Top 10 AI Prompts and Use Cases and in the Education Industry in Durham

Durham's K–20 classrooms and campuses are already seeing practical wins from generative AI - from AI-powered study tools boosting retention at Duke and NCCU to administrators using chatbots for faster student support - and national analysis shows these tools can accelerate lesson planning, research, and student engagement while raising integrity and data-quality questions (see the EDUCAUSE review of generative AI in education: EDUCAUSE review of generative AI in education).

Local practitioners should pair small, measurable pilots with staff training: a concrete option is Nucamp's 15-week AI Essentials for Work syllabus that teaches prompt design and evaluation for real tasks, so districts can pilot tutor/chatbot use cases without overbuying vendor solutions (course syllabus and details: Nucamp AI Essentials for Work syllabus and course details).

For Durham leaders, the immediate “so what” is clear: run targeted pilots, teach prompt literacy, and measure retention and support-load improvements before scaling.

Additional local example: AI-powered study tools in Durham case study.

Program	Length	Key focus	Early bird cost
AI Essentials for Work	15 weeks	Prompt design, practical AI skills	$3,582

“Shouldn't higher education institutions be preparing graduates to work in a world where generative AI is becoming ubiquitous?”

Selection prioritized prompts and use cases that Durham educators can run quickly, safely, and measurably: each candidate was mapped to a concrete instructional or operational goal, vetted against NC State's NC State AI Guidance and Best Practices for Responsible AI Use in Education (including data-classification limits and recommendations to treat AI outputs as drafts) and checked for compatibility with NC State's Approved AI Products List for Campus‑Approved Tools so pilots avoid unnecessary IT Purchase Compliance delays; prompts that operate only on “green” (non‑sensitive) data were elevated to the Top 10 to enable faster trials.

Prompt quality criteria followed NC State's write effective prompts checklist - specificity, role framing, and iterative refinement - and were validated against regional practitioner discussions such as sessions at AIM22 Conference Sessions on Artificial Intelligence for Education.

Prioritization also weighed training needs (can staff learn prompt literacy via short PD) and measurable outcomes (retention, support-load, or grading-time reductions).

The result: a compact, risk‑aware Top 10 that aligns with state guidance, approved tools, and real classroom priorities.

Step	Primary source	Intended result
Map use cases to goals	NC State AI Guidance and Best Practices for Responsible AI Use in Education	Clear pilot metrics
Vetting data & tools	NC State Approved AI Products List for Campus Tools	Green-data prompts to avoid ITPC
Stakeholder validation	AIM22 Conference Schedule and Session Descriptions	Practitioner-aligned prompts

For Durham districts and campus ed‑tech teams looking to raise adoption without blowing the budget, Pendo's all‑in‑one SXM platform gives product analytics, in‑app Guides, and visual Session Replay so teams can spot where students and faculty stall, push targeted help, and measure impact - all from a vendor headquartered in Raleigh, NC that already supports global customers; see how Pendo ties visual replays to analytics in the Pendo Session Replay guide (Pendo Session Replay guide and analytics overview).

Pendo's Orchestrate and AI features accelerate multi‑channel onboarding (email + in‑app) and AI localization has cut guide translation time by 80% for customers like Nelnet, a concrete “so what” for Durham: translate and deploy multilingual, in‑app help far faster to reduce support loads and boost equity.

For districts worried about student data and model use, Pendo documents per‑customer model training and opt‑in AI settings that keep behavioral data scoped to a subscription, helping IT and compliance teams evaluate risk before piloting.

Fact	Detail
Headquarters	Raleigh, NC
Core features	Analytics, In‑app Guides, Session Replay, Orchestrate, Pendo AI
Notable impact	Nelnet: 80% faster guide translation (Pendo AI)
Scale	14,400+ customers; 23 trillion interactions (all-time)

“AI localization saves us SO much time. Before, it would take us 10 minutes per 1-step guide. Pendo AI helped us translate content in 2 minutes or less. Plus, we don't need additional programs to use and save .xliff files, so more PMs can translate.” - Liz Feller, Manager of Design and Analytics, Nelnet

Squirrel AI's intelligent adaptive learning platform combines smart‑learning tablets, a Large Adaptive Model (LAM) and an Intelligent Adaptive Learning System (IALS) to deliver nano‑level, personalized learning paths that adjust in real time to each pupil's gaps and strengths; for North Carolina classrooms this matters because the system reports a 25% improvement in math scores in one semester, offers 24/7 parent access for family engagement, and - for U.S. deployments - stores data within the U.S. under stated compliance practices, easing district vetting.

