 |
|
|
|||||
Below, coauthors Melissa Valentine and Michael Bernstein share five key insights from their new book, Flash Teams: Leading the Future of AI-Enhanced, On-Demand Work. Melissa is an associate professor of management science at Stanford University, where she codirects the Center for Work, Technology, and Organization. Michael is an associate professor of computer science at Stanford, where he is a Bass University Fellow. Both have had their work featured in major publications, including The New York Times and Wired. Whats the big idea? Have you ever wished that you could assemble your version of the Avengers at work? Thats basically what it means to build a Flash Team. Bringing together the right set of experts at exactly the right time to tackle a tough, important job has become a realistic, repeatable goal for leaders todayunlocked by powerful new technological tools that enhance organizational strategy. Listen to the audio version of this Book Biteread by Melissa and Michaelbelow, or in the Next Big Idea App. 1. There are experts everywhere, all the time One of the biggest mindset shifts of flash teams is recognizing that expertise is abundant. Managers have been trained to think that hiring an expert takes weeks of job postings, interviews, and approvals. But whatever expertise you need, you can probably access it in minutes, not months. A founder of a $35 million start-up told us that he had a client who needed to reimagine how to sell a beloved toy truck after their retail stores shut down. Using the flash teams approach, he quickly found a former McKinsey partner in retail, someone from Toys R Us corporate development, and a supply-side expert from Amazon. They had never met before, but they delivered so well that the client rehired them to manage execution. Weve demonstrated that same speed in our classrooms. Ive asked students to hire a professional designer from Upwork and get a finished team logo in under 80 minutes. Every time, theyve done it. Technology lowers the transaction cost of finding, vetting, and convening experts. Leaders can stop assuming that talent is a bottleneck. Once you recognize expert abundance, you can start designing projects differently: taking on bolder challenges, experimenting faster, and pulling in expertise at the moment its needed. This shifts managers from fearing scarce or hard-to-find talent to orchestrating abundant talent. This is possible because modern online labor markets and digital platforms are the infrastructure. They provide access to millions of professionals worldwide, reputation systems that help you assess quality, and fast contracting and payment systems that remove friction. Technology lowers the transaction cost of finding, vetting, and convening experts, so leaders can act on the abundance thats already available. We live in an economy of expert abundance. With the right mindset and tools, you can assemble the right people at exactly the right time. 2. AI can help you design teams and organizations The amazing thing about having flash teams is that, because they are created using computing, you suddenly have the ability for AI to help design your organization: how to staff, how to work together, and when to adapt. As a result, we can solve lots of problems that are organizational and managerial blind spots. Were not talking about theory here, but about practical dials that managers usually leave on the table. AI can influence how our teams and organizations are structured and function: How should this team be collaborating? Should we have horizontal leadership or enforce a steep hierarchy? Who should even be on this team, or are right for this project? Many of the decisions needed to build an effective team can be supported by AI insights. As people, we tend to under-explore. We dont try enough options. We try a couple different things, see what seems to work, and then we say, Yeah, seems good. But this is how we fall into a rut. With AI plus flash teams, you instrument the basics and give the system permission to propose small experiments, such as: Try a Directly Responsible Individual (DRI) for decisions this week. Rotate one member for fresh eyes. Shorten stand-ups and add a mid-week asynchronous check. If it produces improvements, the AI learns to keep it; if it doesnt help, the AI might toss it. As people join or roll off, the recommendations adapt. These kinds of things give us managerial superpowers. AI-enhanced flash teams can make this possible. 