There is a problem at the center of almost every AI transformation conversation that nobody has named quite right. Heather Xu has a name for it: the pace gap.
The technology is moving fast. The people inside organizations are trying to move as fast as they can. The organizations themselves are trying to do the same. But these three things are not moving at the same speed, and the gap between them is not a temporary inefficiency to be optimized away. It is a structural feature of this moment, one that will keep reasserting itself no matter how many change management workshops an organization runs.
Xu is the founder of Design for Flow, a strategic advisory practice built on the human side of AI transformation. She spent twelve years as a director at KPMG building the firm’s learning strategy and transformation capability from the ground up. She is also, by her own description, a systemic sense-maker: someone who gets lost in complexity long enough for something coherent to emerge, then builds frameworks simple enough for others to navigate.
The conversation with Jason Gorman on The Learning Experience Ops Show produced two of those frameworks. Both deserve to be in every L&D leader’s toolkit.
Listen to the podcast on your favorite directory:
Four Ways to Work With AI
The first framework came out of a project Xu did with an L&D team roughly a year ago. She noticed something consistent in how they were using AI: an anchor toward production. Create the learning objectives. Generate the content. Automate the output. The tools were being treated primarily as accelerants for existing deliverables.
She pushed them to think about what they were leaving on the table.
The result was a model she calls the four modes of human-AI interaction, organized around the motions of work. The first is Coach: grow me. This is AI as a guide, a teacher, a thinking partner that helps you develop yourself. The second is Thought Partner: think with me. Not trusting AI to provide wisdom, but co-creating your perspective through iteration, questioning, and exploration. The third is Accelerator: work with me on an output. Bringing your own perspective, pressure testing it, honing it, looking at it from different angles. The fourth is Automator: work for me. The agentic AI movement, delegating a task and reviewing the result.
The insight is not just that these four modes exist. It is that most people and most organizations are dramatically over-indexed toward the last two, while under-investing in the first two.
“The real value-added activities are when people ask the more strategic questions,” Xu says. “Are you anchoring toward using AI for outputs? Could you be using it more to shape your thinking and your ideas?”
The practical implication is significant. When an L&D professional uses AI to generate a course outline, they are working in accelerator or automator mode. When they use AI to challenge their assumptions about what the learner actually needs, to surface research they have not encountered, to stress-test a proposed intervention against the business problem it is meant to solve: that is thought partner mode. The latter produces better work. It also produces a more capable practitioner.
What Your Strengths Actually Are
The second framework is more personal. Xu calls it Amplify, Unblock, Stretch, and Free, and it is built on a distinction she first encountered in a book she technically stole from a Microsoft manager who was not going to read it.
First Break All the Rules, by Marcus Buckingham, introduced her to an idea she has never fully let go: your strengths are not necessarily what you are good at. They are what makes you feel strong. The distinction matters enormously.
“You might be really good at something,” Xu explains, “like I can edit a document to perfection. Is that what brings me to life? No. That final twenty percent of getting something perfect, that is what drains me.”
From that foundation, the personal value framework asks four questions. Where can AI amplify what already makes you feel strong? Where can AI unblock you when you get stuck, whether in decisions, in paralysis, in the infinite scroll of hotel options for a trip to Morocco? Where can AI stretch you past your comfort zone and hold you accountable for the growth you have been deferring? And where can AI simply free you from the things you hate, the admin tasks, the repetitive processes, the work that has never belonged with a human in the first place?
Applied honestly, these four questions produce a genuinely personal AI strategy. Not a job role’s relationship with AI. A person’s.
The Human Connection Team
One of the more striking ideas in the conversation is Xu’s reframe of L&D’s future role. If AI can now do coaching, thought partnership, content generation, on-demand learning, and skills reinforcement at scale, the obvious anxiety is that the learning function is being hollowed out from the inside.
Xu inverts the frame entirely.
“Who else is out there saying: we really need to get our people together to collaborate, to be social, to learn from each other, to build enduring human skills?” she asks. “Can L&D be the human collaboration team?”
It is a provocation worth sitting with. Every trend in organizational life is pushing toward efficiency, delegation, automation, remote-first, async-first, output-first. The counterforce, the argument that human connection is not a soft program but a strategic capability that is actively eroding and needs to be designed for, is not finding a home in most organizations.
L&D might be the function best positioned to make that case. Not because it is the obvious home for it, but because nobody else is making it.
Agency Is Not Self-Motivation
The thread running beneath both frameworks is human agency: the inner drive to be proactive, to take the reins, to maintain some sense of direction and control in a work environment that is changing faster than any individual can fully absorb.
Xu is careful not to treat agency as a trait some people have and others lack. She sees it as something that can be cultivated, and that looks different depending on who you are.
She has identified five archetypes. The disciplined builder who constructs steadily and sees things through. The curious explorer who generates energy through discovery. The systemic sense-maker who needs to bring things into a coherent whole before creating. The social collaborator who needs human interaction to generate momentum. The rapid creator who needs to put something out and see how it lands.
