Snowflake is bolstering its data stack to make the platform better suited to the sheer volume of data produced by AI agents.

The exact impact AI will have on the enterprise labor market is unclear but investors predict trends will start to emerge in 2026.

More than 20 venture capitalists share their thoughts on AI agents, enterprise AI budgets, and more for 2026.

Presented by Capital One Software

---

Tokenization is emerging as a cornerstone of modern data security, helping businesses separate the value of their data from its risk. During this VB in Conversation, Ravi Raghu, president, Capital One Software, talks about the ways tokenization can help reduce the value of breached data and preserve underlying data format and usability, including Capital One’s own experience leveraging tokenization at scale.

Tokenization, Raghu asserts, is a far superior technology. It converts sensitive data into a nonsensitive digital replacement, called a token, that maps back to the original, which is secured in a digital vault. The token placeholder preserves both the format and the utility of the sensitive data, and can be used across applications — including AI models. Because tokenization removes the need to manage encryption keys or dedicate compute to constant encrypting and decrypting, it offers one of the most scalable ways for companies to protect their most sensitive data, he added.

"The killer part, from a security standpoint, when you think about it relative to other methods, if a bad actor gets hold of the data, they get hold of tokens," he explained. "The actual data is not sitting with the token, unlike other methods like encryption, where the actual data sits there, just waiting for someone to get hold of a key or use brute force to get to the real data. From every angle this is the ideal way one ought to go about protecting sensitive data."

The tokenization differentiator

Most organizations are just scratching the surface of data security, adding security at the very end, when data is read, to prevent an end user from accessing it. At minimum, organizations should focus on securing data on write, as it’s being stored. But best-in-class organizations go even further, protecting data at birth, the moment it’s created.

At one end of the safety spectrum is a simple lock-and-key approach that restricts access but leaves the underlying data intact. More advanced methods, like masking or modifying data, permanently alter its meaning — which can compromise its usefulness. File-level encryption provides broader protection for large volumes of stored data, but when you get down to field-level encryption (for example, a Social Security number), it becomes a bigger challenge. It takes a great deal of compute to encrypt a single field, and then to decrypt it at the point of usage. And still it has a fatal flaw: the original data is still right there, only needing the key to get access.

Tokenization avoids these pitfalls by replacing the original data with a surrogate that has no intrinsic value. If the token is intercepted — whether by the wrong person or the wrong machine — the data itself remains secure.

The business value of tokenization

"Fundamentally you’re protecting data, and that’s priceless," Raghu said. "Another thing that’s priceless – can you use that for modeling purposes subsequently? On the one hand, it’s a protection thing, and on the other hand it’s a business enabling thing."

Because tokenization preserves the structure and ordinality of the original data, it can still be used for modeling and analytics, turning protection into a business enabler. Take private health data governed by HIPAA for example: tokenization means that data canbeused to build pricing models or for gene therapy research, while remaining compliant.

"If your data is already protected, you can then proliferate the usage of data across the entire enterprise and have everybody creating more and more value out of the data," Raghu said. "Conversely, if you don’t have that, there’s a lot of reticence for enterprises today to have more people access it, or have more and more AI agents access their data. Ironically, they’re limiting the blast radius of innovation. The tokenization impact is massive, and there are many metrics you could use to measure that – operational impact, revenue impact, and obviously the peace of mind from a security standpoint."

Breaking down adoption barriers

Until now, the fundamental challenge with traditional tokenization has been performance. AI requires a scale and speed that is unprecedented. That's one of the major challenges Capital One addresses with Databolt, its vaultless tokenization solution, which can produce up to 4 million tokens per second.

"Capital One has gone through tokenization for more than a decade. We started doing it because we’re serving our 100 million banking customers. We want to protect that sensitive data," Raghu said. "We’ve eaten our own dog food with our internal tokenization capability, over 100 billion times a month. We’ve taken that know-how and that capability, scale, and speed, and innovated so that the world can leverage it, so that it’s a commercial offering."

Vaultless tokenization is an advanced form of tokenization that does not require a central database (vault) to store token mappings. Instead, it uses mathematical algorithms, cryptographic techniques, and deterministic mapping to generate tokens dynamically.This approach is faster, more scalable, and eliminates the security risk associated with managing a vault.