3. Management classics are still classicjust reimagined. In some ways, flash teams sound like something brand newon-demand experts, AI tools, dynamic org chartsbut the management classics are still classics. They just look a little different in this new world. Take project management. In our research, we studied hundreds of flash teams. The best teams didnt succeed just because they had the right experts. They succeeded because a team manager made sure the pieces came together: synchronizing handoffs, keeping information transparent, and making sure the clients vision stayed connected to the teams daily work. One engineer told us bluntly, The PM (project manager) makes or breaks the team. Or leadership. In one of our experiments, when a client suddenly changed requirements mid-project, the teams that thrived werent those with the flashiest experts. They were the ones where a leader stepped in to integrate different perspectives, rebalance priorities, and help the group adapt quickly. Leadershipthe ability to inspire, coordinate, and adaptstill matters, maybe more than ever. Flash teams give new life to timeless management skills. And integration. Even with great role clarity, unexpected complexity shows up every day. Someone doesnt deliver, or two roles conflict, or the work comes in messy. Thats the residual complexity that only managers can resolve. In one case, a team writing poems for a card game had beautiful but mismatched outputs. They quickly elevated one person into the role of Chief Poetry Officer for a dayjust long enough to integrate the parts into a coherent whole. Thats hierarchy reimagined: temporary, lightweight, but crucial. With flash teams, digital tools support classic management functions. Platforms like Slack or project dashboards give managers real-time visibility across the whole team. AI-enabled systems can help leaders spot when handoffs are slipping, recommend worflow adjustments, or even simulate different team configurations before you commit. The human arts of leadership, integration, and project management get amplified. Flash teams give new life to timeless management skills. The tools may be modern, but the fundamentalsclear leadership, good coordination, thoughtful integrationare still what make or break a team. 4. AI org simulations and organizational what-ifs. Flash teams open this incredible opportunity to have a what-if machine: What if we organized the team this way? Would the team work better or worse? What if we brought this person onto the teamwould it help? What if we split up into two smaller units? Would we move faster and make better decisions? Imagine being a manager and getting a fast, concrete preview of what might happen: what could go wrong, whats likely to improve, and what might get worse. This is possible through the clever application of large language models, such as ChatGPT. We can use this new generation of AIs to create lightweight simulations of your organization. Imagine digital twins: little digital copies of everyone on your team that act and behave roughly the same way that they do. With those simulations, you could put the digital twin of your team or org into different configurationsreconfigure the team, change collaboration rules, and moreand see if they coordinate more smoothly. This is possible through generative agents. These are AI agents that simulate people based on a bit of knowledge about them. Maybe you run a little interview with everyone in the team and use that to create a digital twin of them, or maybe everyone agrees to use a slice of your historical Slack or email to create digital twins of your team. Once you have that, your team can become this dynamic, queryable object: you ask a question, run a quick scenario, and watch how it plays out. Its a rapid, plausible rehearsala what if. This is possible through the clever application of large language models, such as ChatGPT. In this way, we can also catch early warning signs for a team. It allows leaders to flag whether a team is likely to fractureto stop wanting to work togetherby using about 60 to 90 seconds of their chat. A tiny glimpse into how people communicate and coordinate can reveal surprisingly strong signals. Suddenly, we can predict whether this team will work as great long-term partners, or if we should reconsider them. Its almost like organizational speed dating. Imagine having the superpower to create organizational what ifs. It gives you this amazing managerial sandbox. Flash teams turn your org into a safe, queryable what-if machine, so you can prototype structure before you commit to it. 5. You already have a flash teams toolbox You dont need to have a PhD in artificial intelligence to do these things. You can do it today, without any custom software. All you need is the idea and access to a modern large language model like ChatGPT. It turns out that everything we had to spend months coding manually can be generated on the fly by an LLM if you can just be specific about what you need. AI can help you design or refine your team. One option is to get an advanced degree in computer science and learn about networks of multi-armed bandits, then build it internally. But the other option is to just keep a spreadsheet where youve been keeping track of how things are going, and the management decisions youve been making so far. Input that into GPT-5, ask it to implement this approach, and it will do all the math for you. Enjoy our full library of Book Bitesread by the authors!in the Next Big Idea app. This article originally appeared in Next Big Idea Club magazine and is reprinted with permission.