The conventional advice for cultivating agency tends toward the driver archetype: just go build something, just go do it. For people who do not identify with that energy, the advice does not land. Xu’s framework is an attempt to meet people where they actually are.
“Taking that first step is usually enough to just get the juices flowing and momentum going,” she says. “Where you can get out of a sense of paralysis.”
For organizations, the implications are uncomfortable. Most workflows, performance metrics, and learning interventions are not designed around where individuals get their energy. They are designed around roles and outcomes. Closing the pace gap will eventually require closing that one too.
The Walk Between Prompts
Xu ends the conversation with a question that does not have a clean answer yet: as AI takes on more of the cognitive load of work, where do humans go to rest?
She is not speaking metaphorically. Research she has been tracking shows that the elimination of low-cognitive tasks is creating a new kind of exhaustion. Every task is now important. Every interaction demands full presence. The mental resting time that used to exist in the margins of work has been optimized away.
The image Jason Gorman offers is apt: a developer who sends Claude a task, goes for a walk while it works, comes back to review, sends the next task, goes for another walk. It sounds like a joke. It might be closer to a model.
“We need to think about the cognitive burden that is being placed on people as work is intensifying,” Xu says. “If we don’t recognize humans as part of the equation we’re solving for with new technology, we’re going to erode the skills and the value that make our organizations competitive.”
Design for Flow is not a workflow optimization practice. It is an argument that human thriving and organizational performance are not in tension with each other, and that the organizations that figure out how to design for both will be the ones that are still standing when the pace gap closes.
Listen to the full conversation with Heather Xu on The LX Ops Show or wherever you listen to podcasts.
Podcast Transcript– Jason Gorman with Heather Xu
Jason Gorman (00:00)
Welcome to the Learning Experience Ops Show, where we talk about making learning better that’s future ready. I’m Jason Gorman, the founder and CEO of Jackrabbit Learning Experience. I keep coming back to a feeling about our work. We all know the technology is moving fast, and we all know that the people, us, we’re trying to move as fast as we can, and our organizations are trying to move as fast as they can also.
But I think that the technology and the people and the organizations are moving at three different speeds. ⁓ And every organization moves differently. Every person in that organization moves differently. So this creates a bit of a three-body problem for us, right? It’s just impossible to understand the trajectory and future state of what our work is going to look like.
because all of this is the very definition of a complex problem. And my guest today has a name for this. She calls it the pace gap. And more importantly, she has a framework for closing it, not by picking one side over the other, but by treating workflow design as a human capability that is an integrated challenge with the organizational needs. So I’m also excited.
about this conversation for personal reasons. So Heather was introduced to me by my business partner, James Altman, went to graduate school together at Northwestern. And when we were talking, James and I were talking about this podcast and what would be important to talk about. I don’t think I told you this yet, Heather, but he talked about you and a conversation he had with you and your notion of, you know, design being a thing that should be applied to organizations. And so we were, you know, in
Heather Xu (01:38)
I don’t know.
Jason Gorman (01:50)
And that really has become kind of a central idea in this podcast. So I’m glad you’re here. Thanks for coming on. So Heather Xu is the founder of Design for Flow. It’s a strategic advisory practice founded on the human side of AI transformation. She spent 12 plus years as a director at KPMG in the Human Capital Advisory Consulting Practice.
where she built and grew the firm’s L &D strategy and transformation capability. Her work spans organization design, workflow optimization, leadership development, and learning experience design, all centered on one question. How do we create environments that minimize friction, accelerate flow, and enable people to thrive? Welcome to the show, Heather.
Heather Xu (02:36)
Thank you. It’s my pleasure. Sure.
Jason Gorman (02:38)
All right, so let’s jump in with those lightning
round questions. What is your 30 second elevator pitch about what you do?
Heather Xu (02:46)
I think you just said it, really it’s how do we design work that flows? ⁓ It’s figuring out that human side of AI adoption and how do we shift our environment in a way that enables human thriving? So how do we do it in a way that amplifies human potential ⁓ as well as realizes the benefits of the technology?
Jason Gorman (02:48)
Sorry.
All right, and what do you love the most about your work?
Heather Xu (03:14)
I-
I love that it’s like a puzzle to me. So I look at the organizational level of workflows, operating model design, you know, a drive toward organizational outcomes, but then I’m able to go really deep into the human side, the underlying psychology, the mindsets, the motivation, and how do we create, how do we kind of meet in the middle where we create something that flows from an effectiveness and efficiency perspective, but also flows for the human inside of the process where they feel like,
they can move forward with that sense of inner flow where they’re bringing their best selves to the table and I don’t know just kind of thriving at work.
Jason Gorman (03:56)
Well that kind of leads nicely into my next question. When you are in flow, when you’re in that state, what are you most likely to be doing in your work?
Heather Xu (04:05)
I’m, I think what I love most is sort of getting, getting lost in the chaos and complexity about the current state of what exists or what’s evolving. And then like sitting there long enough until something coherent emerges and I can create something simple. That’s not just simple for me to help understand it, but also simple for others to help people navigate it and whether that’s.
