I'm Alison Beard.
And I'm Adi Ignatius, and this is the HBR IdeaCast.
Adi, do you remember back in 2022 when we were thinking about our first all-day live virtual leadership conference and we really didn't know what we were doing at all?
But we hired a freelance consultant and we pulled together people from around the company who were experts in various things.
You know us in editorial, others in marketing, product development design, production.
And then we actually put on the event a few months later.
Yeah, and it went really well.
I remember that extremely well, and it's one of my best experiences at HPR.
And it was partly because it was an all-volunteer team, pretty much of people who had other jobs but who came together for this really important task and we pulled it off.
Yeah.
So that is what our guest today would call a flash team.
And she has studied groups like this who are pulled together very quickly.
They don't normally work together and they develop a camaraderie and teamwork.
They collaborate to put out a product or develop an idea more quickly than one would normally be able to in an organization.
You can do this within organizations, like we did at HBR, but you can also.
Now, thanks to the internet, thanks to online labor markets, you can search all over the world for the right kind of expertise you need.
And you can also get really granular about the type of skills that you're looking for.
And what's more, artificial intelligence can help you do all of this.
So it's interesting because when we formed our own ad hoc events team, we sort of wondered look, we probably need to set up a proper permanent team, right?
So I'd be really interested in the calculation.
You know which is more effective, which is more cost efficient, whether you have the full-time team or whether you just do it on a kind of ad hoc basis, as needed.
So my guest today is Melissa Valentine.
She's a professor at Stanford University.
And I think her argument is that, because work is changing so quickly and you really don't know the skills that you're going to need tomorrow, it actually is more cost effective, more time efficient to build a business around flash teams.
It's not necessarily outsourcing the work, but it's having a really good idea of the talent that you can access outside and inside your organization, what skills they have, and then staff every single project really appropriately.
So I learned a lot.
I think it's something that we might want to experiment with more at HBR.
She, in addition to being a professor, is a senior fellow at Stanford's Institute for Human-Centered Artificial Intelligence.
She's co-author with Michael Bernstein of the book Flash Teams Leading the Future of AI-enhanced on-demand work.
What exactly differentiates the kind of flash team that you're talking about from an ad hoc project team that might be put together within an organization?
So teams are recognized to be super valuable because you get lots of smart people, they get together, they collaborate.
So what's different about a flash team?
Is we started to recognize how technology can give teams different properties, different superpowers?
So one of the technologies is actually just the internet.
And because of the internet, it means that there's like millions of people all over the world who are expert in something that you might need.
So with the internet or with like online labor markets like Upwork, The search for who's available to help join your team is like so much broader.
Two is we're all a lot more comfortable with remote work now.
So you can search from millions of people online.
You can pull them together in a remote way.
And then AI is also available now to sort of help us do the things that make a good team like find people, bring them together, optimize the team staffing, help them talk to each other, help manage the team.
It is a project team, but it's like now with superpowers because of technology.
Why do you think it's important for both organizations as a whole and then individual managers to learn how to source and manage teams in this way right now?
Well, it's funny because when I wrote the Flash Teams book, it wasn't like the book the world was necessarily asking for.
But, like now that it's out, I'm seeing it's the book that the world needs because it's a moment where people are really worried about like jobs.
Will we have jobs in the future?
And I think what the Flash Teams idea really shows is like all of us are entrepreneurs.
Even if you're like in a big company, not necessarily at a startup, all of us have ideas.
All of us can see here's a problem that could be solved.
Here's an idea that could be brought to life.
And so, with flash teams, the barrier to what it takes to bring your idea to life or to solve a problem is so much lower.
I could just get a team started.
I could get a team together in minutes.
And with all of these tools, I can be much more effective in solving problems.
And so the idea then is that there's huge both time savings and cost savings in assembling teams this way.
I can get it cheaper and I can get the project done faster.
It is definitely time savings.
There's possibility for real cost savings.
And I think there's a hope that it can also be a lot more effective.
