Technology
AI for SLPs: How Speech Therapists Are Using AI for Documentation, Note Taking, and Practice
The tools reshaping the SLP workflow, the privacy considerations to watch for, and what AI still cannot do
April 12, 2026•10 min read
The tools reshaping the SLP workflow, the privacy considerations to watch for, and what AI still cannot do
A few years ago, AI was something most speech language pathologists read about in passing. Today it is showing up in the actual workflow. SLPs are using AI to draft session notes, summarize assessment data, generate practice materials, and provide students with feedback on home practice. The tools are improving fast and the time savings are real.
This guide walks through how AI is being used across the SLP workflow today. We will cover the major tool categories, name the products SLPs are actually adopting, flag the privacy considerations that matter, and talk honestly about where AI helps and where it does not.
Documentation is where AI is making the biggest dent in SLP workload. Ambient AI scribes listen to a therapy session (with patient consent), transcribe what is said, and generate a structured clinical note in seconds. The SLP reviews, edits, and signs. Studies in healthcare more broadly have shown ambient scribes can save clinicians one to two hours of documentation time per day. For SLPs writing SOAP notes, progress notes, and session summaries, the time savings translate directly into either more billable sessions or actually leaving work on time.
A few examples of AI scribes and note taking tools SLPs are using include Heidi Health and DeepScribe (both popular among allied health professionals), Ambience Healthcare (widely adopted in larger health systems), Suki AI (voice activated, integrates with major EHRs), and Mentalyc (originally built for mental health therapists, used by some SLPs in private practice). The right fit depends on your setting, your EHR, and whether the vendor supports speech therapy specific note formats.
General consumer tools like ChatGPT or Claude are not HIPAA compliant by default and should not be used with identifiable patient information.
Assessment reports are one of the most time consuming parts of an SLP's job, especially in school based settings where evaluations stack up before IEP deadlines. AI can help in several ways:
The important caveat: AI generated drafts need clinical review. The SLP is still the professional making the diagnostic decisions and writing the goals. AI just speeds up the first draft. For a deeper look at writing strong IEP goals, see our guide to speech therapy IEP goals.
Documentation gets the headlines, but the use case that may have the biggest clinical impact is AI powered practice. Here is the problem: speech therapy progress depends heavily on what happens between sessions. Motor learning research is clear that frequent, accurate practice with feedback drives carryover. But for most students, between session practice does not happen, or it happens without feedback because parents cannot reliably judge whether their child's /r/ or /s/ production is correct.
AI speech recognition can now provide immediate feedback on speech sound accuracy. The student says a word, the AI scores the production, and the student knows in real time whether they got it right. This solves the feedback problem that has historically made home practice ineffective.
LumaSpeech is a purpose built AI articulation and fluency practice tool for SLPs and their students. SLPs assign targets in seconds. Students practice at home with AI feedback on every production. The SLP sees what was practiced and how accurately, so the next session starts with data instead of guesswork.
Read more about how AI articulation therapy uses real time feedback and why speech therapy homework compliance is so closely tied to the feedback problem.
AI practice tools are not a replacement for therapy sessions. They are an extension of them. The clinician designs the program and the AI handles the high volume, between session repetition that motor learning requires.
Generative AI tools like ChatGPT and Claude are useful for the creative parts of session prep, as long as no identifiable patient information is involved. SLPs are using them to:
These uses are generally lower stakes than documentation or assessment because they do not involve patient data. The output still needs clinical review for accuracy and appropriateness, but the time savings on session prep are substantial.
It is worth being direct about this: AI is a tool, not a clinician. The parts of speech therapy that depend on human judgment, relationship, and adaptive expertise are not going away anytime soon.
Distinguishing CAS from severe phonological disorder, or untangling a child's speech profile when multiple things are going on, requires clinical reasoning that AI cannot reliably do.
Trust, rapport, and the moment to moment adjustments a therapist makes based on a child's emotional state are at the heart of effective therapy. AI cannot read a room.
Talking a worried parent through a diagnosis, supporting them through the IEP process, or helping a family adjust to long term communication needs takes empathy and judgment. These are core SLP skills, not workflow steps.
Knowing when to step back, when to push, when to change approaches mid session, when a child needs a break. This kind of adaptive clinical decision making is uniquely human.
The right framing is that AI handles the repetitive, time consuming, and structured tasks so SLPs have more time and energy for the parts of the job that require their expertise.
A short checklist for any AI tool you are considering adopting in your practice:
AI in speech therapy is moving fast on three fronts. Documentation tools are getting better at capturing the nuances of clinical conversation. Practice tools are improving their accuracy on subtle articulation distinctions. And generative AI is making it easier to produce individualized therapy materials at scale.
The SLPs who benefit most from this shift will be the ones who adopt AI strategically, treat it as augmentation rather than replacement, and stay critical about both the privacy and clinical limitations. The goal is to spend less time on paperwork and more time doing the clinical work only an SLP can do.
AI is no longer a future trend in speech therapy. It is in the workflow now, helping SLPs with documentation, assessment, materials, and student practice. The right tools, used carefully, give SLPs back the one resource they have the least of: time. The best use of that time is the part of therapy that has always mattered most, the human work of helping someone communicate.