"We realized that for the scale and speed demands that we had, we needed to build out that capability ourselves," Raghu said. "We’ve been iterating continuously on making sure that it can scale up to hundreds of billions of operations a month. All of our innovation has been around building IP and capability to do that thing at a battle-tested scale within our enterprise, for the purpose of serving our customers."

While conventional tokenization methods can involve some complexity and slow down operations, Databolt seamlessly integrates with encrypted data warehouses, allowing businesses to maintain robust security without slowing performance or operations. Tokenization occurs in the customer’s environment, removing the need to communicate with an external network to perform tokenization operations, which can also slow performance.

"We believe that fundamentally, tokenization should be easy to adopt," Raghu said. "You should be able to secure your data very quickly and operate at the speed and scale and cost needs that organizations have. I think that’s been a critical barrier so far for the mass scale adoption of tokenization. In an AI world, that’s going to become a huge enabler."

Don't miss the whole conversation with Ravi Raghu, president, Capital One Software, here.

---

Sponsored articles are content produced by a company that is either paying for the post or has a business relationship with VentureBeat, and they’re always clearly marked. For more information, contact sales@venturebeat.com.

The two companies are launching the Accenture Anthropic Business Group to bring Anthropic's AI to Accenture's employees.

Empromptu claims all a user has to do is tell the platform's AI chatbot what they want — like a new HTML or JavaScript app — and the AI will go ahead and build it.

AWS has been working with the U.S. government since 2011 and is now building AI infrastructure specifically for the entity.

LinkedIn is launching its new AI-powered people search this week, after what seems like a very long wait for what should have been a natural offering for generative AI.

It comes a full three years after the launch of ChatGPT and six months after LinkedIn launched its AI job search offering. For technical leaders, this timeline illustrates a key enterprise lesson: Deploying generative AI in real enterprise settings is challenging, especially at a scale of 1.3 billion users. It’s a slow, brutal process of pragmatic optimization.

The following account is based on several exclusive interviews with the LinkedIn product and engineering team behind the launch.

First, here’s how the product works: A user can now type a natural language query like, "Who is knowledgeable about curing cancer?" into LinkedIn’s search bar.

LinkedIn's old search, based on keywords, would have been stumped. It would have looked only for references to "cancer". If a user wanted to get sophisticated, they would have had to run separate, rigid keyword searches for "cancer" and then "oncology" and manually try to piece the results together.

The new AI-powered system, however, understands the intent of the search because the LLM under the hood grasps semantic meaning. It recognizes, for example, that "cancer" is conceptually related to "oncology" and even less directly, to "genomics research." As a result, it surfaces a far more relevant list of people, including oncology leaders and researchers, even if their profiles don't use the exact word "cancer."

The system also balances this relevance with usefulness. Instead of just showing the world's top oncologist (who might be an unreachable third-degree connection), it will also weigh who in your immediate network — like a first-degree connection — is "pretty relevant" and can serve as a crucial bridge to that expert.

See the video below for an example.

Arguably, though, the more important lesson for enterprise practitioners is the "cookbook" LinkedIn has developed: a replicable, multi-stage pipeline of distillation, co-design, and relentless optimization. LinkedIn had to perfect this on one product before attempting it on another.

"Don't try to do too much all at once," writes Wenjing Zhang, LinkedIn's VP of Engineering, in a  post about the product launch, and who also spoke with VentureBeat last week in an interview. She notes that an earlier "sprawling ambition" to build a unified system for all of LinkedIn's products "stalled progress."

Instead, LinkedIn focused on winning one vertical first. The success of its previously launched AI Job Search — which led to job seekers without a four-year degree being 10% more likely to get hired, according to VP of Product Engineering Erran Berger — provided the blueprint.

Now, the company is applying that blueprint to a far larger challenge. "It's one thing to be able to do this across tens of millions of jobs," Berger told VentureBeat. "It's another thing to do this across north of a billion members."

For enterprise AI builders, LinkedIn's journey provides a technical playbook for what it actually takes to move from a successful pilot to a billion-user-scale product.

