Next-step language remains a proposal for clinician review.
RelyCare can draft follow-up targets, task contexts, cue-fading ideas, representative stimuli, home carryover, and homework based on the session record and supplied goals. Those suggestions are not treatment orders. The SLP checks whether the proposed level, success criterion, cueing approach, and home activity fit the patient, then edits or removes anything that does not reflect the intended plan of care.
Next Session Target
Continue /r/ at phrase level; target >75% accuracy with max phonemic cuing before advancing to sentence level.
Technique
Introduce structured carrier phrases ("The red rabbit...") to bridge word→generalization gap.
Home Practice
5-min daily "Red Robot" game: find red objects, say "red ___" 10x. No materials needed.
Common Questions
The details people usually want before they click deeper.
Does this replace my clinical judgment?
No. RelyCare drafts structure and organizes audio-supported details, but it cannot know everything the treating clinician observed or intended. You review the transcript-linked evidence, correct the draft, add your interpretation, and approve the final record. It is a documentation assistant, not a clinician or an external clinical validator.
How is this different from a generic AI medical scribe?
Generic scribes are usually designed to summarize broad clinical conversations. RelyCare is organized around SLP documentation concepts: operational targets, task level, scoring basis, first attempts, retries, self-corrections, cue hierarchy, error patterns, representative stimuli, evidence quotes, SOAP structure, and longitudinal patient context.
Will RelyCare save a specific number of minutes per note?
RelyCare does not publish an unmeasured time-saving guarantee. The intended advantage is a review-ready starting point that keeps the session evidence attached. A founder-led pilot lets your clinic compare its current workflow with RelyCare and choose success measures that can be evaluated honestly in your environment.
Can a HIPAA-regulated clinic use the SOAP note workflow?
A BAA is available for eligible clinics under the required contractual and operational conditions. Eligibility does not make every workflow automatically compliant. Your clinic must discuss consent, access, configuration, intended use, and privacy responsibilities before processing real session audio during a pilot.
Keep Exploring