Jason Gorman (04:12)
you
Heather Xu (04:33)
a framework, whether that’s a learning program, whether that’s just a way of thinking, ⁓ guiding principles. That’s, I think, my sweet spot.
Jason Gorman (04:44)
sweet spot to let’s look at other side of the coin here anxiety what is your biggest anxiety for the future
Heather Xu (04:50)
out.
I mean,
right now, FOBO I heard that term recently, the fear of becoming obsolete and definitely, definitely is high up there on the like existential crisis side of things. ⁓ I do find that the more I lean into AI and these emerging technologies and find a way that they amplify my strengths, the more…
Like I’m able to manage that. ⁓ And then I think the other thing is just the fear of not knowing what’s real. I think we’re getting so much AI generated content without the validation or the oversight. And even intentional sharing of AI generated content. ⁓ I don’t think our society is prepared to.
to get to the point where we can’t recognize what’s actually real.
Jason Gorman (05:52)
Yeah.
Heather Xu (05:54)
Yeah,
those are pretty deep fears.
Jason Gorman (05:57)
That’s a pretty deep fears.
All right.
Heather Xu (06:03)
I do
live in Puerto Rico and I’m like, there’s always, you know, we can always grow bananas. We can always grow tree. It’s like, we can live off the land if everything goes to crap.
Jason Gorman (06:16)
That’s good. I suppose I could do that in my yard here in Boston if I just dug up sections of it. Less effective in the winter, I suppose. Last lightning round question, and this can be a work thing or it could be a personal thing, but what is something, a skill that you’re trying to learn right now?
Heather Xu (06:20)
Yeah.
So ⁓ I’m trying to kind of push myself in learning the technical side or like technical capabilities that were.
beyond my reach before, because now with AI, you don’t really need the technical skills to do technical things. So things like building in Claude code, experimenting with different tools. And I’m trying to look at what I always saw were my gaps in capability. So if we take learning program design, for example, the video creation, the interactive element creation, the things that
I would have had to hire somebody to do in the past. Can I now do those things with the tools available and kind of fill my own personal skill gaps?
Jason Gorman (07:23)
already a terrific conversation. It’s just got us a whole bunch more interest in this conversation. So we’ve had conversations before and you told me a bit about your background and you have told me that in the early part of your career you spent years in Shanghai and while you were there you were doing various things and you describe your time there as being in a, I love this term, a career wilderness.
kind of working in lots of different types of jobs. You mentioned like learning, coaching, communications. But I’m kind of curious to hear about some of those experiences that were early there and then kind of what your turning point was to travel out of the wilderness and into the next chapter of your life.
Heather Xu (08:14)
Mm-hmm.
⁓ So I love my formative Shanghai years. I moved there right after college and I was an English teacher for a year. I worked at Microsoft where I did ⁓ coaching for English, for communications and customer service for the engineers there.
⁓ at EF Education, was curriculum designed for some of their English schools, their online school, as well as their in-person schools. And then I pivoted to be a business editor for the transfer pricing tax practice at Ernst & Young. And ⁓ I think part of the, you know, I moved to China because I was drawn to it, because I liked the exploration. I was sort of following the spark that led me to different opportunities.
But as a foreigner, you weren’t really positioned in companies for a longer term career path. So I felt like I was kind of jumping from experience to experience. And ultimately, it got to the point where I felt like I needed to shift and kind of bring it all together. But what I love about it was that it really was kind of cultivating a lot of underlying skills that continue to serve me today.
And I don’t think, I left China and I pursued my master’s program at Northwestern where I met James. But I don’t think it really clicked as to how valuable all of the different experiences were until I was interviewing with KPMG. And they looked at my resume and they said, wow, this is rich experience. Because when I was interviewing for…
some different kind of companies, learning teams, they wanted me to fill a box. And I didn’t have the all the pieces of the box that they wanted me to completely fill. But when I pursued consulting, they saw the value of everything. And I feel like somebody else had to kind of call it out for me to fully realize it. ⁓ But I think it’s a lesson in life that
You know, sometimes we don’t set out to kind of drive this linear path. We don’t want to, and that’s okay. And if we’re following that spark, the thing that lights us up inside, and we’re staying true to that, every experience grows you and helps you become more of who you are.
Jason Gorman (10:38)
want to just kind of focus in on one piece of what you just said. I think you are the third or fourth person that I’ve interviewed on this podcast. This is my 20th-ish episode. You’re the fourth person who said that they ⁓ taught English overseas. And I don’t really know what to make that. That’s also in my background. I was also, I taught English in the Emirates and here in the States.
Heather Xu (10:59)
Yeah.
Jason Gorman (11:08)
I think it’s a curious pattern actually that I’m seeing across this. I think people maybe, I guess in my case, like was interested in education, so like it sort of matches up. ⁓ But then maybe I was more interested in the curriculum development than I was in the teaching ultimately. don’t know, how do you make sense of the teaching part?
