5 min read

Why AI Won't Speed Up Your Scrum Team (The 1999 Trap)

AI speeds up code, but review, QA and release still run at human speed. The bottleneck moves and the Definition of Done breaks first. Part 1 of 3.

Part 1 of 3. Part 2 is Code Got Cheap. Accountability Didn’t. Part 3 is Why AI Is Breaking Your Definition of Done.

Short answers

AI tools don’t remove delivery bottlenecks. They move them. When a team uses AI to write code faster, the extra code lands in peer review, the test suite and Sprint Backlog refinement, and those still run at human speed. Work in progress piles up, the Definition of Done collapses, and the team ends up inspecting less, not more.

I started writing code in 1991 and lived through two startups in the late 1990s, one IPO and one dot-bomb crash. The video uses that “1999 trap” as the pattern: speculative promises first, delivery economics later.

The Scrum link is the three empirical pillars of Scrum Theory: Transparency, Inspection and Adaptation. If your team can’t inspect the code or the requirement with their own eyes, it can’t adapt. The video also separates AI theater (200 backlog tickets nobody asked for) from boring utility (finding the discrepancy across five vendor specs in minutes), and ends where the original chicken and pig cartoons did: an AI model has no skin in the game, so accountability stays human.

Video transcript

Lightly edited from the video’s captions: timestamps removed, punctuation added, Scrum terms linked to the Guide at first mention.

AI doesn’t remove bottlenecks. It moves them.

Your team isn’t shipping twice as much software with AI. They’re just producing 5 times more unvetted code, and burying your Sprint Backlog under automated sludge.

AI tools don’t eliminate delivery bottlenecks. They just move them.

Right now, executive leadership is demanding, quote, “AI transformation,” unquote. Developers are generating features 3 times faster. Product managers are churning out synthetic user stories. As a result, code reviews are overwhelmed, QA is breaking and releases are stalling. Speed without focus is just automated chaos.

If you’re feeling the whiplash, take a breath. We have seen this movie before.

From EDS in 1991 to the dot-com bust

I started writing code in 1991 at EDS, back in the Ross Perot days. I shipped logistics systems at GM’s world headquarters in Detroit. Worked on core, MRP and ERP systems for Domino Sugar in Manhattan, and worked with IBM in Boca Raton on DOS and OS/2.

By the late 90s, I lived through 2 startups. One made it through an IPO, and the other was an absolute textbook dot-bomb crash. Billions of dollars in speculative promises evaporated the exact moment basic delivery economics re-entered the room.

The Enterprise Unified Process, and discovering Scrum

In the early 2000s, while co-authoring the book on the Enterprise Unified Process with Scott Ambler, which was published in 2005, I realized heavyweight processes were suffocating teams. During that research, I discovered Scrum.

In January of 2004, I was trained directly by Ken Schwaber as CSM (Certified ScrumMaster) number 72. By 2006, I became one of the first 25 Certified Scrum Trainers in the world, and created the original Scrum chicken and pig cartoons.

Today, I sit in an active chief product officer seat at Edvizion. The technology changes. The tools evolve. Human alignment and the laws of empirical delivery do not.

Here is what is actually happening to your delivery pipeline right now.

The delivery funnel: why work in progress explodes

Think of your delivery system like a funnel. At the top, you have inputs. At the bottom, you have actual working software in customers’ hands.

When a team uses generative tools to write code faster, they have not delivered value. They’ve just poured a fire hose of uninspected work into the top of the funnel. Your work in progress explodes. Your Definition of Done collapses.

Where does that work land? It crushes your peer reviews. It chokes your automated test suites. It turns Sprint Backlog refinement into an argument over hallucinated edge cases.

Synthetic confidence vs. Scrum’s three empirical pillars

Scrum is built on 3 empirical pillars: Transparency, inspection, adaptation (together, Scrum Theory).

Generative tools destroy transparency by replacing clarity with synthetic confidence. If your team cannot inspect the code or the requirement with their own eyes, you cannot adapt. You are simply running an automated feature factory.

AI theater vs. boring utility

The bubble will burst. The evangelists will vanish. And what’s left behind will be boring utility, just like electricity, broadband and indoor plumbing. We don’t need another autonomous agent demo. We need clean plumbing.

AI theater is generating 200 backlog tickets nobody asked for, or pasting proprietary customer data into public models.

Boring utility is taking 5 conflicting vendor specifications, dropping them into a private secure notebook, and finding the discrepancy in 3 minutes. It’s letting an offline model summarize messy retro notes, so your team can spend their human energy having the hard conversations about priorities.

The Chicken and the Pig: accountability stays human

20 years ago, when I created the original Scrum chicken and pig cartoons, we used them to talk about skin in the game. Who is involved and who is truly committed.

Today, leadership is trying to make an AI model the pig. An algorithm does not have skin in the game. It doesn’t get fired when customer data leaks. It doesn’t lose sleep when production goes down. Accountability remains 100% human.

You don’t need 50 new browser extensions. You need your teams to get back to building working software.

The Saturday Implementing Scrum email

Every Saturday morning around 9:30 Eastern time, I send out my weekly Implementing Scrum email. And every issue, I share one of my original chicken and pig cartoons and reflect on what I’m learning in the real world, working with coaching, mentoring and training clients, and serving as the chief product officer at Edvizion.

And yes, AI is becoming more and more interwoven into how we all deliver products. I share what works, what breaks, and how to keep empiricism at the center of it all.

Join thousands of other leaders at michaelvizdos.com/focus. And if your delivery pipeline is already on fire, you’ll find a direct link on that page to book a confidential mentoring session with me.

Focus. #deliver

Next: Part 2, Code Got Cheap. Accountability Didn’t. walks through what an AI-native delivery loop looks like and why it is not waterfall in markdown. Part 3, Why AI Is Breaking Your Definition of Done asks who decides Done and what to add to your Definition of Done. For the cartoon behind the chicken and pig, see where it came from.

Get the free Study Kit

Full audio book + printable PDF. Includes the Saturday morning cartoons.

Get the Free Study Kit
Michael Vizdos

Written by Michael Vizdos

Certified Scrum Trainer® since 2006. About me · LinkedIn

35+ years working with teams around the world on Implementing Scrum. Found this useful? Feel free to share it with your team, or send feedback. I read all of it.

I sometimes use AI tools to brainstorm, draft and create, including images and video visuals. How I use AI.