Why not just have students do this with ChatGPT?

Because ChatGPT will not tell them what happened next. The simulation runs a consistent engine across fifty topics with authored consequences, so the same strategy always produces the same result and a different one produces a traceably different one. That is what makes judgement gradable.

Why not use the real tools like Semrush, a live AI visibility dashboard?

Use both. The real tools show what is; the simulation shows what if. Students cannot spend $300,000 of a real brand's money to find out whether their allocation was right.

Do students need to have done the reading?

No. Ela, the in-sim copilot, explains any metric on request and cites the source, so students can start cold. Those who have read the book go deeper, faster.

Can students game it with an AI assistant?

The graded artefact is the positioning brief and the allocation memo, marked on reasoning against a four-trait rubric, and decision windows are time-boxed. An AI-written brief that is generic fails the rubric on specificity and differentiation by design.

How long does it take in class?

Roughly eight hours of student time across two rounds, run over two to three weeks. It also works as a single three-hour session using round one only.

What do I actually mark?

The positioning brief (rubric supplied), the Topic Defense, the Allocation Memo, and the quiz items. Results are shown to students after marking so the grade reflects the reasoning, not the number.

Does it work with Canvas, Brightspace or Blackboard?

Students access the simulation through a browser via a cohort link; grades move to your gradebook via export today, with LTI grade passback on the roadmap. Ask us for the Canvas/Brightspace note.

New for Fall 2026

AI Search Simulation

Students take over search marketing at Beckwith & Co., a sixty-year-old optician chain with forty stores and a loyal following with almost no presence in AI search. Ask ChatGPT where to buy glasses online and it names competitors; Beckwith does not appear. Across two rounds and $300,000, students diagnose the gap, split budget across SEO, SEM, GEO and GEM, build the pages and PR pitches, and analyze what the AI says about the brand afterwards.

Screenshot 2026-09-15 at 11.45_edited.pn

Book a Demo

Request Instructor Access

A three-hour session · A four-week module · The full fifteen-week course · Slides, readings, quiz items and assignments personalised to your syllabus ·

Trusted By Leading Global Institutions for Undergrad, MSc, MBA, and Exec Ed Courses

Students learn to earn a ranking.
Almost none learn to earn a citation.

Most digital marketing courses still teach search (SEO/SEM) as a keyword problem: rank for the term, buy the click. That covers Google. It does not cover the answer engines a growing share of customers now ask first — where a brand is either named in the answer or absent from it, and where the brand's own content is only one of the sources the AI checks. This simulation gives students a brand that is strong offline and invisible in AI search, and asks them to close that gap in a way that moves sales, not just visibility.

What do students do?

01 Round one: the launch. $100,000.

Beckwith is taking its first collection online. Students pick which of four customer personas to target, choose three to seven topics to fight for, allocate the budget across SEO and GEO and build the pages and PR pitches. The trap: chasing the high-volume, cheap-glasses topics the brand can never truthfully own.

02 Round two: the paid budget. $200,000.

Round 1's pages and pitches are live. Students keep working the organic side, and now have $200,000 of paid budget on top, split as they choose between Google Ads and ChatGPT Ads. They can buy visibility on the topics organic hasn't reached, or bid on the ones they are already winning. The trap: paying for clicks they were getting for free.

03 The debrief.

Students read a results dashboard that shows what the AI said about the brand before and after, which prompts now cite it, and how mentions and citations laddered up to traffic and sales. Ela, the in-sim research copilot, walks them through it. Students can and should run it more than once — the engine is deterministic, so a different strategy produces a different, traceable result.

What each stage looks like.

The teaching materials come ready to use.

FOR THE PROFESSOR

Start from the teaching pack.

The Instructor Guide, a "Building Your Course" guide, seven chapter decks plus a master deck, the drop-in syllabus block with seven outcomes and a grade weight, and a 105-question quiz bank. A recorded walkthrough covers the mechanics so you can spend your prep on the debrief, not on learning the screens.