The new challenge: a 1.3 billion-member graph

The job search product created a robust recipe that the new people search product could build upon, Berger explained. 

The recipe started with with a "golden data set" of just a few hundred to a thousand real query-profile pairs, meticulously scored against a detailed 20- to 30-page "product policy" document. To scale this for training, LinkedIn used this small golden set to prompt a large foundation model to generate a massive volume of synthetic training data. This synthetic data was used to train a 7-billion-parameter "Product Policy" model — a high-fidelity judge of relevance that was too slow for live production but perfect for teaching smaller models.

However, the team hit a wall early on. For six to nine months, they struggled to train a single model that could balance strict policy adherence (relevance) against user engagement signals. The "aha moment" came when they realized they needed to break the problem down. They distilled the 7B policy model into a 1.7B teacher model focused solely on relevance. They then paired it with separate teacher models trained to predict specific member actions, such as job applications for the jobs product, or connecting and following for people search. This "multi-teacher" ensemble produced soft probability scores that the final student model learned to mimic via KL divergence loss.

The resulting architecture operates as a two-stage pipeline. First, a larger 8B parameter model handles broad retrieval, casting a wide net to pull candidates from the graph. Then, the highly distilled student model takes over for fine-grained ranking. While the job search product successfully deployed a 0.6B (600-million) parameter student, the new people search product required even more aggressive compression. As Zhang notes, the team pruned their new student model from 440M down to just 220M parameters, achieving the necessary speed for 1.3 billion users with less than 1% relevance loss.

But applying this to people search broke the old architecture. The new problem included not just ranking but also retrieval.

“A billion records," Berger said, is a "different beast."

The team’s prior retrieval stack was built on CPUs. To handle the new scale and the latency demands of a "snappy" search experience, the team had to move its indexing to GPU-based infrastructure. This was a foundational architectural shift that the job search product did not require.

Organizationally, LinkedIn benefited from multiple approaches. For a time, LinkedIn had two separate teams — job search and people search — attempting to solve the problem in parallel. But once the job search team achieved its breakthrough using the policy-driven distillation method, Berger and his leadership team intervened. They brought over the architects of the job search win — product lead Rohan Rajiv and engineering lead Wenjing Zhang — to transplant their 'cookbook' directly to the new domain.

Distilling for a 10x throughput gain

With the retrieval problem solved, the team faced the ranking and efficiency challenge. This is where the cookbook was adapted with new, aggressive optimization techniques.

Zhang’s technical post (I’ll insert the link once it goes live) provides the specific details our audience of AI engineers will appreciate. One of the more significant optimizations was input size.

To feed the model, the team trained another LLM with reinforcement learning (RL) for a single purpose: to summarize the input context. This "summarizer" model was able to reduce the model's input size by 20-fold with minimal information loss.

The combined result of the 220M-parameter model and the 20x input reduction? A 10x increase in ranking throughput, allowing the team to serve the model efficiently to its massive user base.

Pragmatism over hype: building tools, not agents

Throughout our discussions, Berger was adamant about something else that might catch peoples’ attention: The real value for enterprises today lies in perfecting recommender systems, not in chasing "agentic hype." He also refused to talk about the specific models that the company used for the searches, suggesting it almost doesn't matter. The company selects models based on which one it finds the most efficient for the task.

The new AI-powered people search is a manifestation of Berger’s philosophy that it’s best to optimize the recommender system first. The architecture includes a new "intelligent query routing layer," as Berger explained, that itself is LLM-powered. This router pragmatically decides if a user's query — like "trust expert" — should go to the new semantic, natural-language stack or to the old, reliable lexical search.

This entire, complex system is designed to be a "tool" that a future agent will use, not the agent itself.

"Agentic products are only as good as the tools that they use to accomplish tasks for people," Berger said. "You can have the world's best reasoning model, and if you're trying to use an agent to do people search but the people search engine is not very good, you're not going to be able to deliver." 

Now that the people search is available, Berger suggested that one day the company will be offering agents to use it. But he didn’t provide details on timing. He also said the recipe used for job and people search will be spread across the company’s other products.

