Ch. 1 — Setting the Stage

Vocabulary Key terms & definitions

Mental Model

An internal representation of how the world works — a set of rules, beliefs, assumptions, or ways of thinking that guide how a person perceives and responds to situations. Mental models are learned, shaped by background, culture, education, religion, political affiliation, and personal experience. The word "dog" means a puppy to one person and a working guard dog to another — same word, radically different mental images. In MIS, mental models determine what gets built into information systems, because systems are designed by people who encode their own assumptions into code, fields, and workflows.

Heath & Heath, Made to Stick (2007); Gallaugher, Ch. 1; slides — "Think of the Word Dog."

Curse of Knowledge

A cognitive bias in which knowing something makes it nearly impossible to imagine not knowing it. As Chip Heath and Dan Heath write: "when we know something, it becomes hard for us to imagine not knowing it. As a result, we become lousy communicators." In MIS, this explains why an expert who has used Excel for 25+ years may be a worse Excel tutor than a TA who learned it last semester — the expert can no longer access the beginner's confusion. IT professionals cursed with deep technical knowledge routinely build systems that make perfect sense to them and none to the users they're designing for.

Heath & Heath, Made to Stick (2007); Gallaugher, Ch. 1; slides — Instapoll #3.

Information System (IS)

Any system — technical or social — that collects, processes, stores, and distributes information to support decision-making, coordination, and control in an organization. Critically, IS are defined by the five components: Hardware, Software, Data, Processes, and People. The phrase "I want to, but the computer won't let me" reflects an IS designed around someone else's mental model of how a process should work. IS are therefore never purely technical — they embed human assumptions about business processes, and are only as good as the information and processes they encode.

Gallaugher, Ch. 1; slides — "What Does That Have to Do with MIS?"

Self-Directed Learning

An approach to learning in which the student takes initiative in diagnosing their own learning needs, setting goals, finding resources, and evaluating outcomes — rather than relying on an instructor to deliver all information. In MIS, self-directed learning is essential because most real-world problems require finding trustworthy, current sources and distinguishing fact from opinion. Key skills: evaluating source credibility, recognizing when information is platform-specific (e.g., Windows Excel instructions may not work on Mac Excel), and expecting conflicting opinions on subjective questions like "which computer is best?"

Gallaugher, Ch. 1; slides — "Self-Directed Learning."

Gartner Hype Cycle

A framework developed by Gartner Inc. that models the typical maturity, adoption, and social application of specific technologies through five phases: (1) Technology Trigger — an emerging innovation generates early coverage; (2) Peak of Inflated Expectations — hype and overblown promises dominate; (3) Trough of Disillusionment — failures and reality checks cause interest to collapse; (4) Slope of Enlightenment — practical applications emerge and genuine value is demonstrated; (5) Plateau of Productivity — mainstream adoption is achieved and the technology delivers consistent value. The framework helps managers avoid both over-investing at the peak and abandoning technologies prematurely in the trough.

Gartner, Inc. — gartner.com/en/research/methodologies/gartner-hype-cycle; Gallaugher, Ch. 1.

Technology Trigger

Phase 1 of the Hype Cycle. A breakthrough, public demo, product launch, or media report generates significant interest in an emerging technology. Commercial viability is often unproven and products typically don't exist yet. Early media coverage and investor speculation begin. For Generative AI: the public release of ChatGPT in November 2022 was a classic technology trigger — millions of users interacted with a powerful AI model for the first time, generating massive media coverage and investor enthusiasm almost overnight.

Gartner Hype Cycle methodology; Gallaugher, Ch. 1.

Peak of Inflated Expectations

Phase 2 of the Hype Cycle. Frenzy of publicity generates overenthusiasm and unrealistic expectations. Early success stories (sometimes real, sometimes exaggerated) are amplified. "This will change everything!" is the dominant sentiment. A wave of speculative investment follows, often disconnected from actual user adoption or demonstrated ROI. For AI: Microsoft's $13B stake in OpenAI, Amazon's $5B pledge to Anthropic, and Nvidia's market cap growth from ~$300B to $3 trillion were all fueled by inflated expectations about AI's near-term revenue potential.