Heather Xu (11:30)
Well, I was a journalism undergrad. So I had more of the language background and then the cultural exploration interest. And when I was teaching English, and I did my CELTA certification as well, which is the Cambridge ⁓ English Language Teaching for Adults certification, which was a really powerful, really, really powerful adult learning experience, like maybe one of the best in my life. ⁓
And I realized that I liked the learning side, but not so much the English side. You know, I get lost in grammar. It’s one thing to be able to write well, it’s another thing to be able to explain past progressive tense ⁓ effectively. So I think that was the thread that kind of carried me was the creation of experiences that help the light bulb moment go off for people. But then how do I do that? Not…
Jason Gorman (12:05)
Yeah. Yeah.
Heather Xu (12:27)
in an English classroom.
Jason Gorman (12:28)
I appreciate the reflection you talk about in that moment where you kind of solidified what you wanted to do, possibly right before ⁓ you started at KPMG. You read a book called First Break All the Rules, which is a fabulous title by Marcus Buckingham. And I’m kind of curious what what was it in that book that
sparked change for you.
Heather Xu (12:58)
⁓ so it’s an interesting story. They, they gave that book out to all of the managers at Microsoft at the time. And one of my friends and he made, yeah, let me borrow it. And I like, I stole it. I never gave it back because it talks about, I still have it. ⁓
talks about how to create a strengths-based workplace that doesn’t define people’s responsibilities by the org chart and the job title, but instead, how do you… ⁓
create an environment where people are able to do what they do best. And more people might be familiar with the Strengths Finder and like Now Discover Your Strengths, like those books by Marcus Buckingham. This is the one that came first. And ⁓ the premise was really like your strengths are what make you feel strong, not what you’re necessarily good at. And I think as a young kind of in my first corporate job at Microsoft, like that really resonated.
But moreover, I looked at the book and I said, how do I do that? Like, how do I, what is the job where somebody is responsible for helping create that sense of like strength-based organization in their company?
And I don’t know that I’ve ever figured out how exactly to do that. You know, I think that that is still, it’s still an ideal view in my mind. But, you know, I think that I bring a lot of that to the work that I do of how do I, how do I create workflows that bring people to life? How do I kind of make sure they’re focused on the right outcomes, but also the human experience.
And so that’s that through that book, I was like, what’s the field that I need to be studying? And that’s how I ended up finding ⁓ the learning and organizational change program at Northwestern.
Jason Gorman (14:42)
That’s great.
Got it. I don’t remember exactly how you said it, but if something like strengths aren’t necessarily the thing that you’re good at, they’re the thing that make you feel strong at work. Can you say more about that?
Heather Xu (15:01)
make you. Yeah, just what your strengths are.
So you might be really good at something like I am for my own personal life. I was a journalism major, I can edit something to perfection, I can hone, you know, the sentence structure, the document.
Is that the thing that brings me to life? No, not at all. Like that final 20 % of getting a document perfect, that is what drains me from the whole experience. And if I could offload that to somebody, AI, ⁓ I would be very happy to do that. ⁓ And so the recognition is like, it’s about energy. Where do you get your energy? What makes you feel strong? It’s not what you might be perfect at.
Jason Gorman (15:23)
I see.
you
Heather Xu (15:50)
And it also might, you might feel strong in some things that you’re not good at, but maybe that’s a sign that you need to cultivate those skills.
Jason Gorman (15:57)
Thank you. All right, well, let’s talk about KPMG. That’s the next part of your story. So you went there. You built their learning strategy capability from the ground up. That sounds like a challenge. So what was the hardest part of building a new capability like that?
Heather Xu (16:02)
Okay.
you
So KPMG, and I love my experience there and I still do a lot of work with the firm. ⁓ They specialize in transformation. So HR transformation, finance transformation, and then the group that I was in, which is now called Human Capital Advisory, they… ⁓
we had a learning capability, which was all focused on how do we support other transformations with learning ⁓ experiences, learning programs. But we didn’t have our own, how do we approach learning strategy and transformation? ⁓ And the reality is that learning organizations, and I think this is one of the bigger challenges, there might be a learning function, but that’s not typically the entire ecosystem of learning that’s happening.
happening at organizations. And so part of it might fall in HR, part of it might fall adjacent to HR, but then also happening within functions with the different departments, also shadow learning organizations. Like there’s a widespread of possibility there. And what we did was kind of take the best of what the firm
how the firm approaches transformation with operating model design and process mapping, process taxonomies, et cetera, and then thought about how do we bring that to a learning function. ⁓ And so that was, I mean, it was a very exciting time. And I’m glad that that capability still exists and is still evolving ⁓ at the firm. And I’m glad I could kind of be part of the inception of it. ⁓ But I think…
So I don’t know if that was the hardest part, but that was sort of how we approached it. And I mean, the hardest part was doing something for the first time without the scaffolding of other people’s experience underneath you. But I think that was also the most exciting part too.
Jason Gorman (18:13)
Sure. Yeah. Yeah, hear that. One thing that I mean, it’s almost like a trope in L &D is this seat at the table. Like we need to get our seat at the table, right? And I mean, when I say trope, it kind of is a trope, but I think it’s also based on a thing that people feel and experience. It’s a real thing, right? That their organizations don’t pay the right kind or enough attention to what they do to understand their capabilities in the right way.
