When ChatGPT and other models came out, Juliette van der Laarse started wondering: how will AI impact my job? While searching for answers, she found grand visions and very specific how-tos, but nothing in between. Even though she’d been there so long colleagues joked she was part of the furniture, ongoing reorganizations made her wary of the consequences that AI could have.
So, next to a busy day job as an engineering manager at a large corporate, she started to build the AI Flower, a framework that would help her determine the impact of AI on her job and the people she managed, and that would guide her in learning if AI was the right solution for the engineering problems she was supposed to solve.
In episode 169, she shares her open-source AI framework that maps 152 IT activities against 25 levels of AI capability to show what evolves, fossilizes, or stays human work. The AI Flower is designed as a roadmap to help small and mid-sized companies be part of the AI transformation without having to spend thousands of Euros on training and consultancies.
We talk about:
- What makes AI not overhyped, but often misused
- What her own redundancy taught her about staying in control of her career
- Why she kept the framework open source instead of charging for it
Plus, find out what happens when your systems and data are not ready for the AI transformation. Spoiler: bias is not the only problem.
Start listening by pressing play below.
Building the AI Flower Framework as the Roadmap for AI Transformation | Show notes for episode 169 with Juliette van der Laarse
- Juliette van der Laarse Founding Director, AI Flower Foundation
- Dirkjan Hupkes Host
- License: All rights reserved
Or keep reading for practical takeaways, highlights, and trends from the episode.
Practical Takeaways for Founders
Even though her background and the word ‘framework’ may suggest a more theoretical conversation for a corporate audience, you shouldn’t let that fool you. Juliette, a former founder herself, shares some highly practical insights that just as easily apply to startups and small companies. Here are three examples.
Don’t be blinded by AI
Many startups are saying they’re AI native and building with AI for the sake of it. But Juliette warns founders to first investigate what problem they’re trying to solve and then make sure that AI is the best solution. Because sometimes, automation or even a simple system fix may work better.
Coaching and mentoring are key to scaling
Many startups already have knowledge-sharing practices in place to ensure that activities are not dependent on a single or a few employees. Juliette takes this one step further. As coaching and mentoring are key to personal and professional growth, she recommends that leaders allow senior employees to spend 40% or more of their time on those activities.
Design AI so it does not self-contaminate
We know that AI models have their own biases and that they can hallucinate. But AI can also contaminate the models with these errors and biases through reinforcement. When we talk about her tips for FundingCoach, Juliette says that awareness of biases is not enough; she tells me to design the tool in such a way that it does not reinforce its own past outputs.
The through-line of these items isn’t about AI as a topic. It’s that in each case, doing the responsible thing costs something (restraint, time, and deliberate design) that the easier path skips, in favor of the long-term result.
🤗 Know a founder who aspires to build an AI-native startup? Put them on the right track to becoming AI native and share these learnings with them.
Or scroll down for magic moments.
Highlights and timestamps
| Time | Highlight |
|---|---|
| 00:00 | Introducing Juliette van der Laarse and the AI Flower framework |
| 05:19 | From childhood programming to AI research and corporate engineering leadership |
| 08:17 | Why Juliette started the framework: clarity, career impact, and AI transformation |
| 10:05 | Why pilot-first AI strategies often stall at the business case stage |
| 13:35 | Why pilots are useful but not enough for enterprise-wide change |
| 16:35 | How the framework addresses fear, trust, and job stability |
| 19:11 | Juliette’s redundancy experience and what it taught her about career resilience |
| 27:33 | Will AI take all jobs, and what has to happen before that becomes realistic |
| 29:51 | System readiness, regulation, and why some work may stay human |
| 32:06 | AI overhype versus useful automation |
| 33:50 | Why shallow IT training can lead teams to overuse AI |
| 36:57 | How mentoring and senior coaching could scale knowledge better |
| 41:10 | The structure of the AI Flower: activities, standards, and excellence |
| 48:02 | AI capability as 25 levels of maturity and scope |
| 51:49 | Mapping AI capability to IT activities and fossilization of tasks |
| 54:57 | How the model shows role evolution across DevOps, test, and infrastructure work |
| 59:27 | Capturing undocumented knowledge and local context in companies |
| 1:03:06 | Building internal knowledge gathering and documentation workflows with AI |
| 1:09:21 | Peer review, critique, and building a board of naysayers |
| 1:15:46 | How the framework changes over time while keeping a stable structure |
| 1:18:41 | Why the framework is open source instead of monetized |
| 1:25:40 | Using the model to map AI risk and suggest controls |
| 1:30:06 | Where companies should start, and the biggest mistake to avoid |
| 1:35:23 | Why Juliette is setting up a foundation around the framework |
| 1:37:35 | Safeguards for an AI funding coach and avoiding self-reinforcing bias |
| 1:41:15 | Where to follow Juliette as the foundation and framework grow |
The Episode In an Infographic
3 Magic Moments In The Episode
When you need a thermometer for your house, would you buy a €300 AI-driven one if you knew that a €10 non-AI-related device could do the same? This comparison comes up when I ask Juliette if AI is overhyped. It’s a funny moment as it magically captures the core of overhyped versus misused. Here are three more magic moments from the episode.
