Informed i’s Weekly Business Insights

Governments around the world face mounting challenges that threaten their ability to effectively deliver essential services. Barriers ranging from fiscal constraints to skilled-workforce shortages and high consumer expectations are forcing agencies to do more with less, while declining public trust reduces their ability to act. The need for efficiency and innovation has arguably never been greater.

Technology has long been heralded as a potential solution, promising modernization, streamlined processes, and productivity gains across the private and public sectors. Recent breakthroughs in AI give reason for optimism—especially the application of generative and agentic AI—around how the technology can improve productivity and other measures. Yet its impact on services to residents remains elusive and many AI deployments seem stuck in “pilot purgatory,” not yet achieving results that reach residents or frontline workers as organizations struggle with data access, workflow integration, model risk, and high ongoing operating costs.

Can governments capture value from AI? It’s rarely, if ever, about the tools or technology alone. Unleashing the potential of AI in the public sector requires rewiring operations by redesigning workflows, adopting fundamentally new ways of working, and engaging the workforce to drive and scale adoption. Even the most advanced tools will underdeliver if they do not address structural issues, including outdated processes, fragmented decision-making, and misaligned workforce models.

That demands a proven, four-part approach undertaken simultaneously, not sequentially:

1. Crafting a strategy that leads with the mission outcome. The focus should be on meaningful outcomes for residents at an appropriate cost, not on tools or technologies.
2. Reimagining required workflows end to end. The time for piecemeal experiments or use cases is over.
3. Building the operating system around the technology. Careful change management will be needed to drive adoption that scales.
4. Keeping humans in the loop for any consequential action. There should be clear lines around what humans do and what AI does.

While AI is already part of people’s lives and has the potential to be a powerful force for good in the public sector, governments have very different regulatory regimes and degrees of organizational and public support. The road to rewiring can therefore be variable and lengthy. But the four-part approach when taken together, be a starting point to unlock AI’s potential to deliver better services to the public while building trust and resilience.

The startup landscape is undergoing its most profound transformation since the internet revolution. This change is powered by a convergence of forces. Large language models (LLMs) have reached a level of maturity that makes them capable of reasoning and planning tasks. When given access to relevant data and permission to act on the user’s behalf, LLMs can act as agents, taking on the role of digital assistant or even digital employee. New frameworks enable multiple agents to coordinate autonomously, creating agentic AI systems, or teams of digital colleagues. Low-cost application programming interfaces (APIs) have made it inexpensive to connect different IT systems so that they can share data and work together. And cloud costs continue to fall even as computing capabilities rise. These trends are producing a second great compression of entrepreneurship: The cost, time, and head count required to build a prototype of a product, test it, and then build an improved version have all collapsed. This compression shifts the competitive baseline. Small startups can enter, expand, or disrupt markets at a pace and cost structure that challenges incumbents to reimagine their operating models.

The Five Forces of Disruptive Change.  Five interconnected forces are creating a new entrepreneurial operating model. These forces strengthen one another, resulting in a compounding advantage over time. Together they are reshaping how a startup builds a prototype of a product or business model, systematically tests it and proves that there is real-world demand for it, and scales it up.

1. Zero-latency iteration.  Zero-latency iteration refers to the feedback loop in developing a digital product. It means that each version of a digital product can be changed instantly upon receiving customer feedback, which is dramatically shrinking the time required to design, launch, and scale up a new product.
2. Automated go-to-market capabilities.  AI-native companies are employing capabilities that allow them to execute go-to-market strategies at a scale and speed that were once available only to well-funded enterprises with large marketing departments. These capabilities are slashing the costs of acquiring and onboarding new customers.
3. Autonomous business functions.  Agentic AI systems are making it possible to plan, execute, and optimize business activities end to end with minimal human involvement. This is dramatically reducing the head count required to build and scale a new business and constitutes one of the most radical differences between traditional startup models and those employing agentic AI systems.
4. Radical capital efficiency.  Thanks to a combination of lean operations, rapid iterations during product development, and autonomous AI agents, AI-native startups need much less capital to launch a new business and become profitable and can raise it faster than conventional startups.
5. The AI-driven flywheel.  By solving the “unsolvable” manual tasks at the center of a company’s enterprise, agentic AI startups gain an early and deep understanding of their customers’ core priorities and continuously improve how they address them. That is one of their main competitive advantages. Indeed, customized AI interfaces are no longer scarce, and off-the-shelf agent-building platforms are increasingly available. These startups’ real strength lies in the deep operational and workflow expertise gained from companies deploying their customized AI agents to solve some of their customers’ most intractable problems in their most economically critical business units.