Heather Xu (18:36)
Mm-hmm.
Mm-hmm.
Jason Gorman (18:44)
And I can
only, it’s hard not to connect that pervasive pre-existing anxiety with the anxiety that you mentioned earlier about like the kind of existential crisis feeling that people in the learning professions I think are having now. It’s pretty widespread. It’s not just you, it’s not just me. It’s pretty much everybody I talk to. And I’m curious. ⁓
Heather Xu (18:58)
Mm-hmm.
Mm-hmm.
Mm-hmm.
Jason Gorman (19:13)
What do you think are the things that we can do to keep ourselves relevant and to be taken seriously with the stakeholders that we work with?
Heather Xu (19:24)
I think there’s a lot of different ways I could approach this. Let me think for a second. So I think the first thing is.
And maybe AI is enabling this a little bit too with being able to take on some of the work that L &D teams have historically been not just responsible for, but like drowning in. ⁓
because I’ve seen the Evolution League and L &D function wants to be strategic and suddenly it’s compliance training and this training and process training and, you know, and sort of all of the intakes of this is a need, this is a need and your short staff, you know, ⁓ your minimal staff is sort of drowning in these requests that aren’t the most value add. So hopefully AI is enabling some offloading. ⁓ But the first step is
understanding the business strategy, 100%, you have to start with where’s the business going? What are their needs? What are their people needs? And how are we uniquely positioned to move the needle? ⁓ If you don’t start there, you’re gonna miss the mark. And I think that lot of, historically what I’ve seen a lot of learning organizations kind of,
evaluate their success by the collective view of their learning programs. So it’s like we have this curriculum of programs and that in and of itself should drive our identity and our success. And I think that that’s changing. And so I would challenge learning leaders to kind of think about is there a different value proposition? Are we really the sum of our learning programs or are we something else?
Jason Gorman (21:08)
Hmm.
Heather Xu (21:12)
And ⁓
I think like I’ve seen it where, you know, teams, when central teams really do show that they understand the business and they understand the needs and they’re bringing a collective creative system, a creative approach that’s actually kind of shifting mindset, shifting behaviors, enabling capabilities. The business recognizes that and sees them as a partner.
Jason Gorman (21:41)
Hmm.
Heather Xu (21:42)
But
if they’re off kind of creating a set of courses that may be in and of itself, they might be really good courses, but if they’re not moving the business needle, then it’s just going to be something on the side that’s taking up the time. It’s not even that an hour long course is kind of a nice to have value. It’s something that’s detracting people from doing their jobs.
Jason Gorman (22:08)
⁓ huh, yeah.
Heather Xu (22:08)
And so
if it’s not adding value, it’s reducing value.
Jason Gorman (22:12)
Yeah, I mean, we’re always, when I say we, L &D specifically in a corporate environment, it’s always a cost line. Right? And so this is something that I think we’re all really aware of. Anybody who works in the corporate environment understands that. And I think the only way out of that is to make it vividly clear how our actions are contributing to revenue.
And that’s really hard. What a challenge.
Heather Xu (22:44)
And I, so one of the things, and I think this might lead to another question, but I might be jumping ahead. ⁓ but I also think that our,
value isn’t just programmatic or it doesn’t need to be. And especially, I mean, we’ve seen this over the years as we’ve kind of moved from classroom to e-learning to micro learning to, you know, to curation to content libraries. Like there’s there’s been already a shift in how we deliver value to our organizations. ⁓ But I.
think, and I’ve thought this for a while and it’s becoming even more important, that human connection is going to be even more important in the future because of AI. And we’re already starting to see it in the research that.
people using AI tools, it’s now they’re becoming less emotionally intelligent when they’re interacting with others because they’re more impatient. They’re expecting people to react and support them as quickly as AI can react and support them. ⁓ And so I think that as we…
As we see this shift where AI can now, I mean, I use it to troubleshoot things. I use it for technical support. I use it to help me learn new capabilities. I use it to help me shape my thinking. We’re seeing AI grow in possibility as a personal coach, a thought partner. ⁓ That really enables learning in the flow of work. when organizations can do that in a way where…
You’re bringing the assets on the knowledge basis to, combination with AI to surface things at the moment of need. That’s an amazing capability that we didn’t have. And so can and should the learning function now kind of elevate out of that or help shepherd that, but also focus on human connection.
Kind of, I don’t know who, what other team is out there that’s saying, you know what, but we really got to get our people together to collaborate and be social and learn from each other and, build those enduring human skills and not lose sight. You know, it’s not, it’s not, we’re not designing or we shouldn’t be designing for human plus AI. We shouldn’t be designing for human plus human plus AI. And so can L and D be the brainchild and the
and the, I don’t know, the human collaboration team ⁓ potentially. And I think nothing’s figured out for the future.
Jason Gorman (25:28)
Yeah.