Reflecting while speaking in front of 2000 people
Earlier this year, Juliette presented her framework at DevWorld in Amsterdam to a room filled with 2000 people. Most people on stage would be busy giving the crowd answers to their questions. Juliette recalls asking herself how many people in the audience had the answer she wanted to hear about. It shows her deep desire to make the framework something that people use and contribute to.
The board of naysayers
It’s easy to fall in love with your own creation and go looking for fans; it is more powerful to ask feedback from haters. In fact, Juliette sees it as a smart way to help her identify her own blind spots. But the invitation is limited to the people who bring a respectful attitude and substantiated arguments why an element of the framework will not work as expected. She’s not only showing the vulnerability to invite feedback, but also the strength to use it to improve her creation.
Talent has choices
When she heard that she would be redundant, it would have been easy to use the company’s social plan and take time to find another position. But Juliette had prepared. She had been calling out the reorganization for a while. So by the time she posted about her redundancy, recruiters from other companies were already lining up. It’s a message to other talents that preparation gives choices, and to companies that talent can be gone if you don’t take care of them.
These magic moments show what happens when you start checking how things really work instead of just trusting them to work as planned.
🗣 What was your magic moment of the episode? Was it one of these or another moment entirely? Let me know by sharing your favorite moment in the comments.
Or scroll down for my favorite quote from the episode.
The Quote from the Episode
“I think the genie is out of the bottle. We cannot put AI back. So let’s see how we can get the most out of it and work toward a better future instead of ruining it along the way.”
— Juliette van der Laarse, founding director of the AI Flower Foundation
When we talk about the magic feeling of creating something that coding can have, Juliette shares this to show what is really at stake here.
Feeling inspired to create a better future? The blog is public, so feel free to share the quote with someone who needs to join as well.
3 Trends I’m Seeing Across Conversations
Like most episodes, this conversation has overlaps with other guests. Sometimes, I see the overlap up front; other times I find out after the fact. This episode is a mixed bag.
Implementing AI is about people
Why would you contribute to something that will make you redundant? It’s a simple question that underlines the importance of bringing people along in transformations like the one we’re going through with AI. Juliette mentions that often there is a pilot and a request for a business case for more, but that often no attention is given to the impact on people, their work and their lives. We heard about the importance of bringing people along in episode 136, where Sophia Zitman mentioned that most initiatives stumble over internal politics and resistance, not because the technology wasn’t working.
Solutions don’t come to those who wait
Juliette was looking for external advice to discover the impact AI would have on her job and that of the people in her team. When she could not find a good answer, she decided to build something herself. We heard this before in episode 163 with Dr. Robin Blackstone, who was waiting for disciples to come down to fix the healthcare reimbursement system. When they didn’t arrive, she decided to take matters into her own hands.
Doing the research first is fundamental
As part of developing research, Juliette spent a year collecting information about activities and standards in IT, automation and transformations. She did so in addition to a demanding day job in service of ensuring that her framework would be as complete as possible. In episode 164, we heard from Mirjam Weemhoff, who did 15 years of research with MRIs to find how the pelvic floor rotates and functions in different positions, before building a new, adaptable pessary that the industry wouldn’t build.
What connects these trends is that it takes ownership and accountability to go for the best possible outcome, whether that is bringing people along on the AI transformation, taking action because no one else does, or doing the research needed to succeed.
🗣️ What was the value you got from listening to this episode? I’d love to hear from you in the comments.
Leave a Reply
Coming Up On Women Disrupting Tech
More than half the people in the Netherlands live with a chronic disease, and most of that is preventable through diet. That statistic is what pushed Anke Kuik, co-founder and CEO of FLTRD, to leave a career in media and advertising and build something in food and health instead.
Episode 170 covers how Anke turned a personal interest in nutrition into FLTRD, an app that scans supermarket products for how ultra-processed they are and points people toward healthier alternatives. She walks through the years spent developing the Pure Score behind the app, the decision to rebuild the product in-house after outsourcing it first, and what she’s looking for in FLTRD’s next round of investors.
The clip below is Anke explaining why she started the company: the health statistics that caught her attention, and her observation that supermarkets are full of ultra-processed products while healthy food has become harder to find than it should be.
It’s a small piece of a much longer conversation about food, health and entrepreneurship. To catch the rest when the full episode drops on Thursday 17 September at 16:00 hours CEST, leave your email address below.
Until then, as always, keep being awesome,
Dirkjan
What I Want to Leave You With
This episode starts on a terrace near Amsterdam where Juliette first explained the AI Flower Framework to me. We had met at a conference years before, where I asked her and a colleague to come on the podcast to talk about practical allyship for women in IT. That episode never happened, but we always kept in contact.