Incumbents Must Rethink Their Architecture.  The rise of capital-efficient, AI-native competitors exposes a structural mismatch between how incumbents were built and how agentic systems operate. Startups form around clean workflows, tight learning loops, and data models designed for systems that learn continuously. Most incumbents, by contrast, evolved for stability, specialization, and control, and they resist efforts at experimentation and innovation.

The obstacle for established companies isn’t a shortage of AI tools; it’s the organizational architecture those tools land on. Employees who have spent a decade mastering stable processes have little preparation for AI-augmented roles. Middle managers, whose authority historically comes from controlling information and resources, tend to balk at anything that threatens their span of control. None of this reflects a lack of capability; it reflects how the roles were designed.

Yet AI-native startups face vulnerabilities that incumbents can exploit. Autonomous systems introduce quality, reliability, and compliance risk: A misinterpreted signal, an incorrect approval, or a faulty configuration can trigger real-world consequences. Demonstrating control across multi-agent systems is nontrivial, especially when the systems rely on AI models that are inherently unpredictable. Unlike traditional software, AI agents don’t always produce the same result twice, making their behavior harder to standardize and verify. Startups must build testing, monitoring, and escalation mechanisms that incumbents have spent decades refining.

What Incumbents Should Do Now.  Skip the pilots that bolt agents onto broken workflows. Instead start where friction in workflows is highest and the stakes are greatest.  Following are the suggestions:  run a vulnerability audit against the five disruptive forces; start with a bounded, high-friction process; re-architect before you automate; partner with AI-native ventures; strengthen data quality and make it clear when tasks should go back to humans; and prepare people for shifting roles.

The structure and work processes of modern teams make it harder for teammates to get to know one another well and build rapport. Companies increasingly form ad hoc teams as new challenges and opportunities arise. Virtual meetings, now routine, tend to be task-focused and offer few opportunities for team members to connect on a personal level.  Further to this systematic attention to how different personality types affect how teams interact and perform is rare.

The author used the widely studied and accepted Big Five personality test to help reveal the psychological differences that can sap collaboration and team productivity. The assessment highlights individual differences in how people think, act, and feel across 30 personality traits in five broad dimensions: need for stability, extraversion, openness, agreeableness, and conscientiousness. This framework has dominated personality research because of its high validity and robustness over time.

The advantage of using personality assessments as diagnostic tools is that they open up rich conversations. Through facilitated discussions, participants can connect their collective personality traits with team tensions or difficulties — whether latent or overt — and defuse them in the process. And for newly formed teams, the same exercise can provide insights that accelerate cohesion. Whether team members collaborate effectively is often rooted in their ability to work through, accept, and capitalize on their differences.

Three Patterns to Look for.  Once the team understands its profile, members can reflect on the implications of the most extreme patterns — specifically, which perspectives they might be overlooking and where they can expect disagreements.  Three patterns warrant particular attention: a skewed cluster, where scores are tightly grouped at one extreme; a dispersed distribution, where the team diverges widely on one trait; and a grouping with a strong outlier. Left unmanaged, these configurations can create blind spots or tensions, leading to three traps that can derail teams:  the like-mindedness trap, the synergy trap, and the outlier trap.

Guidelines for a Productive Discussion.  Facilitators can use the following six steps to help team members reflect on and discuss their similarities, differences, and challenges. 

Address participants’ questions about the goals of the exercise and any concerns about the confidentiality of their profiles.  Ask: What are your main concerns about this session?

Set the ground rules for respectful communication: active listening, confidentiality, and no interruptions.  Ask: How can we encourage colleagues to express their emotions constructively?