No, nothing’s figured out, but I think that’s ⁓ an interesting way to think about the anxiety too, right? I think the anxiety is about will the thing that I’ve been doing be a viable job in the future? The answer is probably no. For many, not everything, but I many things that people who do this kind of work do. But I do think that a lot of our skills are very translatable.
Heather Xu (25:45)
Bye.
Jason Gorman (25:59)
the future and I’ve heard this idea echoed in places. I really would love to focus on you two frameworks ⁓ that you have and I’d love to talk about those. One is a framework that is really about organizing how people can relate to AI, kind of the different modes and those for you, you named a couple in a second ago. Coach, thought partner, accelerator and automator.
Heather Xu (26:09)
Mm-hmm.
Jason Gorman (26:29)
and I would love to understand each one and then how we position ourselves in relationship given what we’re trying to do.
Heather Xu (26:37)
Yeah, and so great. That one was really built around the motions of work and how we align AI. So this is like process, AI enabled process lens of how do we ⁓ think about the interactions that we’re having with AI according to the motions of work. so ⁓ coach is anything related to teaching, guiding, know,
helping us shape our own thoughts. ⁓ And that’s, you know, grow me the motion is grow me, I need to grow. And so I’m going to lean on AI to help me myself grow. Thought partner is think with me. So I’m not fully
trusting AI to give me wisdom ⁓ or in this relationship, I am kind of asking questions, I’m exploring, I’m co-creating my thinking alongside AI. So it can go research things, it can bring things to the table, I can pontificate, it comes back and sort of through that iteration, I sharpen my own perspective. ⁓
accelerator is I’m working on an output and I’m not yet kind of tasking it for creating it in entirety. Maybe I’m coming with my own perspectives, getting some input, we’re accelerating that production of something tangible and maybe pressure testing it, looking at it from different angles, honing it. And then the last one is automation and that’s what we’re seeing with the agentic AI ⁓ movement.
and all of the transformation that’s going on is like work for me. And so this was my lens to think about how the different types of interactions that we can have with AI and where they make sense based on a workflow and what I need as a human in that workflow in that moment. And I did a project with an L &D team about a year ago.
where I, you know, we worked through some of these things. I didn’t have the model yet. And I noticed there was an anchor toward like produce the outcome, automate the process. You know, there was a definite like.
create learning objectives, create content. And I really push them to think about, how do we start the process differently? Help me learn about something, help me pressure test ideas, me think about this differently, and then leading into more acceleration and outputs. ⁓ And the one other thought that I have is that, so I think a lot of companies are moving toward more of the
Jason Gorman (28:57)
Yeah.
Heather Xu (29:23)
automate, et cetera. ⁓ But there’s been recent research that shows that the real value added activities are when people ask the more strategic questions. And so in your workflow, are you anchoring toward using AI as outputs for outputs? And should you, could you be using it more to shape your thinking and your ideas and, you know?
Jason Gorman (29:35)
Hmm.
Heather Xu (29:50)
getting more value from it in all of the four pillars.
Jason Gorman (29:54)
So it sounds like what you’re suggesting here is that in a workflow, when we’re articulating a workflow, that we’re really explicit, right? Step one, maybe we’re doing some discovery and we’re gathering stuff and AI has a place with that. Step two, we’re not asking for the output. And step two, this is where we are testing our assumptions. We are generating lots of different possible directions. We are finding the gaps in the data that we’ve
we’ve gathered, whatever, that there are, there is an explicit step where AI has a valid, important role, but it is one that is like introducing friction that provokes human thought, like interaction with the content and the problem at a deeper level before moving to the solution. Is that, but it’s like a step like, or multiple steps.
Heather Xu (30:51)
Yeah, I think it’s part of the process for sure. might be cyclical. You might go down and then you might come back and you might circle. So it’s not necessarily linear. ⁓ But I think it needs to be part of the process or else we’re just going to surrender our thinking to the machine, which I don’t think builds us up.
It kills our motivation. kills our sense of identity, you know? ⁓ And so I think we can get a lot more value and actually elevate our own skills if we’re approaching it from this way. And I think what I like about it is that it’s just four. Like I tried to make it very simple and I can see myself in these relationships in different ways. ⁓
Jason Gorman (31:30)
and print.
I, ⁓
I see it too, and I definitely feel it. I just did a huge project. I was trying to learn Claude Code and really pushing myself to learn these technologies more deeply. And I really found that ⁓ if I didn’t insert myself in the right ways, honestly, the outputs kind of sucked.
Heather Xu (32:04)
I agree.
Jason Gorman (32:07)
And I think in that too, it made me realize that using AI to do that back and forth, that’s helpful. AI is helpful there. But then it helped me understand. And I think what you’re suggesting is that when people use the tools like that, it will help them realize their value as a stakeholder in the process of doing whatever it is that they’re doing. And that the AI left on its own actually doesn’t do a great job.
It misses things, it doesn’t understand things, it kind of over indexes on things that humans wouldn’t or an expert wouldn’t. ⁓ So I deeply appreciate that. You do have a second framework. It’s a little bit more personal. It’s amplify, unblock, stretch, and free. And I’m curious about this one too.