Gartner Hype Cycle; Gallaugher, Ch. 1; slides — "The Numbers…"

Trough of Disillusionment

Phase 3 of the Hype Cycle. Interest wanes as experiments and implementations fail to deliver promised results. The technology doesn't fail — expectations were simply unrealistic. Media coverage turns negative. Investors who bought at the peak lose money. "Cut investment. It's garbage!" becomes the refrain. The trough is where many genuinely valuable technologies are abandoned by impatient investors and managers who judge the technology by its peak promise rather than its plateau potential. Companies that stay committed through the trough often emerge with sustainable competitive advantages.

Gartner Hype Cycle; Gallaugher, Ch. 1.

Slope of Enlightenment

Phase 4 of the Hype Cycle. Methodologies for applying the technology productively emerge. Second- and third-generation products appear. More enterprises begin cautious pilot programs. The technology's benefits are demonstrated with real-world ROI. "Hmm, this stuff is pretty good. I'm actually seeing benefits." The key shift: use cases are specific and realistic rather than breathlessly general. Best practices and implementation frameworks develop. For AI: enterprise use cases like contract review, code generation, and customer service automation are demonstrating measurable ROI — these are slope-of-enlightenment applications.

Gartner Hype Cycle; Gallaugher, Ch. 1.

Plateau of Productivity

Phase 5 of the Hype Cycle. Mainstream adoption occurs. The technology's market applicability and relevance are clearly paying off. "Yeah, pretty much everyone that needs it uses it. Solid stuff." The technology becomes infrastructure — reliable, widely understood, and broadly deployed. Examples: cloud computing, smartphones, and relational databases have all reached this phase. They're no longer hyped; they're simply how business gets done. The goal for managers is to time entry after the trough but before the plateau — capturing competitive advantage without paying the premium of peak-hype investment.

Gartner Hype Cycle; Gallaugher, Ch. 1.

Real vs. Anticipatory Demand

Geoffrey Moore's expansion of the hype cycle distinguishes three demand drivers: Actual demand — real customers who buy and use the product, generating real revenue; Perceived demand — what firms believe the market wants, often based on analyst reports and competitor moves rather than real usage data; Anticipatory (speculative) demand — early adopters and investors betting on future value, not present utility. A bubble forms when anticipatory demand dramatically outpaces actual demand — asset prices and infrastructure capacity grow faster than real-world use. The Apollo Global Management data (2025) showing AI adoption flattening at ~14% for large firms while infrastructure investment continues to surge is a textbook anticipatory vs. actual demand gap.

Moore, G. — Crossing the Chasm (1991, rev. 2014); Apollo Global Management (2025); Gallaugher, Ch. 1.

Technology Adoption & Diffusion

The processes by which new technologies spread through markets and society over time. Adoption is whether a firm or person uses a technology at all; diffusion is how widely it spreads across a population. Investment ≠ Adoption ≠ Diffusion: a firm can invest heavily in an AI platform (investment), deploy it to 100 employees (adoption), but see only 5 actually using it regularly (diffusion). Apollo Global Management's 2025 data from the U.S. Census Bureau Business Trends and Outlook Survey (BTOS) shows large-firm AI adoption peaking around 14% and beginning to decline — significant investment has not yet translated into deep organizational diffusion.

Rogers, E. — Diffusion of Innovations (1962); Apollo Global Management (2025); U.S. Census BTOS.

Big Tech & AI Investment

Only six firms have surpassed $1 trillion in market capitalization — Apple, Microsoft, Alphabet (Google), Meta (Facebook/Instagram/WhatsApp), Amazon, and Nvidia — and all are in technology. While each reached $1 trillion by dominating different markets, they are increasingly competing with each other in AI. Microsoft invested $13B for a 49% stake in OpenAI; Amazon pledged $5B to Anthropic; Nvidia controls an estimated 80–95% of the high-end AI chip market; Facebook and Apple are paying extraordinary salaries to attract AI talent. All six are directing their largest R&D allocations toward AI — making AI not a single industry but a platform battle among the most valuable companies ever created.