For enterprises building their own AI roadmaps, LinkedIn's playbook is clear:

1. Be pragmatic: Don't try to boil the ocean. Win one vertical, even if it takes 18 months.

2. Codify the "cookbook": Turn that win into a repeatable process (policy docs, distillation pipelines, co-design).

3. Optimize relentlessly: The real 10x gains come after the initial model, in pruning, distillation, and creative optimizations like an RL-trained summarizer.

LinkedIn's journey shows that for real-world enterprise AI, emphasis on specific models or cool agentic systems should take a back seat. The durable, strategic advantage comes from mastering the pipeline — the 'AI-native' cookbook of co-design, distillation, and ruthless optimization.

(Editor's note: We will be publishing a full-length podcast with LinkedIn's Erran Berger, which will dive deeper into these technical details, on the VentureBeat podcast feed soon.)

Microsoft is launching a significant expansion of its Copilot AI assistant on Tuesday, introducing tools that let employees build applications, automate workflows, and create specialized AI agents using only conversational prompts — no coding required.

The new capabilities, called App Builder and Workflows, mark Microsoft's most aggressive attempt yet to merge artificial intelligence with software development, enabling the estimated 100 million Microsoft 365 users to create business tools as easily as they currently draft emails or build spreadsheets.

"We really believe that a main part of an AI-forward employee, not just developers, will be to create agents, workflows and apps," Charles Lamanna, Microsoft's president of business and industry Copilot, said in an interview with VentureBeat. "Part of the job will be to build and create these things."

The announcement comes as Microsoft deepens its commitment to AI-powered productivity tools while navigating a complex partnership with OpenAI, the creator of the underlying technology that powers Copilot. On the same day, OpenAI completed its restructuring into a for-profit entity, with Microsoft receiving a 27% ownership stake valued at approximately $135 billion.

How natural language prompts now create fully functional business applications

The new features transform Copilot from a conversational assistant into what Microsoft envisions as a comprehensive development environment accessible to non-technical workers. Users can now describe an application they need — such as a project tracker with dashboards and task assignments — and Copilot will generate a working app complete with a database backend, user interface, and security controls.

"If you're right inside of Copilot, you can now have a conversation to build an application complete with a backing database and a security model," Lamanna explained. "You can make edit requests and update requests and change requests so you can tune the app to get exactly the experience you want before you share it with other users."

The App Builder stores data in Microsoft Lists, the company's lightweight database system, and allows users to share finished applications via a simple link—similar to sharing a document. The Workflows agent, meanwhile, automates routine tasks across Microsoft's ecosystem of products, including Outlook, Teams, SharePoint, and Planner, by converting natural language descriptions into automated processes.

A third component, a simplified version of Microsoft's Copilot Studio agent-building platform, lets users create specialized AI assistants tailored to specific tasks or knowledge domains, drawing from SharePoint documents, meeting transcripts, emails, and external systems.

All three capabilities are included in the existing $30-per-month Microsoft 365 Copilot subscription at no additional cost — a pricing decision Lamanna characterized as consistent with Microsoft's historical approach of bundling significant value into its productivity suite.

"That's what Microsoft always does. We try to do a huge amount of value at a low price," he said. "If you go look at Office, you think about Excel, Word, PowerPoint, Exchange, all that for like eight bucks a month. That's a pretty good deal."

Why Microsoft's nine-year bet on low-code development is finally paying off

The new tools represent the culmination of a nine-year effort by Microsoft to democratize software development through its Power Platform — a collection of low-code and no-code development tools that has grown to 56 million monthly active users, according to figures the company disclosed in recent earnings reports.

Lamanna, who has led the Power Platform initiative since its inception, said the integration into Copilot marks a fundamental shift in how these capabilities reach users. Rather than requiring workers to visit a separate website or learn a specialized interface, the development tools now exist within the same conversational window they already use for AI-assisted tasks.

"One of the big things that we're excited about is Copilot — that's a tool for literally every office worker," Lamanna said. "Every office worker, just like they research data, they analyze data, they reason over topics, they also will be creating apps, agents and workflows."