There's a lot of tools that can help you get a better group together, like who's going to have more likelihood of success.
Or when you're in flight, you'll have like all these tools to kind of help see.
You know how are things going.
What's the likelihood we're going to be successful?
There's also like definitely some performance help that comes with all of these tools.
We really wanted people to realize how accessible this idea is.
So, even if you're not somebody who's super tech savvy or someone who's really focused on software right now, sort of the mindset that we offer in the book is first, recognizing we call it like experts everywhere, all the time.
It is really easy to find people to help.
And so we try to help people just kind of think about it in terms of problems and tools.
It might help for listeners to just hear an example of a Flash team that tackled a project that might traditionally have been handled within an organization you know, drawing on people in different offices, maybe from that company, but instead was outsourced to this global network that you're talking about and able to accomplish those goals more quickly and more cost-effectively.
So we actually worked with a doctor who worked in an emergency room.
So he was one of our early users and he had this idea for an app in the hospital that was going to really help patients.
So his idea was that when ambulances were approaching the hospital they could have like a GPS kind of signal that was displayed on the emergency room wall.
That was like showing when the ambulance would show up.
Because what he had noticed is that ambulances would show up and then people would gather all the right doctors and the right equipment and stuff.
And he was like, what if we use that 10 minutes to get ready?
So that was his simple idea of what if we could build that app.
Typically the way that would work in an organization is he would pitch it to his boss.
Who might pitch it to his boss?
They would finally talk to IT and then IT would be like, that's not on the roadmap.
Right.
So it would never get built because it's, even though it's a very simple idea and would actually help.
It's just not an organizational priority necessarily.
Yeah.
So this doctor, he was super entrepreneurial and he wanted to just build this app.
So he did convene a flash team using our approach and like using our software.
And within six weeks, they had built the app, a functional prototype.
They'd actually like tested it with some like ambulance drivers and the doctor's local community.
And it was pulled together with 30 experts from all over the world.
I think that he ended up having like a CTO we would call her who was like an engineer in Bangladesh.
And she sort of oversaw the whole thing.
But what was really cool about that one is just how quickly it came together.
He had the idea.
He started using this approach.
And then within six weeks, he has this functional app that he was taking around to VCs and stuff.
But it is something that theoretically he could have pitched to maybe only his manager to take out of some discretionary budget that they might have had at the end of the year.
Have you seen that approach work within organizations or companies where, because it's sort of a smaller investment, it's easier to get approval for an individual manager to create a flash team to work on a particular project that they're excited about?
Yeah.
Yeah.
You've got totally the right spirit with it.
Since then, we've seen so many examples across so many companies where people are doing this.
So that's a great example.
Like maybe a manager has like a smaller fund that could be used for it.
But across all of the different examples.
It can be like a big company, like a Google, for example, like we met people who were doing essentially flash teams within Google.
And so in that case, you're not searching across the whole internet to find someone to come build something with you, right?
You're just looking across the whole Google workforce.
And then using some of the tools of optimizing the team staffing, figuring out who's available when, and all of that is algorithmic.
You can use AI to do it better.
You can use large language models to do it easily.
Do you see this as a tool or process or strategy that is sort of for technology forward companies, or does it work really across industries?
One of the things that has been most exciting about Flash Teams is when we present this to groups.
Almost everybody has an idea of like what they would do with a flash team.
Somebody was an employee at the Minneapolis DMV and they just like sent a LinkedIn DM and was like we could do like a process redesign.
We could fix this dashboard or whatever.
There was a group of mental health experts in Australia who were like we could do like flash teams of like mental health triage, because it's like a huge continent.
So like, how do they find each other?
Yeah.
And it really highlights kind of the spirit of the book is that we're all entrepreneurs.
We are all problem solvers.
The takeoff so far has certainly been more within software, but there's examples of flash teams, of business models.
And there's a company called Catalant that does flash teams of like MBAs, basically.