What fascinated me was how Juliette manages to combine deep knowledge of industry standards and frameworks with the practical, day-to-day impact that a complex technology brings. Add to that the effort she put into building this framework, and you understand why I jumped on the opportunity to have her explain the AI Flower on the podcast.
While listening, I also noticed her humility about the AI Flower. She does not present the framework as the only framework you’ll ever need to implement AI. Instead, she admits that there may be flaws in it and opens it to the public via an open-source construction. What’s more, through the open-source construction, she makes her knowledge accessible to the many small and mid-sized companies that cannot pay for the expensive consultants and €2000 training sessions that come with transformative technology like AI.
So when she mentioned she wanted to learn from Linux, I could not resist linking her with Nithya Ruff the next day. As a result, she will have an opportunity to share the AI Flower at the Linux global conference in Prague in October.
Listen to the full conversation with Juliette van der Laarse on Spotify, Apple Podcasts, or YouTube. And if you are a female founder navigating a funding conversation, take a look at FundingCoach.ai.
About Juliette van der Laarse
Juliette van der Laarse is a Senior Engineering Manager at the largest online retail and e-commerce platform in the Netherlands and Belgium, Founding Director of the AI Flower Foundation, and an O'Reilly author.
She started programming around age eleven, guided by an uncle who worked in IT. That early spark led her to study game technology, where she took her first AI course, and from there into freelance work, an AR/VR startup, and a year as a researcher at the Asimov Institute studying AI's impact on society. She then spent five and a half years at a major international financial services company, first shaping workforce strategy for thousands of engineers, then leading engineering teams across cloud, testing, security, and reliability platforms — and along the way founded and chaired a women-in-IT community of more than 250 members.
It was there, alongside a demanding day job, that she built the AI Flower: an open framework that maps 152 IT activities against 25 levels of AI capability, giving companies and engineers a concrete way to see where AI actually helps, where it doesn't, and how roles change along the way. She introduced the framework on O'Reilly Radar, presented it to 2,000 people at DevWorld in Amsterdam, and will deliver a featured keynote about it at the Linux Foundation's Open Source Summit Europe in Prague this October.
She is now establishing the AI Flower Foundation to keep the framework open and community-governed, so that companies without enterprise consulting budgets can be part of the AI transformation too.
Juliette was recognized as a technology role model by the European Parliament and named to the VIVA400 list of the most inspiring women in the Netherlands, and she is a regular keynote speaker and industry-award jury member.
You can connect with Juliette on LinkedIn.
About the AI Flower Foundation
The AI Flower Foundation is Juliette's effort to turn the AI Flower Framework into something more than a personal project. As of this recording, it doesn't fully exist yet, she's in the process of registering it, working through the Chamber of Commerce and a notary to put the structure in place as an official non-profit. Once that's done, the framework itself will also be published on GitHub under an open-source license.
She's leaning the foundation's governance loosely toward the style of the Linux Foundation, not copying it directly, but taking cues from how an open, community-run project like that operates. A registered non-profit structure gives the work more credibility, makes it easier for outside organizations to collaborate with it directly, and opens the door to sponsorship, without needing companies to fund what would otherwise look like one person's side project.
Central to how the foundation runs are what Juliette calls dialogue tables: structured sessions where people with genuinely opposing views on a specific IT activity are deliberately brought together. The goal isn't quick agreement; she treats disagreement itself as the best safeguard against bias creeping into the framework, since the more people are involved, the more they bring their own assumptions with them. Long term, she'd like other universities and companies to host their own dialogue tables independently, contributing directly to the model rather than routing every addition through her.
For companies that sponsor the foundation, the relationship she's aiming for runs both ways: sponsors would get a say in which capabilities get developed next, and the chance to be first to pilot new work as it's built.
To learn more, visit the website or follow the AI Flower Foundation on LinkedIn.
Listen to Episode 169 on Spotify, Apple or YouTube
Featured Event: TEDx Coolhaven Women
The TEDx Coolhaven Women event takes place on 9 October from 6:30 to 11:30 in the evening at HAL4 aan de Maas in Rotterdam. This year’s theme is The New Builders, a theme that seamlessly aligns with the content on this podcast.
Throughout the evening, you’ll hear powerful TEDx talks from women who are redesigning the systems that shape our lives—from leadership and parenting to women’s health, mental healthcare, and access to investment capital. Together, they’ll explore what it really takes to build a better future.
As a proud media sponsor, I am happy to offer you a chance to be there in person, to listen to groundbreaking talks and meet and network with like-minded people. As a listener to the podcast, you’ll get a 25% discount when you use the code WDT25 at checkout. And even though this is a TEDx Women event, men are very welcome to join.
- 🗓️ 9 October 2026, from 18:30 - 23:30
- 📍 HAL4 aan de Maas in Rotterdam
- 🎟️ You can buy your ticket here.
Leave a Reply