Gallaugher, Ch. 1; slides — "Big Tech and AI" and "The Numbers…"; WSJ Katherine Bindley (Aug 2025).

Deep Dives Beyond the slides

The phrase "I want to, but the computer won't let me" captures a fundamental MIS truth: every information system encodes the mental models of the people who built it. Those builders had assumptions about what fields matter, what processes look like, what data should be collected, and what users would want to do. When a user's mental model of their own workflow doesn't match the developer's mental model baked into the software, the system literally cannot do what the user wants — not because of a technical limitation, but because of a human communication failure that happened months or years before the user ever touched the system.

This is why the five IS components — Hardware, Software, Data, Processes, and People — must be understood together. A technically flawless system built around incorrect assumptions about human processes and mental models will fail. The profession of MIS began with programming (hardware, software, data), but the deeper challenge has always been the last two components: processes and people. Programs encode business process logic. Business process logic comes from people. People are profoundly shaped by their mental models. Therefore IS are, at their core, expressions of human mental models — for better or worse.

The practical implication for every business manager: when you inherit or commission an information system, your first question shouldn't be "does it work?" — it should be "whose mental model does this encode, and does it match ours?" Systems that feel wrong to users aren't broken — they're just built around different assumptions about how work should happen.

The Curse of Knowledge, introduced in Heath & Heath's "Made to Stick," is one of the most reliably observed cognitive biases in organizational behavior. Once you know something deeply, you literally cannot access the mental state of not knowing it. The classic experiment: tap a song's rhythm on a table and ask people to guess the song. Tappers predict a 50% success rate; actual listeners get it right less than 3% of the time. The tapper can hear the melody in their head — listeners hear random taps. The tapper is cursed.

In MIS, this plays out constantly. An IT professional who has been building databases for 15 years cannot fully imagine what it's like to encounter a relational database concept for the first time. They design admin interfaces for themselves, not for the accounting manager who will use them once a quarter. They write documentation in jargon that assumes the reader already understands 90% of the content. They build workflows that make logical sense given their mental model of the process, but feel alien to users who have never thought about data normalization.

The professor-vs.-TA example from the slides is precise: the professor who has used Excel for 25+ years has completely forgotten the experience of not knowing Excel. The TA who learned it six months ago still has access to that confusion — they can explain it from the inside. This is not about intelligence; it's about the irreversibility of expertise.

Counter-strategy: user testing with actual novice users (not colleagues who "aren't very technical"), writing documentation for someone with zero prior knowledge, and deliberately partnering IT designers with the business users who will use the system — not just at the requirements phase, but continuously throughout development.

Gartner's Hype Cycle is not a prediction — it's a pattern-matching tool. Technologies don't follow it mechanically, but the pattern of hype → disillusionment → productivity is remarkably consistent across decades. The internet hit its Peak of Inflated Expectations in 1999–2000 (Pets.com, $50B valuations for companies with no revenue), crashed into the Trough of Disillusionment in 2001–2003, and reached the Plateau of Productivity with e-commerce, SaaS, and cloud services throughout the 2010s. 3D printing, blockchain, the metaverse, and autonomous vehicles have all run similar cycles.

Generative AI, as of 2025, shows strong signs of being at or near the Peak of Inflated Expectations. Indicators: massive, speculative corporate investment disconnected from demonstrated ROI; AI job listings and salaries surging to levels that imply universal AI transformation of all work; Apollo Global Management data showing large-firm AI adoption peaking at ~14% and beginning to decline even as infrastructure investment surges; and a flood of AI startups with business models that amount to "resell OpenAI's API with a prettier interface." The professor's prediction — expect a trough of disillusionment — is consistent with these signals.