FOR THE STUDENT

Arrive knowing the brand and the decision.

A one-page briefing note sets up Beckwith, the four personas and the two surfaces they will be competing on. Inside the simulation, Ela explains any metric on request and cites where the answer came from — so students who have not done the pre-reading can still start, and students who have can go deeper.

Two rounds of the same four decisions:

choose the topics, fund them, build the pages, then read the results.

FOR THE PROFESSOR

For the professor Watch reasoning become visible before results exist.

Browser-only, no installation needed, cohort link and leaderboard provided. On the Creative screen every student writes a positioning brief, and every downstream page and pitch is generated from it. That brief is the artefact you mark — on reasoning, before the results are revealed. Time-boxed decision windows keep the work in the room.

FOR THE STUDENT

Make real trade-offs with real jeopardy.

Only targeted topics can be funded; only funded topics get a page; only those topics move at Results. Keyword-dense copy that helps Google can give the AI nothing to cite; a fast page that converts well may earn no mention at all. Paid buys a position while it lasts and adds nothing to share of voice. Nothing in the interface tells them what to do.

Every outcome traces back to a specific decision.

FOR THE PROFESSOR

Grade the judgement, not the score.

The four-trait rubric — specificity, groundedness, differentiation, coherence — marks the brief at trait level, so feedback tells students which of four things to fix. Two major assignments (Topic Defense, 800 words; Allocation Memo, 1,200 words) extend the sim into graded coursework. Grades go into your LMS gradebook or via the class link.

FOR THE STUDENT

See what the AI says now — and why.

At the end of a round: what the AI said about the brand before, and what it says now. Which customer questions now name Beckwith, and how much of that presence never produced a click. Ela traces each result back to a decision, and a card for each persona shows how that buyer responded. A PDF captures the round. Then run it again, differently.

What does the AI Search simulation teach?

Diagnose a brand's position in search and AI search from real platform data — rank, keyword difficulty, share of voice, cited share and explain what each number means.

Distinguish mentions from citations from recommendations, and know which spend moves each.

Allocate a fixed budget across SEO, SEM, GEO and GEM, and defend the tilt between organic and paid, search and AI.

Write a positioning brief that is specific, grounded, differentiating and coherent — and survives corroboration against outside sources.

Resolve the on-page tension between what ranks on Google and what an answer engine will cite.

Hold the purchase funnel as the point of the work: perception is the lever, sales are the goal.

How is it assessed?

Grades land in your LMS gradebook or through the class link. The pack includes a four-trait marking rubric for the positioning brief (graded on reasoning, not result), two major written assignments with exhibits, a 105-question quiz bank organised by chapter, and grading language you can lift into your syllabus — including a permitted AI-use line. You see every student's brief, allocation and results per round, plus the class leaderboard.

"I was thrilled when I discovered Novela's marketing simulations! They are perfect for my Digital Marketing course, bringing theory to life with real data, authentic interfaces, and AI-powered feedback."

David Rogers

Columbia Business School

Three ways to bring it into your course

Drop it into an existing course

Two to three weeks: replace the SEO unit with the simulation and Chapters 1–4 of the companion text.

Take the four-
week module

The Integrated Search Marketing module: book, decks, quiz items, both assignments, and the simulation.

Teach the whole fifteen weeks

The full AI marketing syllabus, with the first class taught by Clark Boyd free.

The companion textbook

Integrated Search Marketing: SEO, SEM, GEO, and GEM

in the Age of AI

Dr. Michael Pettiette and Clark Boyd. Foreword by Jim Lecinski.

Seven chapters, 108 pages, live on Amazon in Kindle and paperback. The simulation puts the book's ISM 2×2 directly on the Budget screen. Adopting professors receive the PDF free — even if they don't adopt the simulation. Neither is required to run the other.

About the book →

FAQs

Similar Simulations

Google Ads Simulation

Students allocate budgets, manage bids, and balance efficiency with growth — learning how optimisation decisions create second-order effects. Based on Burberry.

See the simulation →

See the simulation