Learn
Home Foundations Glossary Research
Do
Prompts Workflows Tasks
Adapt
Domains Settings Patterns
Verify
Antipatterns Case Studies Policies Resources

Articulation & Phonology

LLM use for articulation and phonological disorder documentation, goals, and materials.

Articulation and phonology are among the most common SLP caseload areas. LLMs can help with goal wording, session note organization, and material generation, but the speech-specific knowledge must come from you.

LLM Strengths in This Domain

  • Generating word lists by phoneme position, syllable shape, or phonological pattern
  • Brainstorming minimal pair contrasts for specific processes
  • Structuring progress notes with target, support level, accuracy data
  • Drafting parent-friendly explanations of sound development
  • Creating home practice activity lists organized by target

LLM Limitations

  • Cannot assess stimulability or phonetic placement
  • May not understand the distinction between articulation and phonological process disorders
  • Word lists may include developmentally inappropriate items without guidance
  • Cannot interpret speech samples or connected speech data
  • May not understand motor-based vs. linguistically-based intervention distinctions

Prompt Templates

Word List Generator

Generate a list of 20 words with /s/ blends in the initial position, appropriate for a 6-year-old. Include a mix of 1- and 2-syllable words. Avoid words that are culturally or contextually unusual for a school-age child. Organize by blend type (sl-, sm-, sn-, sp-, st-, sw-, sk-).

Minimal Pair Generator

I am an SLP working on the phonological process of fronting with a 4-year-old. Generate 10 minimal pair sets contrasting velar and alveolar stops (e.g., key/tea, go/doe). Use words a preschooler would know. Include both initial and final position pairs.

Progress Note for Articulation Session

I am an SLP. Here are my raw session notes for an articulation therapy session (no identifying info included). Organize into: target sound and position, activity type, cueing level, accuracy data (trials and percentage), behavioral observations, and plan for next session. Do not add details I did not provide.

Goal Progressions

Weak → Strong: Sound-Level Goal

Before: “Student will improve articulation skills as measured by therapist observation.”

No target sound. No position. No accuracy criterion. No measurement method.

After: “Given a structured activity with visual supports, [Student] will produce /r/ in the initial position of words with 80% accuracy across 3 consecutive data collection sessions as measured by SLP data tracking.”

Weak → Strong: Phonological Process Goal

Before: “Student will reduce use of phonological processes.”

Which processes? At what linguistic level? What counts as “reduced”?

After: “Given a minimal pair activity with models, [Student] will produce velar consonants /k, g/ in the initial position of words, eliminating fronting, with 75% accuracy across 4 consecutive sessions as measured by SLP data collection.”

Progression Hierarchy

A typical articulation goal should progress through these levels, and each level needs its own measurable goal:

  1. Isolation → single sound production
  2. Syllables → CV, VC, CVC combinations
  3. Words → initial, medial, final positions
  4. Phrases → carrier phrases with target words
  5. Sentences → structured sentences with target words
  6. Conversation → spontaneous connected speech
  7. Generalization → across settings, listeners, contexts

LLMs can help you draft goals at each level. They cannot tell you which level the student is ready for.

Before & After: Session Documentation

Raw Notes

did /s/ blends again, tried sn- and sp- words, she got most of the sp- ones right but sn- was harder, maybe 50% on sn- and 80% on sp-, used picture cards, gave her a verbal model when she got stuck, she was really engaged today, wants stickers

Structured Output

  • Target: /s/ blends (/sp-/ and /sn-/) in initial position of words
  • Activity: Picture card naming with verbal models as needed
  • Performance:
    • /sp-/ blends: ~80% accuracy with verbal model on errors
    • /sn-/ blends: ~50% accuracy with verbal model on errors
  • Behavioral Note: High engagement throughout session. Motivated by sticker reinforcement.
  • Clinical Interpretation: /sp-/ approaching mastery at word level; /sn-/ requires continued practice with increased cueing. Consider adding visual placement cues for /sn-/.
  • Plan: Continue /sn-/ at word level with visual + verbal cues. Probe /sp-/ at phrase level next session.

Common Mistakes When Using LLMs

  1. Accepting generic word lists. Always check that words are age-appropriate, culturally relevant, and phonetically accurate for your target.
  2. Missing phonetic context. A word like “sun” targets /s/ in initial position, not an /s/ blend. The model doesn’t always distinguish this.
  3. Confusing articulation and phonology. If you’re treating a phonological process, your goals and activities should reflect that (minimal pairs, pattern-based) rather than traditional articulation drill.
  4. Overly complex goals. LLMs tend to make goals wordier than necessary. Simpler is better if it’s still specific.

SLP/IO Assistant

Powered by Claude · No PHI accepted
AI assistant for clinical workflow support. Never enter student names, DOBs, or identifiable information.
Hi! I'm the SLP/IO assistant, an opinionated AI grounded in clinical practice. I can help with goal wording, note structure, ethical reflection, and navigating LLMs responsibly. What are you working on?