You Can’t Automate Chaos

Episode Overview

You can’t automate chaos. That’s the line Hannah Garrett is known for, and the argument this episode tests from both sides of the data center boom. Hannah spent five years on the general-contractor side at Hensel Phelps and Clark, then crossed to the owner side in 2021 for programs and portfolios at Meta, Aligned, Microsoft and now Oracle. The conversation opens on the collision she has watched for a decade, tech companies becoming infrastructure builders and construction companies trying to keep up, and lands on her thesis: AI does not fix an operation, it amplifies whatever operation you already have, so the structure has to come first. Along the way: why delayed projects keep surprising everyone, why the simplest system produces the best data, and what a central workplace for capital projects would look like.

Inside the Episode

The main episode follows the show’s four-act arc, from Hannah’s path into data centers, deep into the tech-versus-construction seam and the case for structure, and out through the takeaway and advice for leadership. The bonus cut carries the advice round for students, the bookshelf and the closing riff. The strip maps each chapter’s share of the 53 minutes; timestamps are approximate. Click any block to jump to its chapter.

Each block links to its chapter below: the map mirrors how the released episode flows.

The Guest · Hannah Garrett

Hannah Garrett

Sr. Program Manager, Strategic Portfolio Management · Oracle

Miami, Florida, United States

Program ManagementPortfolio OperationsData CentersOperating ModelsAI Readiness

Hannah has spent ten years managing complex portfolios in fast-moving environments, and she’s done it from both sides of the data center buildout. She came up through the general-contractor world (Hensel Phelps, Clark Construction), ran projects at The Hanover Company, then crossed to the owner side in 2021: global programs and process at Meta, technical program management at Aligned Data Centers, AI data center portfolios at Microsoft, and now strategic portfolio management at Oracle. Her focus: defining how portfolios operate, aligning teams around clear ownership, and keeping execution consistent as scale increases.

“You can’t automate chaos.”Hannah’s line, the post that sparked this episode

Career journey

The Host · Orion Matthews

Orion Matthews

Founder, Queryon · Host, The Major Project Podcast · PMO data & analytics

Eugene, Oregon, United States

Project Controls ReportingPMO Data AnalyticsCapital PortfoliosAI in ProjectsPower BI

Orion is an award-winning programmer, entrepreneur and technology executive who has spent the last decade in the engine room of major capital projects. He founded Queryon in Washington, D.C. in 2015 with a focus on data analytics and software for the construction industry, starting with bellwether client BP, whose project-controls solutions his team deployed globally, from the North Sea to Azerbaijan to Houston. Today he’s a recognized expert in PMO data analytics and project-controls reporting, serving billion-dollar portfolios for owners including BP, Santos, Balfour Beatty and Turner & Townsend, a regular speaker at major industry conferences on AI and analytics in capital projects, and the host of this show.

The Meteor Moment

Hannah’s path in, and what the boom looked like from inside it. Five years covering every side of delivery on the general-contractor side (pre-construction with developers through managing work in the field), then a jump to Meta’s data center portfolio in 2021, “before data centers were really cool.” Then the world noticed.

“It definitely felt like a meteor overnight… Before, it was like, ‘Oh, I work on data centers,’ and they’re like, ‘What is a data center?’, and you have to explain the cloud.”Hannah Garrett · 2:02

1. The guest intro lands the both-halves framing: GC side at Hensel Phelps and Clark, owner side at Meta, Aligned, Microsoft, Oracle. 0:32
2. Life before the boom: explaining the cloud at dinner parties; life after: “internally, it’s just gotten a little more chaotic, because it’s so busy.” 2:14
3. Why she started writing: “All these themes we’re going to talk about today, I’m super passionate about, but no one’s really talking about them.” 4:37

Opposite Organisms: Tech vs Construction

The episode’s frame. Tech is built to pivot; construction is built to reduce risk. Both are rational, and they are now jointly delivering billions of dollars of infrastructure. This chapter unpacks what actually breaks at the seam: who absorbs a change, whose language wins, and why the delivery team keeps taking the blame.

