Chapter 15 · Real-World Applications

Agents are already
at work.

Everything you've learned — perception, planning, tools, memory, guardrails — isn't a lab exercise. Agentic AI is in production across industries right now.

The floor tour

Five industries, one pattern.

We'll walk five aisles — enterprise, software, healthcare, finance, education — and meet the agents actually running in each. Watch how different the domains look, and how similar the agents underneath really are.

🏢 Enterprise

Research, ops, comms.

💻 Software Dev

Write, review, run.

🩺 Healthcare

Research, coding, care.

💵 Finance

Research, risk, rules.

🎓 Education

Tutor, design, assist.

Same core loop underneath — pointed at a different workflow in each aisle.

Aisle 1 · Enterprise

The back office went first.

Anywhere there was reading, summarizing, or routing to do, agents landed early. Six on the floor:

◆Competitor research

Scans the market, gathers what rivals are doing, hands back a briefing.

◆Report summarizer

Digests long reports and decks into the few points a busy reader needs.

◆Doc data-extraction

Pulls structured fields — dates, totals, parties — out of messy PDFs and forms.

◆Email triage

Reads the inbox, categorizes each message, drafts a reply, sends the routine ones.

◆Meeting intelligence

Transcribes the call, pulls out action items, follows up so nothing slips.

◆Sales automation

Qualifies leads, personalizes the outreach, keeps the records up to date.

Aisle 2 · Software Dev

Engineering was an early adopter.

For a simple reason: code can be run, and a result that runs is a result you can trust.

◆Autonomous coding

Writes the feature, adds the tests, debugs failures, refactors — end to end.

◆Code review

Reads every change for bugs, security holes, and style before a human looks.

◆DevOps monitoring

Watches logs and metrics, spots the anomaly, trips the alert before users notice.

A runnable result is a built-in feedback loop — the agent grades its own work.

Aisle 3 · Healthcare

Agents work alongside clinicians.

Taking the reading, coding, and coordinating off their plates — so people can focus on people.

◆Clinical-literature research

Searches the medical literature and summarizes the findings that bear on a case.

◆Medical coding

Reads the clinical notes and assigns the correct billing codes automatically.

◆Patient comms & scheduling

Handles scheduling and answers the common questions patients ask all day.

Aisle 4 · Finance

Reading dense documents, watching for risk.

Both are things an agent can do continuously, at scale, without tiring.

◆Earnings research

Reads earnings reports and models out how different scenarios could play.

◆Risk assessment

Monitors portfolios and surfaces exposure the moment conditions shift.

◆Compliance flagging

Scans activity against the rulebook and flags anything that crosses a policy line.

The edge here is continuous — the agent never looks away from the market.

Aisle 5 · Education

This is where agents get personal.

Meeting each learner where they are — instead of teaching to the middle of the room.

◆Adaptive tutoring

Adapts to each student's level and pace, slowing down or pushing ahead as needed.

◆Curriculum design

Builds lesson plans and assessments aligned to what the class needs to learn.

◆Research assistants

Finds sources, summarizes papers, checks facts so students go deeper faster.

The through-line

Why any of this works.

Strip away the industries and every win comes down to the same three moves:

Compresses

Work that took hours or days — reading, drafting, cross-checking — collapses into minutes. It removes the waiting between steps.

Parallelizes

Tasks that needed a whole team run side by side. One agent per lead, per document, per ticket — work fans out instead of queuing.

Never stops

No nights, weekends, or fatigue. An agent triaging inboxes is as sharp at 3 a.m. as at noon — 24/7, without drift.

Compresses

Hours and days → minutes.

The agent doesn't get faster at one step — it removes the waiting between steps, so a day of work collapses into a coffee break.

By hand hours – days With an agent minutes ▼ same job, compressed

Never stops

On the clock at 3 a.m.

No breaks, no fatigue, no drift. An agent watching logs or a portfolio is exactly as sharp overnight as at noon.

24/7 No nights off No weekends No fatigue always on

The pattern that repeats everywhere

Strip the industry — it's the same creature.

Every agent you met today is one thing wearing five costumes. It:

👁️Perceives

Reads its situation — the inbox, the codebase, the chart, the clinical note.

🗺️Plans

Maps a path toward the goal and re-plans when reality shifts.

⚡Uses tools

Acts on the world — searches, runs code, sends the message, updates the record.

🔄Adapts

Notices what went wrong and course-corrects — pointed at a real workflow.

Perceive · plan · use tools · adapt.

Chapter 15 in one breath

The future isn't coming. It's already clocked in.

Across enterprise, software, healthcare, finance, and education, agents do real work today — researching, drafting, reviewing, coding, monitoring, tutoring. They win the same three ways everywhere: they compress hours into minutes, parallelize what teams used to split, and run around the clock.

Same loop underneath; a real workflow on top.

Next up

Chapter 16 · Your Learning Path.

You've seen the agents at work and the pattern under all of them. Next: how you start building — the path from watching the floor to running your own.

Press S for speaker notes · F for fullscreen · → to advance.