The hidden work (and sweat) powering AI productivity

By Adam Mills

AI Editor

See more of our trusted coverage when you search.Prefer Newsweek on Googleto see more of our trusted coverage when you search.

Newsweek

Editor's note: This is AI Impact, Newsweek's weekly newsletter where each week, we will explore how business leaders are unlocking real value through artificial intelligence.

Seeking Value

An airline refund, a vehicle production plan, an annuity launch and a drug-safety report each leave little room for mistakes. They also illustrate the different demands companies face when putting AI and digital technology to work.

Four interviews in theSeeking Value video series, presented by Cognizant, examine those challenges across industries.

Travelport is preparing for a world where AI assistants could increasingly handle travel purchases. But getting a booking right, processing refunds and managing disruptions requires precision. The company is working with Cognizant and Anthropic to modernize and test the software behind those transactions.

Toyota Motor North America offers a measurable example of technology’s value. Its broader digital transformation, including machine learning for demand forecasting and supply-chain planning, has returned $3 for every $1 invested over roughly five or six years. In conversation with Newsweek, Toyota’s Jason Ballard and Cognizant’s Nishanth Vallabhu share lessons on measuring returns, making the case for investment and bringing employees into the transformation.

Speed was central to Talcott Financial Group’s retail expansion. The insurer connected 12 technology systems to launch three annuity products in approximately one year, then faced the challenge of supporting the platform after it went live. Talcott’s Samir Srivastava and Cognizant’s Haminder Gera discussed what it took to meet that deadline, including a lesson about preparing for the demands that follow a major launch.

Drug safety introduces another consideration: accountability. A conversation featuring GSK and Cognizant examines fragmented safety information, opportunities for automation and a proposed AI-agent model that keeps qualified professionals responsible for regulated decisions.

You can watch the full interviews, moderated by my colleague Gabriel Snyder, Newsweek’s executive editor, enterprise, at the links above.

Signal Capture

Signals from the frontlines of AI adoption

Atlassian: AI Can Save Time—and Create More Work

AI is helping product teams ship faster while requiring them to spend some of the time saved checking agents, reprompting chats and cleaning up generated work.

Atlassian’s “State of Product 2027” survey of 1,000 product professionals found that 80 percent of respondents are shipping faster with AI—but without customers seeing value any sooner. Forty-nine percent said they have had to check agents to keep them on task, 49 percent reported reprompting AI chats and 48 percent retraced an agent’s steps after something went wrong. Sixty-one percent of product managers said AI has created more technical debt. The respondents worked at companies with 500 to 10,000 employees in the U.S., Germany and France.

Axel Sooriah, product management evangelist at Atlassian, said companies need a broader accounting of what AI productivity costs.

“I’d start by counting the cleanup, because if you only measure what AI produces, it will always look like a win,” Sooriah told Newsweek.

Those costs can be easy to miss when productivity is measured through visible output alone. In software development, that can include pull requests—proposed code changes submitted for review.

“Your dashboard can probably tell you how many pull requests went out last month,” he said. “Can it tell you how many hours went into checking the tools that wrote them?”

Review work can also land with someone who has never used an AI tool. Eighty-two percent of the survey’s respondents said team members at least sometimes produce low-effort AI work and revise it based on feedback rather than investing more time upfront, a problem Atlassian calls “work slop.”

Sooriah offered the example of a lengthy specification document arriving shortly before a meeting. It may read well initially until the recipient finds padding and incorrect claims and spends the afternoon repairing something that took far less time to generate. He sees incentives as part of the problem when employees are rewarded for producing more documents, tickets or code without enough attention to whether the material is useful.

“The rule I like is that whoever sends it owns it,” Sooriah said.

Drafting with AI is fine, he said, provided employees read, check and cut down the material before passing it along. Otherwise, the apparent time savings for one person can become review work for someone else.

Sixty-nine percent of respondents said decision-making has not accelerated alongside product development, even as AI makes it easier to produce specifications, prototypes and features.

Sooriah said teams can now draft a specification in minutes or build a prototype in an afternoon. With both taking less time, teams can be tempted to clear backlog times or add features before deciding which ideas deserve to move forward.

“The hard part now is deciding what’s worth building, and for a lot of teams that part hasn’t sped up at all,” he said.

He said the time freed by faster execution is better spent testing multiple versions of an idea, learning which one customers respond to and discarding the rest. Leaders can then judge AI against adoption, retention, revenue or another business measure and track how quickly teams learn whether an idea worked.

“If you’re shipping faster than you’re learning, your customers won’t feel the difference,” Sooriah said.