“Construction is kind of a dinosaur because it’s very reliant on experience, because there is a lot of risk. It’s a physical product. It’s very hard to fix mistakes… You need that standard to exist in order to eliminate the risk, because the risk is very real on the construction end.”Hannah Garrett · 6:05

1. Is the agile mindset right for this world? “There’s a lot of ego involved… they did build amazing things, and they have billions of dollars backing them. But they’ve never worked with physical products like this, where there is real consequence to change, and things are starting to break.” 7:34
2. Why delayed projects keep surprising everyone: “The people actually doing the work have very little say when a change comes down… those conversations are not happening. So the product is delayed, and everyone’s looking at the delivery folks like they did something wrong, when in reality, this conversation was never had.” 9:56
3. The language gap: “risk” and “change order” are heavy words in construction and light ones in software. “This is a big gap in our industry… I’m just not sure who’s going to be the first one to have that conversation.” 11:42
4. Her fix for the whiplash effect: both sides own change; tech runs real change management (“if this delays the project, is this change still worthwhile?”), delivery beefs up how changes get processed. “If you have a structure in place, it’s a lot easier.” 12:37

Structure Before AI

Orion channels the techie (“we’ll 5D-BIM the whole data center, build a world model, let AI forecast everything in real time”), and Hannah answers with the episode’s central argument: AI doesn’t fix an operation. It amplifies whatever operation you already have.

“AI does not fix things, and I think that’s a misconception. It doesn’t fix things, it amplifies things… So it’s really important to set the structure first.”Hannah Garrett · 16:09

1. What she hires for in this environment: “Competency is underrated… take initiative, drive things home, be curious, being able to learn and progress easily.” 18:29
2. On Dorsey’s “hierarchy is an obsolete information-routing protocol”: she can see the central-model future, but “I always call project management the domino for AI: they handle so much complexity. The models definitely have to get better… but it’s definitely possible.” 19:56
3. How companies get amplification backwards: “They’re handing it to their employees and saying, ‘Hey, use this however you can,’ and setting aggressive goals… people spin off in silos. They’re accelerating something, but you don’t know that that something is meaningful.” 20:48

“With five projects you can communicate with everyone involved. When you 10x that to fifty, communication between all those teams is impossible… If companies don’t insert a system before they have that scale, it’s chaos: people are burnt out, they’re in a ton of meetings, you’re having billion-dollar decisions determined in Slack channels. If you implement a good system before you grow, you can grow and scale easily without adding risk.”Hannah Garrett · 22:03 · “chaos is a report card”

Structure Without Bloat

The devil’s-advocate chapter, asked from our audience’s side of the table: project controls already documents everything. Isn’t that exactly how you end up with bureaucratic complexity? Hannah’s answer is a line to keep: standardize everything that should be standard, so people have energy left for what genuinely requires judgment.

“It’s all about meeting in the middle, structure with flexibility. There should be one way to hold meetings, one way to take meeting notes… Every time you have to think, ‘How do I want to do this?’, that creates a decision, and it repeats on every project. Create a standard for everything that should be standardized, so you have more energy to focus on what requires that flexibility.”Hannah Garrett · 23:30

1. The talent cliff is real: a hollowed-out middle, few juniors entering, and heavy turnover. “We have to learn how to teach the new generation in a different way than we’ve ever done before… A good PM gets burnt out and leaves a huge hole, a lot of risk for the business.” 27:12
2. Why PMs burn out: “You’d be surprised how much administrative load they’re taking on: reporting, status updates, chasing people for updates… And you’d think, ‘Oh, have we not automated this?’ And we haven’t.” 28:16
3. Orion’s EHR analogy: doctors who retired rather than become documentation clerks. The same burden is landing on PMs now. Hannah: “I think you nailed it.” 29:49
4. The tool paradox: delivery teams beg “no new tools, please” while leadership buys systems that track every data point. “It sounds counterintuitive, but if you simplify the system, you’re going to get better data… an overly robust system that tracks everything gets you a bunch of data that never gets updated.” 30:27

Visibility & the Central Workplace

The longest and most concrete chapter of the episode: 18 minutes on what leaders should actually be able to see. Portfolio health in five minutes, qualitative signals over vanity KPIs, the contractual walls nobody can tear down, the internal walls everybody could, and the central platform Hannah has never seen built.