Category:
E-Commerce
Meta allegedly stopped internal research on social medias impact on people after finding negative results, a court filing released Friday claims. The filing took place in a Northern California District Court, as a group of U.S. state attorneys general, school districts, and parents launched a suit against Meta, Google-owned YouTube, TikTok, and Snap. The court documents allege that Meta misled the public on the mental health risks to children and young adults who excessively use Facebook and Instagram, even though its research showed that the social media apps had demonstrated harm. “The company never publicly disclosed the results of its deactivation study,” the lawsuit says. “Instead, Meta lied to Congress about what it knew.” The research, code-named “Project Mercury,” took place in 2020. Meta scientists worked with survey firm Nielsen to see what impact deactivating Facebook had on people. According to internal documents, people who stopped using Facebook for a week reported lower feelings of depression, anxiety, loneliness, and social comparison.” According to the filings, instead of pursuing more research, Meta dropped the project, claiming that participants feedback was biased by “the result of the existing media narrative around the company. Politico reported that in a sealed deposition earlier this year, Metas employees expressed concern about the research’s findings. Oh my gosh, y’all. IG is a drug, Shayli Jimenez, a Meta senior researcher, is quoted as saying in internal documents. In response, another employee allegedly said, Were basically pushers.” The Politico story reported that Jimenez said in her deposition that the comments were made “sarcastically.” In a statement, Meta spokesperson Andy Stone said: “We strongly disagree” with the allegations, which rely on cherry-picked quotes and misinformed opinions in an attempt to present a deliberately misleading picture.” Stone continued: The full record will show that for over a decade, we have listened to parents, researched issues that matter most, and made real changes to protect teenslike introducing Teen Accounts with built-in protections and providing parents with controls to manage their teens experiences.” And in a series of posts on BlueSky, Stone also pushed back against the idea that Meta was trying to bury the results of the terminated study with Nielsen, noting that the study found that people who believed using Facebook was bad for them felt better when they stopped using it. “It makes intuitive sense, but it doesnt show anything about the actual effect of using the platform, Stone wrote. However, the latest uproar over Meta’s research is hardly the first time the company’s impact on children’s mental health has been questionedeven by its own employees. In 2021, former Facebook product manager Frances Haugen leaked hundreds of internal company documents to the government, which referenced risks to children. Haugen said the company’s leadership knows how to make Facebook and Instagram safer but refuses to “because they have put their astronomical profits before people.”A growing body of evidence, outside of the company’s own research, has long pointed to the harm that social media may have on children’s mental health. A 2019 study found that teens who spent more than three hours a day on social media may be at heightened risk for mental health problems, particularly internalizing problems.” Likewise, research shows that mental health disorders among today’s youth are at an all-time high and growing. In response to growing concern around children’s mental health, in a 2023 report, then-U.S. Surgeon General Vivek Murthy called on social media companies and policymakers to act, rather than to place the entire burden of limiting kids’ time on social media on parents.
Category:
E-Commerce
We have reached the moment white collar workers have feared for months. Has AI finally come for my job? Companies like Salesforce claim they need fewer human employees to do the work AI can tackle, after laying off thousands. Klarna claims the company was able to shrink its headcount by about 40%, in part because of AI. Duolingo said last spring it will stop using contractors for work that AI can handle. Overall, companies have announced a staggering 700,000 job cuts in the first five months of 2025, an 80% jump from the previous year. The irony is almost poetic. For years, the tech industry assumed robots would come for factory workers first. Amazon’s leaked documents once suggested the company could replace half a million warehouse jobs with automation. Instead, just weeks ago, Amazon laid off 14,000 middle managers while planning to hire 250,000 seasonal warehouse workers for the holidays. The AI revolution, it turns out, is hollowing out the corporate ladder before it touches the warehouse floor. The narrative around artificial intelligence and the job market is challenging for white-collar workers right now. Yet while Silicon Valley sends warnings over which desk jobs AI will consume next, we’re missing an equally important question about the future of AI: What about everyone else? The AI application bubble nobody’s talking about We are currently in an AI application bubble. The last few years of AI innovation have focused almost entirely on white-collar productivity: workplace efficiency tools, revenue-optimization platforms, and communication automation. Many of the major AI innovations from the past two years have been designed for someone working a 9-5 desk job from a laptop. Meanwhile, the people who make up 60% of the American workforce are stuck completing manual onboarding processes, sorting through countless texts to find the right shift, calling in when they need a shift swapped, clocking in on physical time clocks, logging in to web-only portals, and waiting for biweekly paychecks. Were talking about warehouse crews, janitors, delivery drivers, nurses, and game