There's a company called Chief Outsider and they'll pull together a flash team that will make like brand or marketing for you.
So it's really taken off across different industries.
So I want to dig into some of the practicalities of doing this.
And let's talk about recruitment first.
You know, the idea of assembling your team.
We've talked about those online labor markets like...
TaskRabbit and Upwork, and I know there are different ones around the world.
How exactly do you start the process of saying I have this idea.
I don't even know what types of people I need to help me build it.
Which market should I go to?
How do I advertise the project?
Just walk me through that process.
Yeah.
So, having done it a bunch, now we sort of like instinctively and intuitively kind of know how to think about looking for people to help.
Recognizing the need for help, there's probably an expert who could help me with this.
There's probably somebody who knows how to do this.
And then starting to kind of formulate the first question that you have.
In this day and age, the cool thing is you can even take that to a large language model.
Who would be the first team member if I wanted to make this software app?
And I mean, I do that with my exec ed classes now.
I'll ask people like, what problem would you solve and who's the first expert?
And honestly, just asking that question to something like ChatGPT gets you a pretty good sense of the expertise you're looking for.
The next thing that we tend to do is look to something like Upwork like a large, you know, like an online labor market.
And what I like to show people is that Upwork will have these really helpful profiles that tell you a lot about who the person is.
And sometimes it will even have a little area that says available now.
And then you can just interview that person for like 15 minutes.
You pay them for your time and you just kind of get an idea.
You say, you know, I want to be the first generative AI use case in my company.
What should I do?
And you find the right expert and they can give you a sense of like, spend 15 minutes talking to them and you have like a decent sense of it.
And are there really enough people who are making themselves available for gig work?
They aren't full-time employees at other organizations to be your expert on any topic.
I like to do a fun exercise with my undergrads where I'll tell them you need a professionally designed team logo before class is over go.
And they all kind of start being like, you know, there's no way and it's like stressful.
But in every case they've like found somebody within 80 minutes.
Like every team, every undergrad has like found somebody.
What I'm trying to illustrate with that is yes, something like Upwork has millions of people who are available.
It's a very kind of like live dynamic network.
So finding people to talk to and who want to help out is it is it's very fast.
It's very dynamic.
For managers who are trying to assemble this sort of crack team of experts from around the world.
What are some of the biggest challenges that you see them run into and how do they overcome them?
One of the chapters in the book, we have a phrase that I just cannot underline enough, which is you have to get the launch right.
So if you put a lot of thought into the launch, things are gonna go much better.
So things like what software are we gonna work on?
What are some of the team norms?
Like what is some of the team structure?
Just having that kind of like place for people to land.
So they show up with all of this entrepreneurial energy to like help solve a problem.
But it's like helping them know how to get started from the beginning and not spending a lot of time being like are we on Teams?
Are we on Google Meet?
Are we on Slack or whatever it is?
It's like you like deploy kind of this suite of communication tools and like the knowledge repository.
You just have that going at the start and then people get together and they just go.
And are there ever problems, you know, in that you've hired the wrong person and have to get rid of them and replace them, or a person's skills don't really match the project you're working for?
If a manager is like really kind of leaning into the Flash team model, there's like two ways that I think they have to be ready to adapt fast.
One is your plan is certainly not right.
Your plan will certainly need to adapt like the moment you get there.
So you spend a lot of time kind of like getting the launch ready.
You push play and then yeah, somebody shows up and they don't have the skillset, or like one of the tools is down, or like the internet is out somewhere.
And so like from day one, you have to be ready to adapt.
But then the second piece is because you're hiring so fast, sometimes it's not the right fit.
Maybe it's the wrong skill set.
Maybe it's the wrong fit.
But when you're going really fast and kind of like doing like a six-week sprint, when just everyone's trying to get things together, we sort of have this.
You know, hire fast, fire fast.
Now I do need to do a big asterisk here.
We have a whole chapter about like being a good manager, being good to people.
