What is this course about?
Per-seat licensing is structurally wrong for a campus: the population is seasonal, bursty, and mostly idle, so you pay for tens of thousands of seats that never log in. This course builds the three-scenario model — per-seat, usage-based, and owned — on the learner's own numbers, and includes the costs institutions consistently forget: integration, change management, and exit.
Who is this course for?
- CFOs and budget officers
- CIOs and IT finance staff
- Procurement officers
- Provosts building an AI business case
What do I need before starting?
- Bring your institution's headcount figures and any current AI quotes
- Comfort with a spreadsheet
What will I be able to do afterwards?
- Model per-seat, usage-based, and owned scenarios on your own headcount
- Explain why campus utilization patterns break per-seat economics specifically
- Estimate token cost from workload rather than headcount
- Compute a defensible self-hosting break-even including operations
- Present a three-year TCO to a CFO who has seen failed pilots
What does each module cover?
What are the three AI pricing shapes?
35 minPer-seat, usage-based, and owned — and the assumption each one hides about your institution.
Objectives
- Describe each pricing shape and its embedded assumptions
- Identify which shape a given quote actually is
- Recognize hybrid pricing designed to obscure comparison
Topics
Activity. Classify three real vendor quotes by pricing shape and name each one's hidden assumption.
Why does per-seat break specifically on a campus?
45 minSeasonality, bursts around registration and finals, and a population that is overwhelmingly idle.
Objectives
- Characterize campus utilization patterns against a per-seat model
- Compute the effective cost per active user
- Explain the structural mismatch to a non-technical audience
Topics
Activity. Compute your effective cost per active user under a per-seat quote using real login data.
How do you build the headcount-multiplied bill?
45 minThe full population — students, faculty, staff, and alumni — and what each per-seat tier costs across it.
Objectives
- Enumerate every population a licence would need to cover
- Build the multiplied bill across tiers
- Model growth and seasonal variation
Topics
Activity. Build the headcount-multiplied bill for your institution across three vendor tiers.
What does an advising conversation actually cost in tokens?
50 minBottom-up cost modeling from real workloads instead of top-down from headcount.
Objectives
- Estimate token consumption for representative workloads
- Model cost across conversation volume rather than user count
- Account for retrieval and context overhead
Topics
Activity. Measure token consumption for three real workloads and extrapolate to annual volume.
When does self-hosting actually break even?
50 minGPU amortization, utilization, and the operations burden — modeled honestly rather than optimistically.
Objectives
- Model hardware capital and amortization
- Include realistic operations staffing
- Compute the break-even volume against usage pricing
Topics
Activity. Compute your break-even volume and state the assumptions it depends on.
Which costs does nobody budget?
40 minIntegration, change management, evaluation, and exit — the line items that decide whether the project succeeds.
Objectives
- Estimate integration and change management realistically
- Budget for ongoing evaluation and monitoring
- Quantify exit cost and its effect on negotiating position
Topics
Activity. Add the four forgotten line items to your model and see how the comparison shifts.
How do you present this to a skeptical CFO?
45 minBuilding a business case for someone who has already funded pilots that went nowhere.
Objectives
- Structure the case around risk rather than enthusiasm
- Present sensitivity rather than a single number
- Pre-answer the objections a finance officer will raise
Topics
Activity. Present your model to the cohort playing a hostile finance committee.
Building your three-scenario model
50 minThe workshop module: a complete model on the learner's own institutional numbers.
Objectives
- Complete all three scenarios with real figures
- Run sensitivity analysis on the decisive assumptions
- Produce a one-page summary for a decision meeting
Topics
Activity. Finish the model and produce the one-page summary.
What is the capstone project?
Three-year AI TCO model for your institution
Build a complete three-scenario cost model on your institution's real headcount and workload figures, including the forgotten line items and a sensitivity analysis, with a one-page summary for a cabinet decision.
Deliverable: A working model plus a one-page decision summary.
How are learners assessed?
- Model reviewed for arithmetic correctness and stated assumptions
- Sensitivity analysis must identify the genuinely decisive variables
- Hostile-committee presentation assessed on objection handling
What ships with the course?