Upcoming Webinars

What Enterprise AI Looks Like When It Works

On November 4, Dr. Ranjit Tinaikar, host of Newsweek’s “AI Impact Forum,” will sit down with Firdaus Bhathena, executive vice president and chief technology and transformation officer at S&P Global, to discuss how large companies are turning advances in AI into business results, and why trusted, differentiated data is becoming increasingly important.

Drawing on Bhathena’s experience at S&P Global, FIS Global and CVS Health, the conversation will examine how organizations make data governed and reusable, redesign workflows around new capabilities and balance speed with trust, security and resilience.

They’ll also explore how leaders can keep technology investments focused on customer problems and identify where genuine competitive advantage can emerge.

Join the live discussion Wednesday, November 4, at 11 a.m. ET. Register for free.

The AI Election: How Artificial Intelligence Shaped the 2026 Midterms

Newsweek/Getty

What will the 2026 midterms tell us about the emerging politics of AI?

One week after Election Day, Gabriel Snyder, Newsweek’s executive editor, enterprise, will moderate a discussion with policy experts and Washington insiders examining the midterm results through the lens of artificial intelligence.

The panel will feature Nathaniel Persily, professor of law at Stanford Law School; Brent Buchanan, founder and CEO of Cygnal; Nick Penniman, founder and CEO of Issue One; and Andrew Mamo, communications advisor at The Bench.

Using Newsweek’s AI Policy Scorecard as a starting point, the panel will explore where AI policy became a campaign issue, how positions crossed traditional party lines and how debates over regulation, data centers, jobs, privacy and national security played out in individual races.

The conversation will also look at how campaigns and outside groups approached these issues, including the role of industry spending, and what the results could mean for the policy fights ahead.

Register and join the live discussion on November 10.

Enterprise AI in the Physical World: Navigating the Supply Chain’s Messy Frontier

Newsweek

Warehouses, factories and transportation networks put enterprise AI to a different kind of test: digital decisions have immediate consequences in the physical world.

Duncan Angove, chief executive officer of Blue Yonder, joins Dr. Ranjit Tinaikar, host of Newsweek’s “AI Impact Forum,” on November 19 for a conversation about bringing increasingly autonomous systems into complex supply chain operations.

The discussion will cover the role of specialized models, the expertise companies need to preserve as more decisions become automated and whether the economics of enterprise technology could move toward measurable business outcomes such as productivity, operating costs and the movement of goods.

Register to join the live discussion Thursday, November 19, at 12:30 p.m. ET.

Prompt Injection

What’s one recent insight you’ve learned about AI?

David Karandish | Founder and CEO, Capacity

...

“AI can’t handle everything. As much as we push for automation and optimization, there are areas where it falls short, and identifying the threshold at which it’s no longer useful is as important as knowing where it excels. This understanding has shaped the compound startup strategy we follow at Capacity.

AI is genuinely better than humans at solving certain categories of problems, particularly repetitive, rules-based tasks where speed and consistency matter. Where it struggles is with layers of complexity and exceptions that don’t fit a defined pattern. Those are the problems that still require human judgment.

At Capacity, we also recognize there are problems certain companies solve better. They’ve spent years building and refining specific technologies, but they struggle to integrate them into a broader tech stack that delivers a more seamless customer experience. This is where human problem-solving needs to come in.

We’ve spent almost a decade building, acquiring and implementing technologies to address the full customer experience across voice, chat and SMS. That progress required being honest about where we needed help to fill gaps in our product and where we should focus the most human effort.”

Have your own lesson to share? Email me at: a.mills@newsweek.com

Run Log

AI use case of the week

Parts Town and Botifiy Make Product Pages Easier for AI to Read

Product pages built for shoppers can leave some content invisible to automated systems that read websites for search engines and AI tools. Parts Town, a distributor of genuine OEM repair parts for foodservice, residential appliance and HVAC equipment, found that some bots could reach its JavaScript-heavy site while missing information visible to customers.

Linda Ramsey, chief marketing officer at Parts Town, said the company worked with Botify, which helps brands improve their visibility in search and AI systems, to track which URLs Google and AI crawlers visited and which they skipped. Botify also helped provide simplified versions of pages to bots that do not run JavaScript, giving them access to more of the information customers already see.

Millions of URLs created another challenge. Search engines and AI systems do not read every page every day, so Parts Town wanted their limited attention concentrated on pages tied more closely to sales.

“Think of it as a visitor with 10 minutes in a giant warehouse,” Ramsey said. “We wanted them to spend it in the aisles that sell.”

Parts Town reported that clicks to revenue-driving pages rose 31 percent year over year, while OpenAI visits to pages it had prioritized increased 91 percent.

AJ Ghergich, senior vice president of AI and global services at Botify, said retailers should resist redesigning their sites around AI at the expense of customers already using them. Instead, he recommended starting with one important product and checking whether AI systems can find the information needed to surface it, including the product data supplied through feeds.