So like the gig economy can be exploitative and that's like certainly not the intent here.
The hire fast, fire fast is not intended to be exploitative.
It's intended to be.
People really believe in the idea and you've given them a chance to like, really help build something that you also really believe in.
That like the group just coheres.
It's been really fun to like build and like, see how much solidarity can develop in these kinds of groups when people are just really building together.
So if somebody is like out of sync with the group, it doesn't feel good to them.
It doesn't feel good to the group.
So sort of just like recognizing where the energy is really flowing and where things are really working and then just not letting something introduce a lot of dysfunction into a group.
That's pretty task focused.
Don't let things fester.
Yeah.
Taking a step back.
These online labor markets also make it very easy to sort of enter into contractual relationships with people such that you've agreed to pay them for a certain amount of hours or for a certain block of work.
So that is handled up front.
Yeah, so that's one of the technological capabilities online labor markets will give you.
They will do the contracting.
They'll know things about like the local labor laws.
They'll handle like the IP or like the NDA and stuff like that.
So a lot of that is handled in the labor market, which makes it really fast to just get started.
We also see like the application of this idea in just large companies already.
I mentioned a few of them.
And there you already have like full time employees who are already on all the platforms.
So in that case, you're just pulling together a flash team internally.
Then that's also all kind of taken care of already.
Yeah.
And then you know, as a manager in that launch process that's so critically important.
You and your co-author talk about establishing clear roles right at the start, which is sort of advice for any good project manager, right?
Everyone needs to know what they're doing and have sort of clear assignments and no overlaps.
And when there is an overlap it's acknowledged and just sort of set out that framework before everyone starts.
Yeah, Flash Teams really brought our attention to the idea of role clarity, which is as you're saying.
I mean, that's a management classic, right?
Always like a good practice to have.
And the reason it's so important with Flash Teams is is it's a group of strangers.
So they're coming together.
They've never met before.
And it's like already instantly clear who has what expertise, what they're responsible for.
So they can immediately start working as a team.
And there are a lot of good examples of that happening in different places, like on film crews or in hospitals.
So those are places where you have temporary teams like this.
But I will say the other thing that's super helpful about role clarity, because we're talking about not every manager knows exactly who to hire.
If you've never done software before, maybe you don't know what is the difference between UI user interface and UX user research.
But you can just ask somebody that, like they already know, like they have a lot of role clarity so they can start to teach you who they are, what their responsibility is, the kinds of decisions you'll make.
And they'll just help you.
They sort of teach you their roles.
You become kind of like the general contractor and then they teach you what their expertise is.
They teach each other what their expertise is.
So that kind of role clarity just makes the whole thing work together really easily.
Yeah, I love the analogy of a general contractor.
You're the same way a person would bring in tradespeople you know to do tile or carpentry or plumbing or electrical work.
You know, you're doing the same thing for the project that you're working on.
It's just, instead of them coming to a physical place, they're coming to your sort of like online work hub to do that.
Yeah, exactly.
Yeah, exactly.
Yeah.
Aside from, you know, I guess, curiosity and adaptability, which you've addressed, are there any other sort of key skills or attributes that you think really successful Flash team managers have?
When you have energizing vision of what success is, if there's somebody who's there, who's like this is what a solved problem looks like, or this is what it means for this idea to come to life, then that person is often like playing a really key, like integrating function.
A risk with a flash team.
You know people aren't together.
They might be on different time zones, it's just that the effort gets fragmented.
But if you have somebody who has that kind of like driving energy and they're very passionate about solving this problem, then they're going to play this integration role.
They're going to look at how all the pieces fit together.
They're going to troubleshoot.
They're going to problem solve.
So yeah, there's a bunch of studies that we've done over the last 10 years studying flash teams.
And I think that that integrating manager role is really important.
Without that, the flash team is much less likely to succeed.
You also address the idea that they have to insist on documentation.