day parking attendants. These are the people who have been largely forgotten when it comes to how AI can transform their day-to-day jobs without risk of eliminating their roles. Every day, millions of shift-based workers keeping hospitals running, concerts staffed, and factories moving are dealing with broken, archaic systems. They’re waiting for shift confirmations, digging through emails for schedules, and calling managers just to ask, “When am I working next?” By focusing almost all of AIs potential on the white-collar economy, weve left out the workers who are irreplaceable. Building accessible, intuitive tools for non-tech-savvy users has the potential to narrow the global inequality gap while creating a more resilient foundation for technological progress and a more resilient economy. Only 40% of American workers say they have a “quality job” While office jobs dwindle, demand for human workers continues to grow. Restaurants need servers. Construction sites need carpenters. Hospitals need nurses. And in turn, the people doing these jobs need shift accessibility, work-life flexibility, and the ability to get paid quickly after shifts so they can continue to participate in the shift work economy and keep the world moving. The human cost of not having a better way to work is striking. A recent Gallup and Jobs for the Future study found that just 40% of U.S. workers have what they define as a “quality job.” The rest face unstable schedules, limited growth opportunities, and financial insecurity. Not because they lack motivation or work ethic, but because the systems that support frontline work haven’t kept pace with the demands of modern life. When workers play significant roles, have preschedules, and receive fair pay, they’re more engaged, more productive, and lead higher quality work lives. What we learned building technology for Uber drivers We know what’s possible when technology is actually designed for frontline workers, because we’ve lived it. While leading product development for the Uber for Drivers app, the two of us spent years focused on the driver experience. Drivers had to navigate complex processes: onboarding, completing background checks and vehicle inspections, selecting preferences, and receiving payments. Uber’s success was powered by a phenomenal self-service app that gave drivers the agency, control, and flexibility they needed in their lives. That experience taught us that technology has the potential to dramatically improve frontline work, and the emergence of AI gives us an opportunity to do that once again. Tools like smarter scheduling systems that account for worker preferences and availability, AI-powered training programs that adapt to individual learning styles, communication platforms that actually work for teams that don’t sit in front of computers all day, and predictive systems that can optimize logistics and reduce physical strain. The technology exists. The investment, however, is still lacking. Irreplaceable The Essential Economy that we are talking about includes sectors like construction, manufacturing, transportation, etc., and represents $7.5 trillion in output per year, which is 27% of Americas GDP, equating to 52 million jobs and two million businesses. If you were to add healthcare, retail, and all public servicesconsidered by many to be critical, hourly work sectors of the economythe size jumps to $12 trillion of GDP, 95 million jobs, and three million businesses. Without people to fill these roles, not only are essential services not being provided, but the US economy also suffers greatly. With each technical revolution, we’ve always seen that collaboration with the technology yields better results than we can without it, or it can without us. Instead, what if AI innovators asked, How can we use AI to make these jobs better, safer, and more productive while also making workers’ lives easier? Consider a warehouse worker trying to swap shifts to attend a child’s school event. In most facilities today, this involves a series of text messages, phone calls, and manual approvalsa process that might take days and often fails. AI could handle this in seconds, reaching out to available workers who have relevant experience and required certifications, sharing shift details, and filling the shift.The worker doesn’t lose their job; they gain flexibility and dignity. Consider a nurse who needs more hours as bills are adding up. He signs up with a new staffing agency so he can pick up extra shifts here and there. Today, onboarding entails manual back-and-forth with the agency and waiting days for assignments. AI can dramatically speed up his time to first shift, verifying his license instantly after he uploads it, offering digital onboarding tailored to the units where hell be picking up shifts, and matching him with shifts that work for his busy schedule. Instead of frustration and delays, the nurse begins with confidence and is able to start earning and covering his bills faster. Applying AI to the roles that need it today As the tech industry grapples with shifts toward white-collar jobs and AIs role, we have an opportunity. The same sophisticated AI systems that can automate corporate reporting can be adapted to optimize shift schedules. The same machine learning that powers chatbots can improve safety protocols. The same natural language processing that summarizes emails can help workers with limited English proficiency better understand their rights and benefits. The current moment of disruption in white-collar work is painful for millions of people, and that pain deserves recognition and resolution. At the same time, it also creates an opening to ask bigger questions about where AI should be applied and who it could serve. The AI revolution isnt going away anytime soon. This is our opportunity to choose how we use it, and who benefits.
Category:
E-Commerce
All news |
||||||||||||||||||
|
||||||||||||||||||