Discuss the risks and benefits of traits where the team is tightly clustered on one side or widely dispersed across the scale.  Ask: How could the spread of scores on each trait help or hinder our interactions?

Relate observations about the assessment findings to previous team difficulties.  Ask: Does this resonate with past instances of tension, inertia, or misunderstanding?

Invite team members to suggest ways of addressing those challenges.  Ask: What are some steps team members can take to manage their personalities more effectively?

Recap the team-level lessons and proposals.  Ask: What was the biggest surprise, and which takeaways will you prioritize?

Employers continue to report talent shortages while millions of capable job hunters struggle to find work. At first glance, the contradiction makes little sense. Yet it reflects a growing disconnect between how organizations source talent and the realities of today’s workforce.

Yet early-career professionals are not the only people encountering hiring barriers. Experienced candidates also struggle to gain interviews, often because of assumptions about age, career aspirations or technology skills rather than objective capability. Equally qualified older applicants receive substantially fewer interview callbacks than younger applicants.

The organizations best positioned to thrive will be those that treat talent sustainability as a leadership strategy rather than simply a hiring strategy. That requires rethinking not only where talent comes from but also how potential, capability and contribution are recognized throughout a whole-life career.

Organizations today face two significant blind spots that hinder their ability to attract and retain talent. First, most talent strategies ignore the global demographic reality. Second, recruiting practices continue to focus disproportionately on a mythical hiring sweet spot, overlooking capable people whose age, experience or career stage falls outside traditional expectations.

7 Leadership Moves That Strengthen Talent Sustainability.  

1. Expand where you look for talent.  Talent sustainability begins with recognizing that capability can come from many places. Graduates, career changers, return-to-work professionals, internal candidates and experienced workers all represent valuable sources of capability. Organizations that repeatedly recruit from the same narrow hiring profiles reduce their capacity to innovate, solve problems and respond to change.
2. Create capability, don’t just acquire it.  Hiring experienced people is only one way to strengthen an organization. Apprenticeships, stretch assignments, mentoring, job rotations and meaningful early-career opportunities all help create the expertise organizations will need in the future. The strongest organizations develop talent rather than simply competing for it.
3. Make capability renewal continuous.  Skills, technologies and business needs evolve constantly. Organizations that invest in continuous learning, reskilling and upskilling are better equipped to respond to changing customer expectations, technological advances and economic disruption. Organizational strength is built through continual renewal, not one-time training.
4. Design for whole-life careers.  Career paths are becoming longer and less linear. Flexible work, internal mobility, career transitions and re-entry pathways help people continue contributing throughout changing life circumstances while enabling organizations to retain valuable knowledge and experience.
5. Remove hidden capability barriers.  Review hiring, promotion, development and performance practices for assumptions that unnecessarily narrow access to opportunity. Organizations make better decisions when they focus on skills, potential and contribution rather than rigid career expectations or outdated assumptions.
6. Build a culture where everyone learns.  Learning should not be limited by age, tenure or job title. Organizations that encourage continuous learning, mutual mentoring and knowledge sharing become more adaptable while helping employees grow throughout their careers.
7. Build a capability ecosystem.   Talent sustainability extends beyond recruitment. Organizations create long-term resilience by connecting hiring, learning, leadership development, internal mobility and knowledge sharing into a system that continuously creates, renews and retains organizational capability.

Jeff Bezos won’t pretend he didn’t see this day coming. Sure, when he was sitting in his Bellevue, Wash., garage in 1995, at a makeshift desk made from a wooden door slab, he probably wasn’t imagining that his nascent online bookshop would grow into the largest company in the world. Instead, he may have been puzzling over how to set up extension cords to keep his computers and servers running without tripping his home’s circuit breakers. Or he may have been mulling over changing the name of his tiny startup, then called Cadabra, to Amazon.

But three decades later, now that his startup has become an “everything company” that reportedly delivers as many packages as the U.S. Postal Service; now that the Amazon Web Services (AWS) cloud business powers a third of the internet; and now that Amazon has topped the Fortune 500 and ascended to the No. 1 spot in the Global 500—making it the biggest company in the world—Bezos admitted to Fortune: “It’s not like it’s a complete surprise.” 