There was a funny anecdote where the project managers just kept saying to everyone on Slack yep, just look at the docs, look at the docs, look at the docs, because that's where everything is recorded, to make sure people stay on the same page, right?
That's totally true.
And that's actually the story of the same guy that I was talking about at the beginning.
And he's one of our favorite examples because he was like so patient throughout that.
And he did all of the like kind of back channeling where he was asking all of these questions.
And like when the answers weren't coming together, he just kind of stayed persistent.
You know, really playing that like integrating leadership part.
But you're totally right.
Like documentation is absolutely key because everyone is online.
And people in different time zones need to know what's happened while they've been gone.
You know, I've interviewed leaders of all remote organizations and they say that's absolutely critical because otherwise, you know, the lack of coordination creates overlap work, miscommunication.
You know just a lot of mistakes that wouldn't otherwise happen.
Yeah, totally.
It is foundational to remote work.
And I'm not sure if this is where we're headed.
The fun thing about it, I think, is that becomes like the data source.
That then, is like feeding all of the AI tools that are going to help with more things like you know, people knowing what they're supposed to do, like what decisions have been made, and so forth.
It becomes like the trail of decisions and then also the data set that the AI will draw on to like help people coordinate better.
Yeah.
So talk about that, how you incorporate AI into designing a better, you know, not just team because of the people you've recruited or AI agents you've recruited, but because of the way the team is working together.
I was saying like the launch is everything and sort of the individual onboarding similarly needs to be very effective.
When we first did a Flash team, we were relying on like 2017 UI.
So it's, you know, like a description of what your task is.
But now the onboarding is using LLMs.
It can be like very dynamic, very interactive.
You click into the Flash team, you're brought to the platform where everybody is and you're like, welcomed by your own like onboarding agent who, like tells you what you're doing and like what platform we're on, who's your manager, who's your teammates, what everybody's doing.
I mean, and that is happening in like non-Flash teams to like larger companies, that kind of like onboarding, very interactive onboarding.
Are there tools that project managers or Flash team managers can use to sort of foster the kind of teamwork you're talking about the launch and then the coordination?
One of the examples that we have is a bot that can kind of help you decide how to experiment with your team design.
So let's say that you started out and you decided to launch your team pretty decentralized.
And then over time, you're like, we're not going fast enough.
I think maybe we could try like a more centralized or hierarchical team structure or whatever.
You could develop a spreadsheet using even ChatGPT that will recommend to you when's the right time to try a redesign.
And then how do you measure how the redesign went?
And then how soon can you try another redesign?
And then how do you tell it how it went?
So that kind of feedback loop where the manager tries something, collects information on how it went, tries something else.
I mean, that's available to anybody.
And we really talk about just the capabilities and the kind of easy software hacks that are in reach for anybody.
And you can go pretty high tech with some custom software as well.
So our lab does a lot of proof of concept software.
We're not doing any like enterprise grade software that like somebody should come and like buy.
If they want to support our research, they can get.
It's certainly available.
But ours is very like proof of concept.
We'll just be like, can you convene a team on demand hiring everybody in like 15 minutes?
And then we prove it's possible.
Are there any other ways that you see AI enabling more effective flash teams?
Well, a fun one that we've been talking about a lot lately is AI simulations.
Yeah. and my co-author will talk about them as like what-if machines.
So, using like AI simulations, you can sort of take the five people you're considering hiring, input their roles, some of their experience and you can simulate like what happens if this team is given this task.
What happens if this team has this structure and you can sort of simulate whether it's going to work.
And then that helps you kind of like reverse engineer and like design in a certain way.
That's been a pretty fun one.
And what about incentivizing teamwork?
Because you're basically collecting a bunch of gig workers who are independent contractors and they want to get their work done.
And ideally, I'm sure they want the project goals to be met, but once their work is done and they're paid, they don't necessarily need to see it through and help their colleagues who are on this flash team with them.
So how do you make sure that people really do feel like a team?
Yeah, I think that the nature of remote work can kind of lend itself to that sort of fragmentation.