The integration offers significant technical advantages, he argued. Because Copilot already indexes a user's Microsoft 365 content — emails, documents, meetings, and organizational data — it can incorporate that context into the applications and workflows it builds. If a user asks for "an app for Project Spartan," Copilot can draw from existing communications to understand what that project entails and suggest relevant features.

"If you go to those other tools, they have no idea what the heck Project Spartan is," Lamanna said, referencing competing low-code platforms from companies like Google, Salesforce, and ServiceNow. "But if you do it inside of Copilot and inside of the App Builder, it's able to draw from all that information and context."

Microsoft claims the apps created through these tools are "full-stack applications" with proper databases secured through the same identity systems used across its enterprise products — distinguishing them from simpler front-end tools offered by competitors. The company also emphasized that its existing governance, security, and data loss prevention policies automatically apply to apps and workflows created through Copilot.

Where professional developers still matter in an AI-powered workplace

While Microsoft positions the new capabilities as accessible to all office workers, Lamanna was careful to delineate where professional developers remain essential. His dividing line centers on whether a system interacts with parties outside the organization.

"Anything that leaves the boundaries of your company warrants developer involvement," he said. "If you want to build an agent and put it on your website, you should have developers involved. Or if you want to build an automation which interfaces directly with your customers, or an app or a website which interfaces directly with your customers, you want professionals involved."

The reasoning is risk-based: external-facing systems carry greater potential for data breaches, security vulnerabilities, or business errors. "You don't want people getting refunds they shouldn't," Lamanna noted.

For internal use cases — approval workflows, project tracking, team dashboards — Microsoft believes the new tools can handle the majority of needs without IT department involvement. But the company has built "no cliffs," in Lamanna's terminology, allowing users to migrate simple apps to more sophisticated platforms as needs grow.

Apps created in the conversational App Builder can be opened in Power Apps, Microsoft's full development environment, where they can be connected to Dataverse, the company's enterprise database, or extended with custom code. Similarly, simple workflows can graduate to the full Power Automate platform, and basic agents can be enhanced in the complete Copilot Studio.

"We have this mantra called no cliffs," Lamanna said. "If your app gets too complicated for the App Builder, you can always edit and open it in Power Apps. You can jump over to the richer experience, and if you're really sophisticated, you can even go from those experiences into Azure."

This architecture addresses a problem that has plagued previous generations of easy-to-use development tools: users who outgrow the simplified environment often must rebuild from scratch on professional platforms. "People really do not like easy-to-use development tools if I have to throw everything away and start over," Lamanna said.

What happens when every employee can build apps without IT approval

The democratization of software development raises questions about governance, maintenance, and organizational complexity — issues Microsoft has worked to address through administrative controls.

IT administrators can view all applications, workflows, and agents created within their organization through a centralized inventory in the Microsoft 365 admin center. They can reassign ownership, disable access at the group level, or "promote" particularly useful employee-created apps to officially supported status.

"We have a bunch of customers who have this approach where it's like, let 1,000 apps bloom, and then the best ones, I go upgrade and make them IT-governed or central," Lamanna said.

The system also includes provisions for when employees leave. Apps and workflows remain accessible for 60 days, during which managers can claim ownership — similar to how OneDrive files are handled when someone departs.

Lamanna argued that most employee-created apps don't warrant significant IT oversight. "It's just not worth inspecting an app that John, Susie, and Bob use to do their job," he said. "It should concern itself with the app that ends up being used by 2,000 people, and that will pop up in that dashboard."

Still, the proliferation of employee-created applications could create challenges. Users have expressed frustration with Microsoft's increasing emphasis on AI features across its products, with some giving the Microsoft 365 mobile app one-star ratings after a recent update prioritized Copilot over traditional file access.

The tools also arrive as enterprises grapple with "shadow IT" — unsanctioned software and systems that employees adopt without official approval. While Microsoft's governance controls aim to provide visibility, the ease of creating new applications could accelerate the pace at which these systems multiply.

The ambitious plan to turn 500 million workers into software builders

Microsoft's ambitions for the technology extend far beyond incremental productivity gains. Lamanna envisions a fundamental transformation of what it means to be an office worker — one where building software becomes as routine as creating spreadsheets.