Heather Xu (32:59)
So this one I call the personal value framework for AI. And this isn’t about your workflow necessarily. The first one is, let’s think about the workflow and think about the AI relationships. This one is more about my own personal value proposition with AI. And so if we take Amplify and we circle back to that strength-based workplace, and your strengths are what make you feel strong,
What are your strengths? What are you naturally good at? And how can you, what are the opportunities for you to work with AI to amplify your strengths? So I, like I said, I like to kind of get lost in the complexity and like come back up and build a model. ⁓ AI can help me not get too lost in that complexity. Like it can.
Go find the insights. can consume them. I can think about them, discuss them. I can like pull that thread of like, I think there’s something here and it mirrors me and it helps me sort through my own thoughts. I like, it brings me a level of coherence where I have an idea and I’m actually able to follow it through to something.
in a way that I haven’t been able to do before. And so I find my own personal amplification there. ⁓ But I also use AI to unblock me. So like, where do you get stuck?
in decision making, in what are your tendencies for paralysis? And I’m planning a trip to Morocco in a couple of weeks, and there’s a million different hotel options. And I might get lost. I might spend hours in the past, but instead I shortlist them for me. Give me five. ⁓ And it helps me move past whatever might be getting me stuck. ⁓
The third one is stretch. So what are those things that it’s like the beyond the edge of your comfort zone, your growth opportunities where you need a little bit of a push and how can AI actually support you in that and maybe nudge you past that. And so for example, for me, I’m a
I like, I’m a big picture divergent thinker. have a lot of ideas. The discipline side of like content creation is a struggle. ⁓ Namely the word discipline, especially if I’m not operating within a structure, like an external structure when I’m inventing it myself. ⁓ So AI actually helps me.
like stay consistent, which is really good. Like it’s just re-info- it provides some level of like, it’s holding me accountable for those things, which is really good. And I think that might look different for different people. and then, amplify, unblock stretch. ⁓ and free. Like what are the things that you really hate to do and can AI set you free from that? for me, little admin tasks, are there ways to automate them? Can I set up a system now?
Jason Gorman (35:38)
Hmm. Hmm.
Mmm.
Heather Xu (36:05)
I don’t have to think about some things anymore. AI kind of hold that for me and surface them up when I need to.
Jason Gorman (36:13)
There’s a Yes, definitely different for everybody and actually you know all of this so much the last couple things that we talked about which is kind of theme of human agency and creating room and opportunity to kind of grow ourselves to make sure that we are part of our process that we are even
Heather Xu (36:13)
I think it looks different for everybody though.
Jason Gorman (36:39)
even personally kind of unlocking things for ourselves that may be hard to unlock without the use of AI or more labor to unlock without the use of AI. ⁓ So this idea of human agency as an enduring skill to cultivate, I’m curious if you could just talk a little bit more about this notion of
of human agency and then what are practical things that people can do to cultivate this in ourselves? What are some day-to-day things that you would recommend?
Heather Xu (37:21)
It’s a topic I’ve been thinking about for…
like quite a few months now and I’m really starting to see some articles surface about it, which is interesting. But I think about agency as like our inner drive ⁓ to be proactive, kind of take the reins and have some sense of control in our direction and in the work that we’re doing. And underneath agency, I mean, there’s some psychological, psychology models. ⁓
that are really relevant here, but underneath it I think is this idea of
of intrinsic motivation as well ⁓ as being something that’s propelling us. And then I think agency is also really critical to our wellbeing. So if we’re operating in a place where we, like our sense of agency is deteriorating, I think our wellbeing is going to deteriorate as well. ⁓ So that’s how I would define agency. And I think why it’s important now more than ever is one, I mean, if we think about our junior team members, they’re kind of
entering into a workplace environment, the apprenticeship ⁓ scaffolding is eroding. So they’re not getting the experiences that help them build capabilities. their managers are probably swirling and overwhelmed. And so they’re probably getting messages like, be proactive, step up, like, own your, I like to say, own your impact, you know? ⁓
But how do you do that if you don’t have a compass or a North Star or the muscles to help you navigate the uncertainty? And then I also think as we’re delegating more decisions to AI, our own sense as leaders, our own sense of agency is, and I mean, I was just reading about this today in a HBR article that like leaders feel less agency and it’s causing real leadership crisis.
so, but it’s great that we, that we notice that agency is something that we need. I took it one step further to really think about like, so what does it mean to cultivate agency and how do you start and how, how might that look different for different people?
⁓ because I think of like conventional wisdom of like, just go build something, just go do it. And I, that never resonated to me, like with me that, that driver mentality. ⁓ and so I, I’ve been thinking a lot about like the different energies or archetypes of agency, things like, you know, you might be a disciplined builder, ⁓ who likes to construct things like, like really kind of source things, build a plan, see it through execution and, in a steady
Jason Gorman (39:53)
Hmm.
