Week 14 — Speech Workshop / Rehearsal Studio · "Draft & Record Your Special-Occasion Speech"

Course: Public Speaking — Fundamentals of Oral Communication (COMM 1) · Silver Oak University (fictional sample) · Prof. Marchetti
Objective: Objective 8 — plan and deliver a special-occasion speech with occasion fit, vivid language, appropriate brevity, and tone/mood match · SLO A (deliver) & SLO B (analyze — self-assess fit/brevity/vividness)
Worth 50 points · Speech Workshops group = 15% of the grade · Workshop 14
Format this week: draft-record-self-assess delivery drill — a phone camera or Zoom and an approved chatbot.

This week's Workshop is the rehearsal pass for your Special-Occasion Speech. You'll draft, record, and self-assess a 60–90-second toast or tribute before you submit the final version for Assignment 14. Think of the Workshop as your warm-up take — the assignment is your best take. Both happen this week.

---

Part 1 — The Big Picture

Special-occasion speeches are some of the most meaningful words anyone is ever asked to speak — and some of the most easily botched. The toast that drifts into a 12-minute ramble. The tribute so full of generic compliments that no one feels anything. The introduction that's actually a 3-minute autobiography of the introducer.

What makes a special-occasion speech land is not eloquence. It is one true, specific detail — a moment, a habit, a decision — that lets the whole room feel like they understand this person, even if they just met. Finding that detail and shaping it into 90 seconds of spoken language is the skill you're drilling this week.

The guiding question: When I watch my own toast or tribute back, does it feel generic — or does it feel like it could only be about this specific person in this specific moment?

---

Part 2 — The Drill: Draft, Record, and Self-Assess Your Special-Occasion Speech

Choose the same occasion you're using for Assignment 14 — or a comparable one (same type, similar stakes).

Do this:

Step 1 — Find your one specific detail. Before you pick up keywords, answer this question in one written sentence: What is one specific, true thing you know about this person (or occasion) that the whole room could picture? Not "she's always been dedicated." Something real: "She drove three hours to be at my recital and pretended her flight was delayed so I wouldn't feel bad she had to cancel something to come." Write the sentence.

Step 2 — Plan your keywords (one per beat).

Fill in this planning scaffold:

Beat	What it contains	Your keyword(s)
Vivid specific detail (or hook for introduction)	The one true specific thing
What it reveals (or topic/credentials for introduction)	What this shows about who they are
Wish / forward look (or audience reason for intro)	Where they're headed; why we should be excited
Close / signal (the raise, the handoff)	"Please raise your glass to…" / "Please welcome…"

Step 3 — Record your first take. Speak to the lens like it's the room at the real occasion. Aim for 60–90 seconds. Don't re-record yet — watch this take first.

Step 4 — Watch it once all the way through. This is the hard part. Watch like a coach, not a critic.

---

Part 3 — Self-Assessment Scaffold (fill in after watching your first take)

Rate each on a 1–5 scale (1 = needs major work, 5 = strong) and add one line of what you actually saw or heard:

What to watch for	Score (1–5)	What I noticed
Occasion fit — is this clearly the right type of speech for the occasion?	___	______
Tone / mood match — does the tone feel right for this occasion (warm, reflective, celebratory)?	___	______
Specific vivid detail — did you say something specific and true, or did it drift generic?	___	______
Organization — was each beat present and in the right order? Did you close decisively?	___	______
Brevity — did you land in the 60–90-second window?	___	______
Delivery — audible, conversational, looking at the lens, speaking from keywords?	___	______

Your ONE thing: after reviewing the scaffold, which single element would make the biggest difference to your next take? Write: "Next take I will ______."

---

Part 4 — Analysis Questions

Answer in a sentence or two each:

1. Was the specific detail you chose actually specific enough to land? What would you hear in the room — laughter, a catch in the throat, a nod of recognition — if you delivered it at the real occasion?
2. Which was harder this week: finding the right detail, or delivering it without reading? Why?
3. How is calibrating the tone for a toast different from calibrating the tone for a tribute at a memorial? What do you change between those two occasions?
4. After your "one thing to fix," do a second take. Did the single-change focus actually help?

---

Part 5 — Rehearsal-Coach Moment (BYOAI)

Bring in your approved chatbot (Gemini, Claude, or ChatGPT) as a rehearsal coach.

1. Share your planning scaffold keywords and a quick description of how your first take went (or paste the detail + beats as you drafted them), and ask: "I'm practicing a 60–90-second toast/tribute for my public speaking class. Here's my plan and how it went — give me specific, actionable feedback on whether the vivid detail lands and whether the tone matches the occasion."
2. Read its feedback and try its best concrete suggestion in your next take.

Critical step — keep the AI's role in bounds: the chatbot can help you sharpen language and check structure, but it should NOT invent new specific details about the person you're honoring. If it suggests adding a story about them that you didn't give it, that is a fabrication. Anything specific in your speech about a real person must come from what you actually know.