LAM (launched Jan 2024) raised question‑accuracy rates from 78% to 93% by leveraging exclusive data from 24 million students and 10 billion learning behaviors, making Squirrel AI a practical pilot option for PreK–5 math interventions in Durham districts.

Review adaptive‑learning market context and vendor implications in the EY white paper on AI adaptive learning and see Squirrel AI's platform overview for technical and privacy details: Squirrel AI adaptive learning platform overview and privacy details and EY white paper on AI adaptive learning market context.

Key metrics: Reported math improvement - 25% in one semester; Learning behaviors - 10 billion; Student data set - 24 million students; Learning centers - 3,000+; Nano‑level objectives - 10,000+.

Amira Learning pairs advanced speech recognition with Science of Reading pedagogy to give every early reader a 1:1, AI‑guided oral reading tutor that listens, diagnoses, and delivers micro‑interventions in real time - making it practical for Durham districts to scale individualized practice without hiring large tutoring teams; independent and state studies show measurable gains at modest dosages (Amira's research summary reports matching human tutoring after 30 sessions and ESSA Tier‑1 ratings) and state reviews (North Dakota, Utah, Texas, Louisiana) link 15–30 minutes/week of use to substantial score improvements, so the concrete “so what” for NC school leaders is this: a low‑cost pilot of 15–30 minutes/week per student can generate reliable screening data and faster tiered interventions.

Learn the aggregated evidence on Amira's efficacy at the Amira research hub and read the state efficacy summaries for details on usage and outcomes.

Metric	Reported result
Recommended student dosage	~15–30 minutes/week (peak: 2–4 short sessions/week)
North Dakota	High users grew roughly 2× as fast as non‑users
Utah effect size	~0.40–0.43 for recommended usage
Texas STAAR impact	Avg. ≈36 STAAR points per student; ~9 percentile ranks for dosage users
Evidence rating	ESSA Tier 1; matched human tutoring after ~30 sessions

“I have seen firsthand that teachers and students who use Amira with fidelity achieve remarkable reading growth, reinforcing the impact of consistent, research-based literacy instruction.”

University of Michigan's campus-owned U‑M GPT offers a concrete template for Durham-area institutions weighing an in‑house LLM: ITS and U‑M's GenAI team combined licensed commercial models with internal controls to create a “closed AI” portal that emphasizes privacy, accessibility, and affordability, supports file uploads for summarization and extraction tasks, and enforces usage limits to curb misuse - details and usage guidance are documented in the University of Michigan U‑M GPT in‑depth guide (University of Michigan U‑M GPT in‑depth guide) and the EDUCAUSE case study on campus-built generative AI explains why a campus-built approach helps protect user data while broadening access (EDUCAUSE case study: How and why the University of Michigan built closed generative AI tools).

The practical takeaway for North Carolina colleges and districts: a secured, campus‑owned gateway can preserve FERPA‑level uses and still scale - U‑M reported heavy uptake (roughly 15,000 users/day) and a mix of foundational and reasoning models to serve teaching, advising, and operational workflows, so pilot designs should budget for model selection, prompt limits, and clear review processes to ensure accuracy and equity.

Fact	Detail
Text prompt limits	≈75 prompts/hour (text models)
Image prompt limits	≈10 prompts/hour (DALL·E 3 / image models)
File upload caps	Up to 10 files/chat; PDFs: 50 pages or 10 files
Adoption snapshot	~15,000 users/day reported
Data sensitivity	Approved for “moderately sensitive”/FERPA uses

“AI will not take jobs away from you. But people who know how to use AI might.” - Ravi Pendse, U‑M VP for IT and CIO

Teacher copilots like GPT‑4o for text and DALL·E 3 for images can automate first drafts of lesson plans, handouts, differentiated activities, and classroom visuals - turning hours of prep into editable starting points - so Durham teachers can reclaim time from the oft‑cited “10 hours, 40 minutes” workday by automating routine drafting and material creation (Edutopia article on ChatGPT time savings for teachers).

