AFM 241 · Exam Cheat Sheet
Emerging Tech / Business IT · Weeks 1–7 · the frameworks + the traps, distilled
90 multiple-choicelectures + readings + tutorialsBMC NOT tested
Decode every question →
value vs hype
sustaining vs disruptive
amplify vs abdicate
marginal vs full cost
1 · DisruptionChristensen · the biggest block
The Innovator's Dilemma: good companies get beaten by cheaper, "worse" products because the same rational financial logic that wins their best customers blinds them to disruption from below.
Why incumbents fail (the chain): a "why did they fail?" question wants a link here, not "bad management." RPV = Resources (what you have) · Processes (how you work) · Values (what you'll fund, the part that rejects low-margin disruptors).
Classify any disruption, ask WHO it targets (not how good it is):
↑ UP
Sustaining
Better product for your top, best-paying customers. The norm, fine.
"cheaper-better-faster" vs rivals
↓ DOWN
Low-end
"Good enough" for over-served existing customers, cheaper model.
Walmart · Netflix · phone cameras
↗ OUT
New-market
Serves non-consumers who couldn't buy before (no $/skill/access).
Sony radio · Southwest vs the car
400% vs 24.6%
Marginal vs full cost (US Steel). Reusing the old plant shows a fake 400% ROI (marginal cost $50/ton); the new capability's real basis is 24.6% (full cost $270/ton). Trap: the 400% answer. DCF terminal value = CF ÷ (r−g).
2 · Gartner Hype Cyclejudging & timing new tech
Expectations spike on hype, crash, then climb on real value. Tested: the ORDER, the Trough never precedes the Peak. In the Trough, genuine-fit adopters push on; hype-chasers bail. (6 = Beyond → Swamp, then Cliff.)
1
Trigger
buzz, few users
2
Peak
"silver bullet" hype
3
Trough
reality; shakeout
4
Slope
real uses emerge
5
Plateau
mainstream, boring
6
Beyond
Swamp → Cliff
3 · Don't get trappedthe highest-value content, concept → ✗ wrong pick → ✓ truth
Concept
✗ Wrong answer
✓ Actually true
Low-end vs new-market
Walmart = new-market
Low-end (over-served existing customers)
Uber
Uber disrupted taxis
Share grab via regulatory arbitrage, not disruptive
Sustaining vs disruptive
A breakthrough/startup = disruptive
Disruptive = simple, accessible, affordable; a process
Marginal vs full cost
Reuse old plant: 400% ROI
Cost a new capability at full cost (24.6%)
Parmenides' Fallacy
"Do nothing" = stable status quo
The do-nothing path is a non-linear decline
Hype Cycle order
Trough before the Peak
Impossible, hype (Peak) precedes the let-down
Swamp vs Cliff
Still used daily but costly = Cliff
That's the Swamp; Cliff = discontinued, resale-only
Gartner vs Forrester
Magic Quadrant = Offering × Strategy
That's Forrester's Wave; Gartner = Execute × Vision
Phoenix vs rebrand
Renaming/patent tweak = Phoenix
Phoenix needs a genuinely new capability
SOC 2 Type 1 vs 2
Type 1 is tested over time
Type 1 = point-in-time design; Type 2 = over time
ML vs GenAI
GenAI's goal is prediction
GenAI = generation; ML is the predictor
RPA vs Claude
RPA automates analysis/judgment
That's Claude; RPA = structured, rule-based steps
Amplify vs abdicate
Trust the polished AI output
Unverified reliance = abdication; verify & own it
Org Type B
Type B's risk = adopting too late
Type B's danger = adopting too EARLY (lured by hype)
4 · AI Toolkitspell-outs + what each actually is
ML vs GenAI
ML (Machine Learning): learns from data to predict. Goal = predictive accuracy.
GenAI (Generative AI): LLMs on the transformer (2017); creates output. Goal = generation.
RPA vs Claude
RPA (Robotic Process Automation): rule-based, structured tasks; fixed cost; devs build.
Claude: augments cognitive work; variable cost; anyone builds.
6 Pillars of LLM security
SOC 2 (audit of vendor controls; Type 1 = design / Type 2 = over time) · permissioning · refuse prompt-injection · admin oversight · zero data retention · audit logging.
The 4Ds of AI Fluency
Delegation (what to hand AI) · Description (the goal, prompting) · Discernment (trust it? = skepticism) · Diligence (own it = you sign the file).
5 · Numbers & namesif a figure/name appears, it's usually one of these
400 / 24.6%
US Steel ROI, old vs new plant
2017
Transformer (powers GenAI)
~200
Skill description chars (what + when)
$69M
Beeple NFT; market $67M→$14B→crash
$3.5B / $11.4B
DCF terminal value CF÷(r−g)
Logic Theorist
first true AI (1955–56)
A · B · C
org types: aggressive / majority / conservative
5 moats
Momentum · Tech · Ecosystem · New-tech · Model
AFM 241 · figures per course slides · pair with the 13-page reference for depth
Trough never precedes Peak · 400% is the trap · Walmart = low-end