"Just like how 20 years ago you put on your resume that you could use pivot tables in Excel, people are going to start saying that they can use App Builder and workflow agents, even if they're just in the finance department or the sales department," he said.

The numbers he's targeting are staggering. With 56 million people already using Power Platform, Lamanna believes the integration into Copilot could eventually reach 500 million builders. "Early days still, but I think it's certainly encouraging," he said.

The features are currently available only to customers in Microsoft's Frontier Program — an early access initiative for Microsoft 365 Copilot subscribers. The company has not disclosed how many organizations participate in the program or when the tools will reach general availability.

The announcement fits within Microsoft's larger strategy of embedding AI capabilities throughout its product portfolio, driven by its partnership with OpenAI. Under the restructured agreement announced Tuesday, Microsoft will have access to OpenAI's technology through 2032, including models that achieve artificial general intelligence (AGI) — though such systems do not yet exist. Microsoft has also begun integrating Copilot into its new companion apps for Windows 11, which provide quick access to contacts, files, and calendar information.

The aggressive integration of AI features across Microsoft's ecosystem has drawn mixed reactions. While enterprise customers have shown interest in productivity gains, the rapid pace of change and ubiquity of AI prompts have frustrated some users who prefer traditional workflows.

For Microsoft, however, the calculation is clear: if even a fraction of its user base begins creating applications and automations, it would represent a massive expansion of the effective software development workforce — and further entrench customers in Microsoft's ecosystem. The company is betting that the same natural language interface that made ChatGPT accessible to millions can finally unlock the decades-old promise of empowering everyday workers to build their own tools.

The App Builder and Workflows agents are available starting today through the Microsoft 365 Copilot Agent Store for Frontier Program participants.

Whether that future arrives depends not just on the technology's capabilities, but on a more fundamental question: Do millions of office workers actually want to become part-time software developers? Microsoft is about to find out if the answer is yes — or if some jobs are better left to the professionals.

May Habib, co-founder and CEO of Writer AI, delivered one of the bluntest assessments of corporate AI failures at the TED AI conference on Tuesday, revealing that nearly half of Fortune 500 executives believe artificial intelligence is actively damaging their organizations — and placing the blame squarely on leadership's shoulders.

The problem, according to Habib, isn't the technology. It's that business leaders are making a category error, treating AI transformation like previous technology rollouts and delegating it to IT departments. This approach, she warned, has led to "billions of dollars spent on AI initiatives that are going nowhere."

"Earlier this year, we did a survey of 800 Fortune 500 C-suite executives," Habib told the audience of Silicon Valley executives and investors. "42% of them said AI is tearing their company apart."

The diagnosis challenges conventional wisdom about how enterprises should approach AI adoption. While most major companies have stood up AI task forces, appointed chief AI officers, or expanded IT budgets, Habib argues these moves reflect a fundamental misunderstanding of what AI represents: not another software tool, but a wholesale reorganization of how work gets done.

"There is something leaders are missing when they compare AI to just another tech tool," Habib said. "This is not like giving accountants calculators or bankers Excel or designers Photoshop."

Why the 'old playbook' of delegating to IT departments is failing companies

Habib, whose company has spent five years building AI systems for Fortune 500 companies and logged two million miles visiting customer sites, said the pattern is consistent: "When generative AI started showing up, we turned to the old playbook. We turned to IT and said, 'Go figure this out.'"

That approach fails, she argued, because AI fundamentally changes the economics and organization of work itself. "For 100 years, enterprises have been built around the idea that execution is expensive and hard," Habib said. "The enterprise built complex org charts, complex processes, all to manage people doing stuff."

AI inverts that model. "Execution is going from scarce and expensive to programmatic, on-demand and abundant," she said. In this new paradigm, the bottleneck shifts from execution capacity to strategic design — a shift that requires business leaders, not IT departments, to drive transformation.

"With AI technology, it can no longer be centralized. It's in every workflow, every business," Habib said. "It is now the most important part of a business leader's job. It cannot be delegated."

