 |
|
|
|||||
I spent nearly a decade as an intrapreneur inside the worlds largest global holding companies. On paper, it looked a lot like entrepreneurship: validate an idea, conduct research, raise or allocate funds, build capabilities, codify processes, launch SaaS platforms, measure value creation, and implement a communication plan. In practice, it was very different. Big organizations are optimized for productivity and predictability, not the full lifecycle of experimentation that product building requires. That law of nature creates a constant source of friction between innovation and day-to-day business. A new MIT study puts numbers to what many of us have experienced: 95% of enterprise GenAI pilots fail to deliver measurable business impact, despite billions invested. The problem is less about model quality and more about the learning gap: Tools and organizations do not naturally adapt to one another, so in-house pilots never become production systems. MIT and other researchers highlight consistent fault lines: Flawed integration: Pilots sit on the side and never embed into real workflows. The companies that do see impact redesign processes and roles around AI rather than sprinkling models on top. Learning gaps and culture: Organizations treat AI like a oneoff project, not an evolving capability, so teams do not learn with the tools. Misallocated budgets: Spending skews to sales and marketing experiments while the highest ROI is often in backoffice automation that reduces outsourced processes and eliminates manual work. Build versus buy: Buying from specialized vendors and partnering works about 67% of the time, compared to internal builds succeeding roughly onethird as often. Shadow AI risk: employees use personal chatbots at most companies, which muddies impact measurement and raises compliance risk. Reports find widespread unsanctioned use. These patterns are not unique to AI. I saw the same dynamics at play when launching products within corporations long before the AI wave became the center point of the software conversation. The code is never the blocker to success. Its all about incentives. Billable hours and shortterm deliverables are naturally at odds with the patience, rework, and staged learning a product needs. Without a protected path from pilot to scale, even strong concepts suffocate in a productivityfirst culture. Context from prior waves reinforces this current moment in time: an MIT SloanBCG study found only about 10% of organizations realized significant financial benefits from AI, with success tied to how well humans and AI learn together. A year later the research emphasized that organizations capture value when individual workers also feel empowered and gain competence and autonomy from the tools. Even now, adoption at scale remains limited: One recent, large CIO survey reported only 11% had fully implemented AI due to security and data readiness constraints. What successful programs do differently The efforts that work do not live as science projects. They integrate early, align incentives with outcomes, and earn trust on the front line. They move quickly from test to tool. The playbook looks like this: Start with a workflow, not a model. Redesign the process where the decision happens, then fit AI to it. Treat AI as infrastructure that changes who does what and when. Pick one painful, measurable problem. Scope narrowly, ship a useful tool, and iterate in place. Tie success to a business owners KPI. The MIT study notes that the winners execute against specific pain points rather than broad ambitions. Choose to build, buy, or partner with discipline. If timetovalue matters, lean into vendors with proven outcomes, then extend. The success gap between vendor solutions and internal builds is material. Shift investment to the quiet ROI. Target backoffice and operational automation where savings are concrete and compounding. Use those gains to fund the next wave. Make learning a firstclass objective. Pair tool learning with organizational learning: training, job design, accountability, and feedback loops. Bring shadow AI into the light. Set clear guardrails, offer approved tools, and measure use so value shows up in the P&L instead of slipping through side channels. The takeaway here is not that AI is overhyped; it is that experimentation without integration rarely creates transformation. Leaders who treat AI like infrastructure, align incentives to outcomes, and build learning into the operating model will escape the pilot trap. The rest will keep adding to the graveyard. James Chester is cofounder and CEO of WVN.