---

Part 6 — AI-Critique Moment (required — the BYOAI judgment step)

Special-occasion speeches are the week where the chatbot's default failure mode is especially visible: generic flowery sentiment. It defaults to phrases like "May your future be filled with joy, laughter, and love" because it doesn't know the person. Here's how to catch it:

1. After the chatbot gives its feedback, check: did it give you anything truly specific, or did it offer generic suggestions ("add more warm details," "make it more personal") that don't actually help?
2. Push it: "Give me a concrete revision of my specific-detail sentence that makes it even more vivid and shows rather than tells — based only on what I gave you." See whether it can do it without inventing new facts.
3. Write 2–3 sentences reporting: one example of generic or hollow feedback the chatbot gave (or generic language it suggested), and what specific, actionable feedback would have said instead.

Key distinction this week: generic sentiment in the speech ("she's always been wonderful") is the student's mistake. Generic feedback from the AI ("great job, sounds heartfelt") is the chatbot's mistake. Both are worth catching — but they're different failures.

---

Part 7 — What to Submit

Submit a single document (or text entry) with: your planning scaffold (Part 2), your self-assessment scaffold (Part 3) with your "one thing," your Part 4 analysis answers, and your Part 6 AI-critique paragraph. Include your recording (upload or link) if your section requires it. Due Sunday, Dec 6, 11:59 p.m. (50 points).

---

Instructor answer key & model responses — REMOVE BEFORE PUBLISHING TO STUDENTS

Students record their own speech, so there is no single correct content. The key grades the quality of the self-assessment (is the student honest about what works and what doesn't?) and the quality of the AI-critique (does the student catch the chatbot's generic sentiment?). All persons and occasions in model answers below are fictional.

Model planning scaffold (illustrative — fictional person and occasion):
- Vivid specific detail: "Marcus stayed in the lab three extra hours the night before the demo just to triple-check the data — not because he had to, because he can't leave a thing half-done."
- What it reveals: "That's the whole person — that he gives more than is expected, always."
- Wish: "May every team he works on be lucky enough to have someone who cares that much."
- Raise: "Please raise your glass to Marcus."

Model self-assessment (illustrative):
- Occasion fit: 5 — "It's a toast at a team celebration, this is exactly the right type."
- Tone/mood match: 4 — "Mostly warm, but I rushed the 'what it reveals' beat and it sounded flat."
- Specific vivid detail: 4 — "The lab-stay detail is specific but I want to add the triple-check for even more texture."
- Organization: 5 — "All four beats present; closed with the raise."
- Brevity: 3 — "I ran to 1:55 — need to tighten."
- Delivery: 3 — "Made eye contact well, but my pace was too slow on the wish beat."
- One thing to fix: "Next take I will tighten the middle two beats so the whole thing lands under 90 seconds."

Model AI-critique (illustrative): "After my first ask the chatbot responded with: 'That's a lovely toast — it sounds warm and genuine! Consider adding more personal anecdotes.' That's hollow — 'sounds warm and genuine' tells me nothing about whether the specific detail actually lands, and 'add more personal anecdotes' is useless without knowing what details I already have. Useful feedback would have been: 'The lab-stay detail is strong; the phrase what it reveals reads as a label rather than a feeling — can you replace it with something the room can hear, like what it actually feels like to be on a team with someone who does that?'"

Expected answers:
- Part 4: (1) full credit for naming what the room would feel — not just "good" or "bad"; (2) either answer fine if reasoned; (3) must name a real tonal difference — intensity of humor, presence or absence of a raise, level of solemnity; (4) credit for a genuine second take and honest reflection.
- Part 6: full credit for a specific catch — the most common failure is the AI offering "sounds heartfelt" or "good detail" with no explanation of what actually makes the detail work or not.

Grading rubric — 50 points

Criterion	Full	Partial	None
Planning scaffold + recording (Parts 2–3) — scaffold completed with real keywords; specific detail written out; recording made and watched (15)	15	8–12	0–6
Self-assessment quality (Parts 3–4) — honest, specific, names one concrete improvement; not "it was fine" (15)	15	8–12	0–6
Rehearsal-coach engagement (Part 5) — used the coach, tried a concrete suggestion, kept AI's role in bounds (10)	10	5–8	0–4
AI-critique (Part 6) — names a specific instance of generic/hollow AI feedback and what specific feedback would say instead (10)	10	5–8	0–4

Quality gate (self-checked): the four rubric criteria sum to exactly 50 (15 + 15 + 10 + 10 = 50). ✓

This Workshop asserts no external statistics, quotations, or source citations — the material is entirely the student's own personal knowledge of a real or fictional occasion. There is nothing to fabricate or misattribute from an external source. The AI-critique moment explicitly targets the chatbot's generic-sentiment failure mode (rather than the fabricated-citation risk, which was the focus of the research weeks). The model answers use a fictional person and fictional occasion; no real person is named or quoted.