The statement represents a direct challenge to how most large organizations have structured their AI initiatives, with centralized centers of excellence, dedicated AI teams, or IT-led implementations that business units are expected to adopt.

A generational power shift is happening based on who understands AI workflow design

Habib framed the shift in dramatic terms: "A generational transfer of power is happening right now. It's not about your age or how long you've been at a company. The generational transfer of power is about the nature of leadership itself."

Traditional leadership, she argued, has been defined by the ability to manage complexity — big teams, big budgets, intricate processes. "The identity of leaders at these companies, people like us, has been tied to old school power structures: control, hierarchy, how big our teams are, how big our budgets are. Our value is measured by the sheer amount of complexity we could manage," Habib said. "Today we reward leaders for this. We promote leaders for this."

AI makes that model obsolete. "When I am able to 10x the output of my team or do things that could never be possible, work is no longer about the 1x," she said. "Leadership is no longer about managing complex human execution."

Instead, Habib outlined three fundamental shifts that define what she calls "AI-first leaders" — executives her company has worked with who have successfully deployed AI agents solving "$100 million plus problems."

The first shift: Taking a machete to enterprise complexity

The new leadership mandate, according to Habib, is "taking a machete to the complexity that has calcified so many organizations." She pointed to the layers of friction that have accumulated in enterprises: "Brilliant ideas dying in memos, the endless cycles of approvals, the death by 1,000 clicks, meetings about meetings — a death, by the way, that's happening in 17 different browser tabs each for software that promises to be a single source of truth."

Rather than accepting this complexity as inevitable, AI-first leaders redesign workflows from first principles. "There are very few legacy systems that can't be replaced in your organization, that won't be replaced," Habib said. "But they're not going to be replaced by another monolithic piece of software. They can only be replaced by a business leader articulating business logic and getting that into an agentic system."

She offered a concrete example: "We have customers where it used to take them seven months to get a creative campaign — not even a product, a campaign. Now they can go from TikTok trend to digital shelf in 30 days. That is radical simplicity."

The catch, she emphasized, is that CIOs can't drive this transformation alone. "Your CIO can't help flatten your org chart. Only a business leader can look at workflows and say, 'This part is necessary genius, this part is bureaucratic scar tissue that has to go.'"

The second shift: Managing the fear as career ladders disappear

When AI handles execution, "your humans are liberated to do what they're amazing at: judgment, strategy, creativity," Habib explained. "The old leadership playbook was about managing headcount. We managed people against revenue: one business development rep for every three account executives, one marketer for every five salespeople."

But this liberation carries profound challenges that leaders must address directly. Habib acknowledged the elephant in the room that many executives avoid discussing: "These changes are still frightening for people, even when it's become unholy to talk about it." She's witnessed the fear firsthand. "It shows up as tears in an AI workshop when someone feels like their old skill set isn't translated to the new."

She introduced a term for a common form of resistance: "productivity anchoring" — when employees "cling to the hard way of doing things because they feel productive, because their self-worth is tied to them, even when empirically AI can be better."

The solution isn't to look away. "We have to design new pathways to impact, to show your people their value is not in executing a task. Their value is in orchestrating systems of execution, to ask the next great question," Habib said. She advocates replacing career "ladders" with "lattices" where "people need to grow laterally, to expand sideways."

She was candid about the disruption: "The first rungs on our career ladders are indeed going away. I know because my company is automating them." But she insisted this creates opportunity for work that is "more creative, more strategic, more driven by curiosity and impact — and I believe a lot more human than the jobs that they're replacing."

The third shift: When execution becomes free, ambition becomes the only bottleneck

The final shift is from optimization to creation. "Before AI, we used to call it transformation when we took 12 steps and made them nine," Habib said. "That's optimizing the world as it is. We can now create a new world. That is the greenfield mindset."

She challenged executives to identify assumptions their industries are built on that AI now disrupts. Writer's customers, she said, are already seeing new categories of growth: treating every customer like their only customer, democratizing premium services to broader markets, and entering new markets at unprecedented speed because "AI strips away the friction to access new channels."

"When execution is abundant, the only bottleneck is the scope of your own ambition," Habib declared.