Practical entry points include using an editable lesson‑plan prompt template to produce objectives, standards alignment, assessments, and accommodations, then refining outputs for local curricula and student needs (TCEA lesson-planning prompts and editable ChatGPT template).

Pair text generation with campus policies: image generators can speed visual creation but may be rate‑limited in campus deployments (see image prompt caps in campus LLM examples), so pilot scope and review workflows should be defined before classroom rollout (University of Michigan campus LLM and GPT guidance).

The concrete “so what”: use AI to cut planning time, then spend the saved minutes on targeted interventions or equity‑focused differentiation - always review and align AI outputs to standards and student data.

The British University Vietnam's Artificial Intelligence Assessment Scale (AIAS) shows how a clear, tiered policy can turn GenAI from an integrity risk into an instructional asset: BUV defined a five‑point scale (Level 1 “No AI” → Level 5 “Full AI”) so faculty can specify permitted AI uses, and after launching the AIAS academic‑misconduct cases tied to GenAI fell from over 100 to zero while pass rates rose 33.3% and mean grades climbed 5.9% - a concrete “so what” for Durham schools seeking both integrity and equity gains, since the scale also helped English‑as‑additional‑language students demonstrate understanding via multimodal AI‑enhanced submissions; read the BUV case study for details and implementation notes (BUV AI Assessment Scale case study) and compare this approach with local pilots such as Nucamp's overview of AI‑powered study tools in Durham when designing an NC‑friendly rubric that clarifies expectations, supports multilingual learners, and reduces complaint and adjudication workload.

Metric	BUV result
GenAI academic violations	Over 100 → 0 (post‑AIAS)
Pass rate change	+33.3%
Mean grade change	+5.9%
Assessment scale	5 levels (No AI → Full AI)

NC State Extension's practical playbook for safe prompting makes two things non‑negotiable for Durham educators: run pilots on campus‑approved products using your NC State account, and never paste sensitive or “red/purple” data into a public model.

The Extension guidance catalogs approved generative tools (examples include OpenAI ChatGPT Teams/Enterprise, Google Gemini, Microsoft Copilot, and Grammarly) and stresses that many approved tiers require paid accounts - free versions are not recommended because they may use university inputs to train public models - so plan budgets for campus‑sanctioned subscriptions and update data‑sharing settings if a free account is used.

Prompting best practices are equally concrete: treat AI outputs as editable drafts, ask the model to iterate and cite sources, avoid personal identifiers in custom instructions, and consult OIT's data classification before any pilot.

For implementation resources and the campus approved‑products list, see NC State Extension's AI Guidance and the NC State Approved AI Products overview.

Approved tool (examples)	Account requirement	Data allowed
OpenAI ChatGPT (Teams/Enterprise)	Use NC State account; paid tiers recommended	Green (non‑sensitive) only; red/purple needs ITPC
Google Gemini (Chat / Workspace)	Campus‑approved; follow OIT settings	Green/Yellow; sensitive use requires review
Microsoft Copilot (M365)	Campus subscription; configured by IT	Green/Yellow; higher sensitivity requires ITPC
Grammarly Pro / Education, Zoom AI Companion	Paid/education licenses preferred	Green only unless approved

“AI is a tool to augment your expertise and ingenuity, not replace them.”

Durham and other North Carolina education purchasers can accelerate safe AI adoption by borrowing the GovAI Coalition's practitioner-ready governance kit - policy templates aligned to the NIST AI Risk Management Framework, an AI Policy and Governance Handbook, an AI Incident Response Plan, and standardized vendor tools such as the AI FactSheet and Vendor Agreement - so districts avoid reinventing procurement language and force consistent vendor disclosures (vendors may register and submit AI FactSheets via the Coalition's registration process).

Because the Coalition already counts North Carolina members (City of Durham, City of Raleigh, County of Wake and the State of North Carolina Department of Information Technology among others), local teams can use the GovAI deliverables to compare real contracts in the AI Contract Hub, require FactSheets up front, and shorten legal and privacy review cycles - practical wins when pilots must protect student data and meet FERPA‑level expectations.

These resources are intended as templates and not legal advice, so pair them with district counsel before signing vendor agreements.