Heather Xu (40:12)
way. ⁓ You might be a curious explorer that’s out there and covering ideas. ⁓ You might be a systemic sense maker, which I relate to a lot, which is like I’m exploring, but I’m also trying to bring them into a cohesive whole before I can create. And you might be somebody who thrives on social collaboration. And so you need like the human to human interaction in order to kind of get your first step going and get momentum going.
Or you might be a rapid creator. So somebody who needs to like, just do things, put it and put it out there, see how it lands, see what people think.
Those are different, I think, energies for taking the first step. Like really your work probably involves all of them, but what’s your starting point? And if you can understand yourself, think about your strengths, think about what you gravitate to, taking that first step is usually enough to just get juices flowing and momentum going, where you can get out of a sense of paralysis.
Jason Gorman (40:55)
Mm-hmm.
I, you actually basically answered my last question already in that answer, but I’m gonna try to tweak it a little bit, which was, ⁓ so you kind of described what the individual can do. What are some things that the organization can do to help set the stage to allow people to take those steps?
Heather Xu (41:21)
Thank
I think we have to move toward more agile work environments. You know, I’ve been a part of multi-year transformations where you kind of move from vision to, you know, design to build to execution and to ramping up people in the meantime. so two years after you start, have your, your, your, your system and your people ready to roll in a new way of working that.
that’s changed, that, we ⁓ kind of come in full circle to the pace gap. Like that’s the issue is you might put out a workflow, that workflow that we created for the learning function with AI, like it’s irrelevant now. That was a year ago. Everything’s changed. I would do it completely different today. So I think one…
Jason Gorman (42:19)
Yeah. Yeah. Yeah.
Heather Xu (42:30)
I mean, it’s tough. I think this is the question that doesn’t have an answer yet, but we need to figure out where people do have autonomy and decision making. Maybe there’s more, like we have to figure out where people get to lean into the process and aren’t just managing their agents. ⁓ And we need to really think about the cognitive burden.
that is being placed on people as work is intensifying as their responsibilities that no longer do they have as many like low cognitive tasks that allowed their mind to rest. It’s, know, just elevating every everything is now important and people are
There’s the term brain fry, people are getting brain fry from too much AI. And if we don’t recognize humans as part of the equation that we’re solving for with new technology, we’re going to erode the skills and the value that make our organizations competitive.
And so I don’t know if there’s a golden ticket answer to the question, but I think the first step is thinking about humans in the part of the process.
Jason Gorman (43:53)
I heard somebody talking about using a power user, a coder, using Claude and having the experience of think about what it is that they need to get done. They go to Claude, they explain it, and then they go for a walk. And then when they come back, Claude is done. Claude has done the work, right? And then they review the work. They think about it, right? They decide the next thing that needs to happen.
They have a conversation with Claude about the Claude gets started, takes another bit of time to go for another walk. And I do, this kind of sounds like, I mean, that seems very funny and probably not realistic for most people. But I think in your answer, that’s kind what I’m hearing. We as humans, we need things like walks to help us think more clearly, to be more productive. That it can’t just be work, work, work. It cannot all be optimized all the time.
And that as an organization, when we’re thinking about how to do this in the right way, how do we make space to allow people to have those kinds of things, right? ⁓ So I ⁓ appreciate that. And I hope that there is not just like a ⁓ rush into more and more productivity ⁓ and more and more efficiency, because I just don’t.
I just don’t, can’t imagine that working out the right way. And it sounds like that’s exactly what these frameworks of yours and how you’ve described agency is kind of pointing to. There’s a real danger for all of us.
⁓ There’s one last question. It has nothing to do with AI. It has nothing to do with workforce. No, no, I have not. So this comes from Simon Greenwold, who is a giant audiophile. And so I was unsurprised by his question. So his question for you is, what is your desert island album? Meaning you can only listen to one album for the rest of your life. What would it be?
Heather Xu (45:33)
Well, yeah, you haven’t heard my answer yet. No.
So this is where I am going to bring AI into the answer because I’ve been taking some of my inspirational contemplative writing and my poetry and actually using an AI tool to turn it into songs.
Jason Gorman (46:00)
all right.
Wow.
Heather Xu (46:13)
And ⁓ it’s been really powerful to hear my words that I wrote myself back before AI, like presented to me in a different way. ⁓ So I would make an album of that and I would take it with me because that writing is really empowering. And I think I would do that on a desert island.
Jason Gorman (46:24)
Hmm.
Thank you for giving me a wonderful, surprising answer at the end of this conversation and so many amazing insights through the course of this conversation. I really appreciate it. Thank you so much for coming on to the show, Heather.
Heather Xu (46:52)
I’ve had so much fun. This has been great. I love it. It’s good to have some human to human interaction, even if it’s remote.
Jason Gorman (46:54)
It has been great.
Indeed,
Absolutely. It was kind of the point. We just need to be talking to each other now. You said that there’s like the value of connection is going up. We’re going to need more of it in the future. And I think the AI is going to be, it’s going to be tempting to just like, just be locked in conversation with AI, but we can’t, we gotta do this stuff. thanks. Thanks for doing this.
Heather Xu (47:19)
See you.
Thank you so much, Jason. I’ve loved it.