“If a bot can’t read it, it can’t sell it,” Ghergich said.

Starting with one important product keeps the work tied to a concrete business question: is the information that matters most to customers also visible to the systems increasingly helping them find it?

Have an interesting AI use case to share with me? Email me at: a.mills@newsweek.com

Context Window

■ Alphabet reported $9.01 billion in negative free cash flow as heavy capital spending on data centers and AI infrastructure weighed on results, while Google Cloud revenue more than doubled year over year. [Data Center Dynamics]

■ A new Scale AI and NIST benchmark found that frontier models often answered chemistry questions even when they identified them as potentially dual-use, while their assessments of risk showed significant variation from expert judgments. [FedScoop]

■ Anthropic and OpenAI proposed different approaches for stopping dangerous frontier AI systems in California, with Anthropic backing court-ordered suspensions and OpenAI favoring mandatory shutdown controls enforced by developers. [Computerworld]

■ Bank of America is using model routing to direct queries to more cost-effective AI systems, while Capital One is deploying smaller, fine-tuned models to reduce latency. [Banking Dive]

■ A U.K. government survey of 3,503 businesses found AI adoption doubled to 32 percent from 16 percent in 2025, but most gains remained incremental and only 14 percent considered AI highly or critically important to their operations. [GOV.UK]

Transfer Protocol

Tracking executive moves across the AI landscape

Vassili Patrikis, former chief AI and data officer for Microsoft’s Federal National Security Group, has joined Strider as chief technology and AI officer, leading its engineering, technology and AI strategy and scaling its strategic intelligence platform.

Jason Lew, who most recently served as head of engineering at a stealth AI startup, has joined Parallel Bio as chief AI officer, directing its AI work across the automated lab, data infrastructure and models built on human immune data.

Sami Mian, elevated from executive vice president of innovation and AI and distinguished engineer at PrismHR, has been appointed chief AI officer, responsible for governance, education, enterprise deployment and intelligent capabilities across its technology ecosystem.

Dan Cane, ModMed co-founder and former co-CEO, has moved into the company’s first chief AI officer role, overseeing AI across products and internal operations with direct responsibility for its revenue cycle management AI platform.

Michael Patterson, former vice president of AI data platforms and analytics at Perficient, has joined Definian as senior vice president, Databricks & AI platforms, expanding its Databricks practice and helping customers put governed data to work across analytics and AI.

Tap or click here to get this newsletter delivered to your inbox.

Magic Moment

What’s the most fun or unexpected way you’ve used AI lately?

Eugene Sayan | CEO, Softheon

...

“Two examples come to mind, one at work and one very personal. At Softheon, we named our AI platform HAL, short for Human Augmented Learning. If you remember HAL 9000 from 2001: A Space Odyssey, you’ll appreciate the reference. There’s some humor in the name, but there’s also a serious message for our engineers. When you’re building AI that makes decisions affecting people’s healthcare, you have to think about what can go wrong, not just what the technology can do.

In my personal life, I’ve been using AI alongside my Oura ring to better understand my health and daily habits. During the pandemic, I became diabetic. I changed my eating habits and made walking a priority, targeting 10,000 steps a day.

AI has helped me make sense of the information I’m tracking and think through changes to my meals and routine. Over time, my A1C came back into a healthy range. I’m proud of that. I’ve spent more than 25 years building technology to help people access healthcare, but there’s a difference between talking about technology that can improve people’s health and experiencing those possibilities yourself.

For years, I’ve believed healthcare must focus on keeping people healthy rather than treating them once they’re sick. My experience with diabetes showed me firsthand how that shift actually works.

Anyone can look up a standard diabetic meal plan, but generic advice rarely accounts for daily life. What stood out was AI’s ability to act as an active sounding board. By synthesizing my Oura sleep and activity trends alongside my daily glucose responses, it helped me connect dots I would have missed, like how poor sleep altered my carb sensitivity, or how timing a brisk 15-minute walk right after lunch stabilized my afternoon spikes. It transformed overwhelming, static data into practical, sustainable micro-adjustments.

In healthcare, a 99% success rate isn't enough when high-stakes clinical decisions are on the line, but where AI truly shines is in this personal, proactive space. Imagine a future where intelligent systems don't just alert us to illness, but seamlessly guide millions of people through preventative, daily micro-interventions before chronic conditions ever take hold. That is where the technology's real power lies.”

Experience some AI magic? Tell us about it at a.mills@newsweek.com

Tap or click here to get this newsletter delivered to your inbox.

View 0 Comments

Request Reprint & Licensing

Submit Correction

View Editorial & AI Guidelines

Read More on AI