Category:
E-Commerce
As the founder, chair, and CEO of the Exceptional Women Alliance, I am fortunate to be surrounded by extraordinary female business leaders. Our purpose is to empower each other through peer mentorship that provides personal and professional fulfillment within this unique sisterhood. Joanna Massey, PhD, is one of those business leaders, and she is not afraid to challenge the status quo. She is a corporate board director, Fortune 500 executive, and expert in corporate governance and crisis communications. With advanced degrees in business, law, and psychology, she brings a unique, interdisciplinary perspective to one of the most pressing issues of our time: how to protect free speech in the digital age without sacrificing public safety and democracy. Q: You wrote a policy paper for Cornell Law School on regulating free speech. What do people get wrong about the First Amendment? Massey: In the United States, you can say what you want, but you are still responsible for the damage your words do. Thats the part people forget. The First Amendment protects your right to speak freely without the government punishing you. It doesnt protect you from the consequences of what you sayor from being banned by private-sector businesses, like Facebook, Twitter/X, and TikTok. They set their own rules, and if you break them, you deal with the penalties. Q: You say in your work that Americans misunderstand what liberty means. Can you explain? Massey: Liberty was never meant to be limitless. Our founding fathersThomas Jefferson and James Madison, among othersbelieved that freedom had to be balanced with responsibility. The Constitution wasnt written to give one person the right to dominate another. It protects us from the government, but it also protects us from each other. So, when you spew hate online because you dont like how I look, who I love, or what I believe, that isnt exercising your rights. Its infringing on mine. Q: So how do we define the line between free expression and harmful speech today? Massey: Right now, our speech laws focus on intent. The courts want to know, Did you mean to incite violence? Who is going to say yes to that? Its also an outdated standard because the issue today is not the intent behind attacking an individual or group of peopleit is the cumulative impact of the speech. One cigarette doesnt cause cancer, but cumulatively, secondhand smoke doeswhich is why we regulate it. Your freedom to smoke stops when it endangers me. Now, apply that to hatred. One racial slur doesnt cause a riot, but unchecked and repeated hate does. Based on our Constitutional rights, your freedom to spew hate stops when it takes away my ability to live safely and freely. A good example is the false rumors that spread in 2024 about Haitian immigrants in a small Ohio town. Even after officials and business leaders debunked the lies, threats escalated until schools closed, offices shut down, and the entire community was destabilized. Speech today doesnt live in isolationextremism unfolds through a steady stream of posts, shares, and content that doesnt break current laws but collectively causes harm. Q: Why is social media dividing people? Massey: Human beings are biologically hardwired for survival, and our brains dont know the difference between a tiger and a tweet. When someone criticizes our beliefs or lifestyle, our brain reacts as if we are under physical attackby banding together, retreating into tribes, and protecting our side as if our lives depend on it. Platforms give us endless ways to find our people and feel safe inside bubbles that affirm our beliefs. Those algorithms are also programmed to shut out dissenting views and lifestyles, so we dont experience other perspectives in a neutral way. Q: Youve coined the term mass incitement. What does that mean? Massey: Mass incitement happens when platforms or public figures repeatedly amplify false or inflammatory content until millions are echoing it, creating a collective force that makes violence or discrimination more likely. Q: Some say users just need to be more skeptical about the media they consume. But is the fix that simple? Massey: That is a convenient argument, but it misses the point. The real problem is impact. You cannot exercise your rights by infringing on minethat runs counter to the promises of the Constitution. Up until now, we have been blaming our division on politics, but the problem isnt red (Republican) or blue (Democrat). Its green (money). Social media companies make money every time we click, and people stay engaged longer when theyre upset. Thats why the algorithms promote outrage, not accuracy. These platforms arent neutral. Theyre profiting from our disagreements. We regulate television, radio, and phone lines to protect the public interestbut somehow, weve left algorithms completely unchecked. That legal void is fueling chaos. Q: What reforms would actually make a difference? Massey: The answer is modernizing our laws to reflect the reality of mass incitement. That means updating FCC authority, reforming Section 230 of the Communications Decency Act, and holding social platforms accountable the same way we do other producers of products that cause cumulative, foreseeable harm. Q: Free speech absolutists say any regulation is a threat to democracy. How do you respond to that? Massey The real threat to democracy is weaponized speech. Misinformation fuels division and violence, hate speech becomes normalized, and society starts to break down. So, calling hate speech free speech is like calling an assault self-expression. The Constitution protects us from harm, including the harms suffered by victims of hate speech. We have to reconcile that with how much protection hate speech is given today. The answer is to create guardrails that keep speech free and fair. We banned cigarette ads on TV. We rated movies. We censored shock jocks. And the First Amendment survived all of it. It will survive hate speech regulation, as well. Larraine Segil is founder, chair, and CEO of The Exceptional Women Alliance.
Category:
E-Commerce
Rivian is recalling 24,214 R1S and R1T electric vehicles due to a software defect that may cause its hands-free Highway Assist system to misidentify lead vehicles, the U.S. National Highway Traffic Safety Administration said on Friday. The issue affects certain 2025 Rivian vehicles running an older software version in the United States, the NHTSA said. Rivian has issued an over-the-air software update to fix the problem, the NHTSA added. The defect was identified after an incident involving a 2025 R1S model vehicle, where the system misclassified a low-speed vehicle and the driver failed to maintain control. Automakers have increasingly competed to roll out advanced driver-assistance features like lane-keep assist and adaptive cruise control. Rivian has also been working on hands-free and “eyes-off” systems as part of its push into autonomous driving technology. Preetika Parashuraman and Rajveer Singh Pardesi, Reuters
Category:
E-Commerce
All news |
||||||||||||||||||
|
||||||||||||||||||