“We keep going back to, ‘We just need a new dashboard to show this and this’, but we’re ignoring that it’s the quality of the data that’s missing. A new dashboard gets spun up every day. At the end of the day it’s the reliance of the data, and tracking what’s important.”Hannah Garrett · 33:27

Her four steps for taking over a portfolio

From Hannah’s writing, walked through on-air at 37:05–40:41 · “You have to understand the process end-to-end, get to know everyone involved, then find the gaps… If you continuously do that, you become a great PM.”

1. Measure the qualitative: “How many of your PMs are quitting? How much repeat business are you getting? Are client calls actually going well? It’s a sense of ease versus a sense of chaos and tension.” 35:16
2. The contractual walls: owner → leaseholder → general contractor. “There’s no contract between the GC and the owner… they’re not privy to that information. I do think it’s going to limit us, but I don’t see a way around it.” 41:54
3. The walls we could remove: “Take one project at an owner: physical security, architectural, delivery, all these internal teams, and internally they have no visibility into what the other teams are doing… We should have extreme transparency amongst internal teams before we even talk about external ones.” 42:14
4. Why the simple fix never ships: “We like making incremental improvements, but we don’t want to shake the boat… it’s a huge change, and I don’t think anyone wants to be the one to make it.” Plus the quiet ego tax: “How stupid would they feel coming up with the idea to use this very simple tool? It’s not an ego thing, but it is.” 44:53
5. Orion’s confession from the controls side: RBS, WBS, CBS, OBS; “at some point you step back and you’re just like, what have we done? We’ve created all this entropy, and we’re just trying to build five data centers.” Hannah: keep the skeleton whole, drop the side missions: “‘We need a RACI for this’: if the process is built out start to finish, a lot of that is really not needed.” 46:26
6. Where it lands: keep construction’s hierarchy where risk demands it, take tech’s leanness everywhere else, “finding the best of both worlds and combining them.” 48:36

“When I was at Meta, what we realized very quickly is there’s no good way to see the status of all your projects in one place. There’s no central platform. And this is what I will harp on until the day that I die: you need a central workplace… I really like simple systems, Asana, Monday; they can be set up to handle the complexity. What’s important is that it’s all in one place. And that does not exist today. I’ve never seen it done.”Hannah Garrett · 43:29

The Takeaway & Advice

Acts III and IV arrive compressed and strong: the one idea to leave with and what leadership should do on Monday. The direct advice for students and career shifters, the audiences this show is built for, runs as the opening of the bonus cut.

1. Orion’s synthesis: “clean up our systems” has been a 20-year drumbeat nobody acted on because nothing broke. “AI is going to force that change, and it’s going to have to happen fairly rapidly.” 50:06
2. Hannah on where change starts: “It wasn’t a priority in leadership’s eyes, and it has to be top-down… to completely shift the way that everyone else thinks.” 50:37

The Big Ideas

Three ideas carry the episode. Sketched here the way a portfolio room would whiteboard them; the third one belongs to the bonus cut.

From the Bonus Episode

A companion cut released alongside the main episode: the advice round for students and career shifters, Hannah’s bookshelf, and the wrap-up riff that refused to wind down, in which a bookshelf question reignites the conversation into its best stretch: complexity used to be a moat, AI drained it, and the new moat is speed. Quotes are transcribed from the bonus and lightly condensed; timestamps are approximate.

“Complexity has been companies’ moat: ‘you can’t understand what we do, how will you compete?’ Now you have AI, so you can figure it out. That’s no longer a moat. I think now the moat is speed, and ironically, the faster you are, the less complexity you have.”Hannah Garrett · bonus · 5:27

Bookshelf & Links

Everything recommended in the episode and the bonus, plus where to find the people and the show.

Quote Bank

The best on-the-record lines from the episode and its bonus cut, gathered in one filterable place. “Highlight” marks the lines most worth scrubbing back for; quotes lightly condensed, timestamps keyed to the published videos and approximate.