Resource	Purpose
GovAI templates and resources for AI policy, governance, and incident response	AI Policy, Governance Handbook, Incident Response templates (NIST‑aligned)
GovAI Coalition overview and vendor registry information	Membership, deliverables, and vendor registry information
AI FactSheet and standard Vendor Agreement templates	Standardized vendor disclosures to support procurement vetting

Durham practitioners can skip the blank‑page problem by pairing curated prompt libraries with state PD: use ready templates for lesson planning, rubrics, accessibility scaffolds, parent communications, and assessment that are already field‑tested - examples include the aiforeducation K‑12 prompt collection for lesson plans and rubrics (aiforeducation K‑12 prompt collection for lesson plans and rubrics), Panorama's classroom‑focused prompt sets and an AI Roadmap with 100+ prompts for district leaders (Panorama AI prompts for K‑12 and district AI Roadmap), and a concise educator prompt library that teaches prompt engineering basics (iLearnNH K‑12 educator prompt library and prompt engineering guide).

Pair these templates with NCDPI's living generative‑AI guidance and Wednesday webinar series - practitioners who attend live webinars (certificates available) can earn PD while testing turnkey prompts; a concrete “so what”: borrow a rubric or lesson‑plan prompt today and run a 2‑week pilot, capture time‑saved and student engagement gains, then scale with district PD and vendor vetting through statewide resources.

Source	What it offers	NC relevance
aiforeducation K‑12 prompt collection	Curated K‑12 prompts (lesson plans, rubrics, accessibility)	Ready templates for classroom pilots
Panorama AI prompts and district AI Roadmap	30+ prompts + AI Roadmap (100+ prompts)	District‑scale prompt collections and privacy-conscious product options
NCDPI generative AI resources and webinar series	Generative AI guidance, webinar series, and cohort materials	State PD, recorded webinars, and implementation templates for NC schools

Centering students means making core materials truly accessible: use AI to generate timely multilingual translations, plain‑language summaries of assignments, accurate captions and transcripts, and descriptive alt text so non‑English families and students with disabilities can engage on equal terms; run a short, measurable pilot (for example, translate back‑to‑school packets and track a two‑week change in family engagement or appointment attendance) before scaling to district systems.

Local evidence already shows practical impact - AI‑powered student study tools are boosting retention at Duke and NCCU - so pairing small translation pilots with the state's guidance and research-backed inclusive practices is a pragmatic equity play for Durham leaders (see open access social & behavioral sciences research for inclusive pedagogy and local AI partnership recommendations).

Useful starting resources: research compendium of social and behavioral sciences (research compendium of social and behavioral sciences on inclusive education), a local case summary of AI tools in Durham education (local case summary: AI‑powered student study tools in Durham schools), and guidance on building cross‑sector pilots (guide to building cross‑sector AI pilot programs in Durham).

Durham educators should sequence three practical moves now: run short, measurable pilots; invest in job‑embedded PD; and adopt clear governance before scaling.

Start with a 2‑week classroom pilot using vetted prompt templates or an accessibility translation test and capture concrete metrics - teacher planning time saved and student engagement, or for reading tutors follow Amira's recommended ~15–30 minutes/week dosage to measure learning gains - then use those results to justify subscriptions.

Pair pilots with NCDPI's living guidance and webinar PD so staff earn certificates while building AI literacy, and shorten procurement timelines by requiring vendor FactSheets and contract templates from practitioner kits.

Together, these steps turn policy into action: small pilots produce evidence, PD builds staff capacity, and GovAI‑style procurement templates protect student data while accelerating useful tools into classrooms.

Next step	Quick action	Resource
Pilot classroom use	Run a 2‑week prompt/rubric pilot; track teacher time‑saved and student engagement; for reading tools measure 15–30 min/week usage	NCDPI Generative AI Guidebook for K-12 schools
Professional development	Attend NCDPI webinars (live or on‑demand) and claim PD certificates to build prompt literacy	NCDPI AI resources and webinar series for educator professional development
Governance & procurement	Require AI FactSheets and use standardized contract templates to speed legal/privacy review	GovAI Coalition policy templates and vendor vetting resources

“Generative artificial intelligence is playing a growing and significant role in our society.”