So part of the launch, I think, is thinking about where you need people to feel like they're together, where you need cohesion, and thinking about if there's a way to, as you said like, set up the you know, the incentives or the accountability so that it's like a group level incentive or like a group level accountability.
And looking at all your experiments, looking at scenarios where this has been accomplished in companies, do you think that eventually it could be applied to almost every kind of work that's happening within our organization, provided people can work remotely to do it?
I think the vision that we have with Flash Teams kind of the vision of it.
You know, there's all of this kind of like team science and organization science.
There's like a recognition of like how we're effective together as a team or how we're effective together as an organization.
And so the last part of the book really like thinks about if you were to use AI to sort of augment team science, or use AI to kind of augment organization science, then it means your teams are going to be much more effective.
Your teams are going to act and coordinate much more intelligently and also your organizations.
The part where you like get people together on the fly is important.
And recognizing that when they're working online, you're getting all of this data, and the data can then be used as sort of like a team science exercise.
I think that's the vision of like where it could head.
And so I think that yeah, for any industry, certainly they could be using data to like, structure and design their teams more effectively.
Yeah.
So if you're advising a C-suite leader who wants to sort of bring the culture or strategy of flash teams to the group, what do you recommend they do?
How do they introduce the idea?
I think the first step to me would be looking at what data do you have on your workforce?
So, for example, do you already have a database that tells you everybody's skills, everybody's availability, their roles, their expertise?
Once you start to see a data set of a large group of people and you start to see how many skills and how much expertise there is, and then you can start to think about more dynamically matching people to problems.
That tends to be the kind of missing step that I see in a lot of companies.
Yeah, absolutely.
And where the gaps are, so that you do then turn to an online labor market to sort of fill in people as needed.
And then maybe one of those people who helps you for several projects is a good potential hire because you've already worked with them.
Yeah, exactly.
Yeah.
Yeah.
So what about the broader implications here, though?
You know, because you're recommending that organizations try to move into this flash era, which inherently means that they have fewer full time employees, because they're pulling together teams ad hoc with the skills they need at the moment.
But what does that mean for workers?
You know in terms of job stability and benefits.
You know like health care and retirement.
How does that change the labor market?
I think there is a real kind of need and opportunity for kind of society and leaders as a whole to look at, You know, this question of like employment and look at all of the kind of social safety net that goes with employment at present.
Society as a whole is like thinking about gig work, full-time employee, you know, temporary employment and all of that stuff.
I think that the Flash team's perspective I think we've tried to like really decouple that from like full-time employment.
You can take, for example like, even just like a job like Stanford, like I do, like a different project every year, but I have full time employment.
So I think the pieces of Flash teams that we see as visionary and we see as useful for society is solving problems better and like using data, as I was saying, to sort of like structure your team better.
However, it doesn't need to play out in a way that is like precarious gig work.
As you said, you know, it can be done within organizations.
If you have a large global organization you can staff projects more dynamically almost in the way that consultancies always have, even though they've sort of required physical presence generally, to send client teams online to a particular location or destination or group of C-suite executives that they're working with.
It's sort of like that model for any kind of work.
Exactly.
Well Melissa, thank you so much for talking to me today and having me wrap my head around this whole idea.
I hope I get a chance to experiment with a Flash team soon.
Thanks for having me on.
That was Melissa Valentine, an associate professor at Stanford University and a senior fellow at its Institute for Human-Centered Artificial Intelligence.
She's the co-author with Michael Bernstein of Flash Teams.
If you found this episode helpful, share it with a colleague and be sure to subscribe and rate IdeaCasts in Apple Podcasts, Spotify or wherever you listen.
And I'll see you next time.
Thanks to our team Senior Producer Mary Du, Audio Product Manager Ian Fox and Senior Production Specialist Rob Eckhart.
And thanks to you for listening to the HBR IdeaCast.
We'll be back with a new episode on Tuesday.
I'm Alison Beard.