Facilitator guide
Session-by-session running order, discussion prompts, and the questions that reliably derail a room.
Learner workbook
Exercises, checklists, and the templates each module's activity produces.
Hands-on lab environment
A sandboxed ibl.ai deployment so exercises run against real agents, not screenshots.
Assessment bank
Scenario questions and rubric criteria mapped to each stated learning outcome.
Source bibliography
Every primary regulation and standard cited on this page, linked and dated.
Which AI agents does this course use?
The hands-on modules run against agents already deployable on the ibl.ai platform for higher education.
Where does the course material come from?
Every module is grounded in primary sources — the regulation, standard, or research itself, not a summary of it. Each was resolved at authoring time.
- AI Index Report
Stanford HAI
Inference cost and compute trend data underpinning the token modeling.
- Ideas Made to Matter
MIT Sloan
Research on enterprise AI adoption economics and pilot failure rates.
- Transformers documentation
Hugging Face
Technical reference for self-hosting requirements in Module 5.
- AI Risk Management Framework
NIST
Frames the evaluation and monitoring cost line in Module 6.
Delivery notes
Binding guidance for anyone preparing and delivering this course.
- Ship the spreadsheet as a real artifact with working formulas, not a screenshot. It is the most-reused deliverable in the whole higher-ed catalog.
- Module 5 must be honest about when self-hosting loses. A course that concludes self-hosting always wins is a sales deck, and finance officers will detect it immediately and discount everything else.
- Use current published list prices and date them visibly. Pricing moves fast; an undated figure makes the whole model look stale within a year.
- The hostile-committee role-play needs someone who genuinely knows institutional finance. A sympathetic facilitator produces a useless rehearsal.
- Cross-link to the ibl.ai cost calculators rather than duplicating them, and keep the per-seat framing consistent with the rest of the corpus.
Why run AI training on a platform you own?
You own the course, not a licence to it
Course content, learner data, and the platform run inside your perimeter — you own all the code and the data.
Model-agnostic delivery
Run the course's AI components on any LLM — Claude, GPT, Llama, Gemini, Command — and switch anytime.
No per-seat training licences
Usage-based or self-hosted, so cost tracks actual use rather than headcount.
Deploy anywhere
Cloud, private VPC, on-premise, or fully air-gapped — including for cohorts that cannot use public AI tools.
Frequently asked questions
What does the The Campus AI Cost Model: Per-Seat vs Owned Infrastructure course cover?
Per-seat licensing is structurally wrong for a campus: the population is seasonal, bursty, and mostly idle, so you pay for tens of thousands of seats that never log in. This course builds the three-scenario model — per-seat, usage-based, and owned — on the learner's own numbers, and includes the costs institutions consistently forget: integration, change management, and exit. It runs 5 hours across 8 modules across 8 modules, at foundational level, and closes with a capstone: Three-year AI TCO model for your institution.
Who should take The Campus AI Cost Model: Per-Seat vs Owned Infrastructure?
It is written for CFOs and budget officers, CIOs and IT finance staff, Procurement officers, Provosts building an AI business case. Prerequisites: Bring your institution's headcount figures and any current AI quotes; Comfort with a spreadsheet.
Can we run this course on our own infrastructure?
Yes. ibl.ai is model-agnostic and deploy-anywhere — cloud, private VPC, on-premise, or fully air-gapped — and you own all the code and the data. Cohort data, submissions, and any material learners upload stay inside your perimeter, which matters for higher education teams that cannot send work to a public AI tool.
How do we get access to The Campus AI Cost Model: Per-Seat vs Owned Infrastructure?
Request access and we will set it up for your cohort — hosted by ibl.ai, or running against your own deployment. Tell us the group size and timing you need, and whether it should run inside your own perimeter.
How much does AI training for higher education cost on ibl.ai?
There is no per-seat pricing — you pay for usage or self-host and pay only for the infrastructure, so a 5,000-person rollout does not cost 5,000 licences. 1.6M+ users across 400+ organizations run the platform this way, including NVIDIA, MIT, and Syracuse University.