And anyhow, Bezos added in April, sitting at that original door-desk in his study in Washington, D.C., bigness was never the point. “I don’t want us to take pride in being big,” Bezos said. “I want us to take pride in servicing customers. And it turns out, if you service customers really well, that will drive growth.” 

“Customer obsession” has long been a mantra for the 62-year-old Amazon founder. (In fact, he used the phrase 10 times during the conversation.) And with good reason: It’s what propelled the company to the top of Fortune’s annual list of the largest companies in the world by revenue, a spot held by Walmart for over a decade. Amazon will likely become the first company to reach $1 trillion in revenue in the next few years. 

Bezos, one of the richest people in the world, handed the CEO role to AWS chief Andy Jassy in 2021, but he still serves as Amazon’s executive chair, and he’s often asked about the company’s next big projects. Satellite internet? Advertising solutions? Health care? Yes, yes, and yes. 

But the company’s biggest venture by far is a bid to own the future of AI. Amazon devoted $131 billion to capital expenditures in 2025 and estimates that it will spend some $200 billion in 2026—largely on AWS and generative AI. 

“A few of our offerings have become durable pillars, things like Marketplace and Prime and AWS,” Bezos said. “What I see right now is that our chips business, our silicon business, is lining up to be our next pillar.”  If Bezos is right—and his track record suggests he often is—Amazon’s current enormous size could soon look modest.

Of course, for a company this dominant, there are always risks. In addition to battling Walmart in the retail sector, Amazon faces tough competition in the AI arms race—and there’s no denying that Amazon has some catching up to do. 

Bezos’s decision to step away from the CEO role and focus on his other tech companies created concerns that the founder was taking his foot off the gas at Amazon just as the AI race was starting, said futurist and NYU Stern School of Business professor Amy Webb: “I would be curious to know if the next 10 years of the company is going to be as dramatic and exciting as the past 10 years,” she said. 

With all the big plans ahead, Bezos said he’s also laser-focused on what’s not going to change: that foundational ethos of low prices, fast delivery, and a vast selection—an approach that’s just as powerful when applied to chips and servers as it is in the world of books and groceries. It’s what put Amazon on the map in the first place, and what keeps customers coming back. 

It may look now like the success of Bezos’s model was a foregone conclusion, but getting investors to take a chance on an online-only store in the early days of the internet was not easy. Before Amazon’s launch in 1995, Bezos recalls having to explain to people what the internet was. He spoke with some 60 investors to raise a million dollars. Twenty-two said yes, with checks of roughly $50,000 each. 

Amazon went public two years later, on May 15, 1997, at a price of $18 and a valuation of nearly $440 million. (It’s now valued at $2.6 trillion.) Initially, investors were cautious. A 1999 Barron’s cover story titled “Amazon.bomb” questioned whether the startup would ever turn a profit. 

Bezos concedes that Amazon’s trajectory since its launch was improbable, to say the least. “You could not at that time have predicted the magnitude of change that would occur—and anybody who did predict that magnitude of change would probably have been quickly institutionalized and sent to the mental hospital,” he said. “It wouldn’t have been credible or believable.” 

Amazon didn’t become profitable until the early 2000s, years after expanding its product offerings beyond books. It was the launch of cloudcomputing division AWS in 2006 that changed everything. Today, millions of businesses and governments use AWS cloud infrastructure and services to power everything from streaming videos on Netflix to ordering food on DoorDash. Operating income for AWS reached $45.6 billion in 2025, on revenue that rose 20% to $128.7 billion from the prior year. 

Even Jeff Bezos couldn’t have predicted that an online bookstore would reshape the entire retail sector.

Beyond AWS, a string of consumer-facing initiatives extended Amazon’s reach over the next two decades. 

Bezos points to the company’s famed leadership principles, with their emphasis on constant innovation, as the key to its success. “A lot of companies will tell you they’re customer-obsessed, but they’re really competitor-obsessed,” Bezos said. “You can’t be customer-obsessed unless you love inventing…You have to do new things. And Amazon is culturally very good at both of those things.” 