What this means for CIOs: Building the stadium while business leaders design the plays

Habib didn't leave IT leaders without a role — she redefined it. "If tech is everyone's job, you might be asking, what is mine?" she addressed CIOs. "Yours is to provide the mission critical infrastructure that makes this revolution possible."

As tens or hundreds of thousands of AI agents operate at various levels of autonomy within organizations, "governance becomes existential," she explained. "The business leader's job is to design the play, but you have to build the stadium, you have to write the rule book, and you have to make sure these plays can win at championship scale."

The formulation suggests a partnership model: business leaders drive workflow redesign and strategic implementation while IT provides the infrastructure, governance frameworks, and security guardrails that make mass AI deployment safe and scalable. "One can't succeed without the other," Habib said.

For CIOs and technical leaders, this represents a fundamental shift from gatekeeper to enabler. When business units deploy agents autonomously, IT faces governance challenges unlike anything in enterprise software history. Success requires genuine partnership between business and IT — neither can succeed alone, forcing cultural changes in how these functions collaborate.

A real example: From multi-day scrambles to instant answers during a market crisis

To ground her arguments in concrete business impact, Habib described working with the chief client officer of a Fortune 500 wealth advisory firm during recent market volatility following tariff announcements.

"Their phone was ringing off the hook with customers trying to figure out their market exposure," she recounted. "Every request kicked off a multi-day, multi-person scramble: a portfolio manager ran the show, an analyst pulled charts, a relationship manager built the PowerPoint, a compliance officer had to review everything for disclosures. And the leader in all this — she was forwarding emails and chasing updates. This is the top job: managing complexity."

With an agentic AI system, the same work happens programmatically. "A system of agents is able to assemble the answer faster than any number of people could have. No more midnight deck reviews. No more days on end" of coordination, Habib said.

This isn't about marginal productivity gains — it's about fundamentally different operating models where senior executives shift from managing coordination to designing intelligent systems.

Why so many AI initiatives are failing despite massive investment

Habib's arguments arrive as many enterprises face AI disillusionment. After initial excitement about generative AI, many companies have struggled to move beyond pilots and demonstrations to production deployments generating tangible business value.

Her diagnosis — that leaders are delegating rather than driving transformation — aligns with growing evidence that organizational factors, not technical limitations, explain most failures. Companies often lack clarity on use cases, struggle with data preparation, or face internal resistance to workflow changes that AI requires.

Perhaps the most striking aspect of Habib's presentation was her willingness to acknowledge the human cost of AI transformation — and insist leaders address it rather than avoid it. "Your job as a leader is to not look away from this fear. Your job is to face it with a plan," she told the audience.

She described "productivity anchoring" as a form of "self-sabotage" where employees resist AI adoption because their identity and self-worth are tied to execution tasks AI can now perform. The phenomenon suggests that successful AI transformation requires not just technical and strategic changes but psychological and cultural work that many leaders may be unprepared for.

Two challenges: Get your hands dirty, then reimagine everything

Habib closed by throwing down two gauntlets to her executive audience.

"First, a small one: get your hands dirty with agentic AI. Don't delegate. Choose a process that you oversee and automate it. See the difference from managing a complex process to redesigning it for yourself."

The second was more ambitious: "Go back to your team and ask, what could we achieve if execution were free? What would work feel like, be like, look like if you're unbound from the friction and process that slows us down today?"

She concluded: "The tools for creation are in your hands. The mandate for leadership is on your shoulders. What will you build?"

For enterprise leaders accustomed to viewing AI as an IT initiative, Habib's message is clear: that approach isn't working, won't work, and reflects a fundamental misunderstanding of what AI represents. Whether executives embrace her call to personally drive transformation — or continue delegating to IT departments — may determine which organizations thrive and which become cautionary tales.

The statistic she opened with lingers uncomfortably: 42% of Fortune 500 C-suite executives say AI is tearing their companies apart. Habib's diagnosis suggests they're tearing themselves apart by clinging to organizational models designed for an era when execution was scarce. The cure she prescribes requires leaders to do something most find uncomfortable: stop managing complexity and start dismantling it.