The most important insight for a business student: the Hype Cycle doesn't mean the technology is worthless. It means the timeline and scope of impact were overstated. The internet didn't "change everything" in 2000 — it changed everything by 2015. Generative AI will likely follow a similar arc: genuinely transformative, but on a longer and more uneven timeline than the hype suggests. The profitable move is usually to watch the hype peak, survive the trough, and invest during the slope of enlightenment when valuations are rational and use cases are proven.

One of the most important analytical distinctions in technology strategy is between investment, adoption, and diffusion — three measures that are often conflated but tell very different stories. A company can invest $50M in an AI platform (investment), deploy it to 5,000 employees who have accounts (adoption), and discover that only 200 people use it more than once a week (diffusion). The investment and adoption numbers look impressive in a press release; the diffusion number is what actually determines business impact.

Apollo Global Management's 2025 analysis of U.S. Census Bureau Business Trends and Outlook Survey data found that AI adoption rates for large firms (250+ employees) peaked around 14% and began declining — even as AI investment by those same firms continued to surge. This is a classic anticipatory vs. actual demand gap: firms are investing based on what they believe the market will require (anticipatory demand) before actual usage by their own employees has demonstrated clear value (actual demand).

Jamie Dimon's perspective on AI (referenced in the slides) represents the measured view of a major institutional leader: AI is genuinely important and JP Morgan Chase is investing in it seriously, but the transformation will be gradual and sector-specific, not universal and immediate. This is the "slope of enlightenment" mental model — the technology is real, the value is real, but the timeline requires patience and specificity rather than general enthusiasm.

For any technology investment decision, insist on measuring all three: investment (dollars committed), adoption (users with access), and diffusion (users who actually use it regularly and report value). The gap between adoption and diffusion — sometimes called the "adoption chasm" — is where most enterprise technology initiatives fail quietly, long after the announcement press release has been forgotten.

Hype Cycle Stages Reference table · all five phases

Phase	Dominant Sentiment	What's Actually Happening	Gen AI Example
1. Technology Trigger	"This will change everything!"	Breakthrough generates early media coverage; commercial viability unproven; products often don't exist yet	ChatGPT public release, Nov 2022; demos of image generation, code writing, essay drafting
2. Peak of Inflated Expectations	"Uh, where's the ROI?" (investors) / "This will replace all jobs" (media)	Frenzy of publicity; overenthusiasm; speculative investment surges; early success stories amplified	$13B Microsoft/OpenAI; $5B Amazon/Anthropic; Nvidia to $3T market cap; AI job salaries surging to $1M+
3. Trough of Disillusionment	"Cut investment. It's garbage!"	Experiments fail; media turns negative; interest wanes; producers fail or exit; survivors improve product	AI adoption flattening at ~14% for large firms (Apollo, 2025); many AI startups failing; ROI elusive at scale
4. Slope of Enlightenment	"Hmm, this is actually pretty good. I'm seeing benefits."	Practical applications emerge; best practices develop; second-gen products appear; cautious enterprise pilots	AI for contract review, customer service automation, code generation — specific, measurable, proven use cases
5. Plateau of Productivity	"Yeah, pretty much everyone that needs it uses it. Solid stuff."	Mainstream adoption; market applicability clearly understood; technology becomes infrastructure	Not yet reached — likely 5–10 years away for GenAI; cloud computing is a current example of this phase

Source: Gartner, Inc. — gartner.com/en/research/methodologies/gartner-hype-cycle; applied examples from Gallaugher Ch. 1 slides and Apollo Global Management (2025).

The Key Managerial Insight: The Hype Cycle doesn't predict whether a technology will succeed — it predicts the emotional arc surrounding it. Technologies that reach the Plateau of Productivity are genuinely valuable. The question is whether your organization invests at the speculative peak (paying a hype premium and absorbing the disappointment) or at the slope of enlightenment (paying a rational price for a proven technology). Patience and skepticism during the peak are competitive advantages.