No company can keep soaring forever. “Jeff has even said this: There will be a day when Amazon goes the way of the dodo,” he said. “Amazon will cease to exist at some point. It’s inevitable.” Perhaps to stave off that extinction, Amazon’s corporate culture is notoriously hard-driving and competitive.

Meta is building an AI service that it says could eventually help entrepreneurs start and run entire businesses—and charge them only when it produces results.

“Over time, we’d like to build this into a business-in-a-box service that can help you start and run a whole business using Meta’s platforms,” Mark Zuckerberg told investors during Meta’s second-quarter earnings call Wednesday.

More than one million businesses now use Meta’s agents on WhatsApp and Messenger each week, Zuckerberg said. Meta is also rolling them out on Instagram.

The agents can already talk to customers and complete sales. Meta is developing tools that summarize those conversations, identify what customers want, provide competitive intelligence, and suggest ways to grow. Zuckerberg said future versions could “digest what happened overnight” and surface the most important information to business owners.

Meta expects to charge through subscriptions and volume-based fees. Zuckerberg said some products could eventually resemble its advertising system, where “businesses only pay us when we achieve results for them.”  The plan offers a possible payoff from an AI-heavy infrastructure buildout that consumed nearly all of Meta’s quarterly operating cash.

Zuckerberg also floated selling APIs, productivity tools, and computing capacity to large customers. But Business Agents represent Meta’s most concrete attempt to turn that infrastructure into a new product: Businesses are already using them, and Meta now has a plan to make them pay.

The author has spent years on both ends of the cold pitch. HeI sent them to land coverage and clients, and because of the work he does, he receives a steady stream of them, too. That second seat taught him more than the first. Watching hundreds of pitches land in his own inbox showed him, in seconds, why almost all of them get deleted and why a rare few earn a reply.

The deletions are not close calls. Most cold outreach fails the moment it arrives, because it was clearly blasted to a list and could have been addressed to anyone. The reader feels that instantly. Gartner found that most B2B buyers now actively avoid suppliers who send irrelevant outreach, and that bad prospecting damages a relationship rather than starting one. The good news is that the pitches that work are not cleverer or better written. They just do three simple things that the rest skip.

1. They prove you did your homework.  The single biggest predictor of a reply is evidence that the sender actually knows who they are writing to. Not a merge field with the first name. Real, specific proof — a reference to something the person published, a detail about his/her company, a reason this message is landing in his/her inbox and not someone else’s.  According to the author when he pitches, he earns the first sentence before he writes anything else. He finds one true, specific thing about the person — a recent article, a product launch, a shift at their company — and he leads with it. It tells the reader, in one line, that the message was meant for them. That tiny act of research is what separates a note that feels like a conversation from one that feels like spam.  This is also the cheapest edge available, because so few people bother.
2. They make one clear, small ask.  The second habit is restraint. The pitches that die try to close the whole deal in the first email. They ask for a 30-minute call, attach a deck, list every feature and end with three different links. It is exhausting to read, and exhausting gets deleted.  The ones that work ask for one small thing.  Lower the cost of saying yes, and more people say it. Think of the first message as earning the right to a second one, not winning the entire relationship in a paragraph. Once someone replies, even briefly, you are no longer cold. You are in a conversation, and conversations are where deals and coverage actually get made.
3. They follow up without being annoying.  The third habit is the one most people quit before reaching. They send a single email, hear nothing and assume the answer is no. Far more often, the answer is “I was busy and your email got buried.” Silence is rarely rejection. It is usually just an inbox doing what inboxes do.  A short, polite follow-up recovers a surprising share of those lost conversations. The discipline is in the tone. A good follow-up adds something — a new angle, a fresh piece of context, a quick reason the timing might now make sense. A bad one just whines, “Did you see my last email?” One adds value, and the other adds pressure and only one gets a reply.  There is a line, of course. One or two thoughtful follow-ups spaced out over a couple of weeks is persistence. Five in five days is harassment, and it burns the relationship for good. Aim to be the kind of sender you would actually want to hear from: present and useful, not desperate.