AI Adoption & Big Tech Investment vs. reality

Firm	AI Strategy	Key Investment / Signal	Hype Cycle Position
Microsoft	Embed AI into every product (Copilot in Office, Teams, Azure); own the enterprise AI stack	$13B for 49% stake in OpenAI; Copilot integrated across Microsoft 365	Riding the peak; early enterprise Copilot adoption data mixed on ROI
Amazon / AWS	Provide AI infrastructure; build proprietary models (Titan); invest in frontier labs	$5B pledge to Anthropic; AWS Bedrock AI platform; Trainium / Inferentia AI chips	Infrastructure bet — positioned to benefit at any cycle stage if AI grows
Nvidia	Dominant AI chip supplier (H100, Blackwell GPUs); 80–95% high-end AI chip market share	Market cap: ~$300B (2020) → $3T+ (2024); entire revenue tied to AI infrastructure demand	Most exposed to hype cycle — if enterprise AI adoption stalls, GPU demand craters
Alphabet (Google)	Defend search dominance; build Gemini; operate TPU AI chips; DeepMind research	Gemini integrated into Search, Workspace; TPU custom AI ASICs; Waymo autonomous vehicles	Defensive investment — protecting core search revenue while building new AI products
Meta	Open-source AI (Llama models); AI for ad targeting; AI-generated content on platforms	Paying "ludicrously high" salaries for AI talent; Llama models released free to developers	Unique bet: open-sourcing creates an ecosystem Meta can monetize through its platforms
Apple	On-device AI ("Apple Intelligence"); privacy-first; Neural Engine hardware since 2017	Paying extraordinary salaries for AI talent; Apple Intelligence in iOS 18; partnership with OpenAI	Most conservative — betting on local AI that doesn't require cloud, protecting privacy brand

Apollo Global Management Finding (2025): Based on U.S. Census Bureau Business Trends and Outlook Survey data, AI adoption rates for large firms (250+ employees) peaked at approximately 14% and began declining — even as investment continued to surge. This is the textbook definition of anticipatory demand outpacing actual demand. Investment ≠ Adoption ≠ Diffusion. A trough of disillusionment may follow before a more sustainable slope of enlightenment takes hold.

The 5 IS Components Why MIS requires more than technical knowledge

Component	What It Is	Mental Model Connection	Example Failure
Hardware	Physical devices: computers, servers, phones, sensors, network equipment	Hardware choices encode assumptions about who uses systems and how — desktop-first design excludes mobile workers	Building a system on desktop hardware when all users are in the field with phones
Software	Programs and applications: OS, databases, enterprise apps, custom code	Every line of code is a decision made by someone with a particular mental model of the business process	Excel treating 0 as FALSE and 1 as TRUE — a mental model from 20th-century programmers baked into every spreadsheet
Data	Raw facts, figures, records: what gets stored, in what format, with what labels	What data gets collected reflects whose information needs were considered — often not the end user's	A CRM that captures "first name / last name" fails for cultures with different name structures
Processes	The workflows, procedures, and business rules the system automates or supports	Processes come from people — they reflect how someone thinks work should happen, often not how it actually happens	"I want to, but the computer won't let me" — system enforces the designer's process, not the user's reality
People	Users, administrators, designers, managers — everyone who interacts with the IS	People are where mental models live; the Curse of Knowledge means experts build for experts, not for users	Professor (25 yrs Excel) worse at explaining Excel than TA (6 months Excel) — curse of knowledge in action

The Central Insight of Chapter 1: MIS is not primarily about technology — it is about the intersection of technology and human communication. Information systems are only as good as the information they transmit and the processes they encode. Both are created by people with mental models. Understanding mental models, recognizing the Curse of Knowledge, and appreciating how these biases get baked into systems is the foundational skill for anyone who will commission, manage, or use information systems in a business career.

Setting theStage

Mental Models, IS Design & "I Want To, But the Computer Won't Let Me"

The Curse of Knowledge: Why Experts Make Bad Teachers and Bad Designers

The Hype Cycle Applied: Where Is Generative AI Right Now?

Investment ≠ Adoption ≠ Diffusion: The AI Measurement Problem