Adam Mills

By Adam Mills

AI Editor

0

Share

Newsweek is a Trust Project member

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.

Read More on AI

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

Good morning and thanks for joining me.

A pattern has kept surfacing in my recent conversations with executives: The closer AI agents get to acting on someone’s behalf, the more specific the questions become.

With Nandan Sheth, CEO of Splitit, the question was: What has to happen when an agent reaches checkout, and how much authority should it have? Bill Robbins, CEO of Menlo Security, put it in security terms: What happens when an agent can act without a human checkpoint? Mike Keogh, president of Provalus, raised the operational version: When AI prioritizes fraud alerts, which decisions still need a person to investigate, question or escalate?

I’ve started to see those conversations as pieces of the same problem. The more work we hand to an agent, the more important it becomes to define what it is allowed to do, what still requires a human touch and how control gets returned when something goes wrong.

That was the backdrop for my latest conversation with Mike Storiale, senior vice president of AI technology and transformation at Synchrony, a consumer financing company.

His focus is on shopping and payments, where consumers are already beginning to use AI as a conversational way to find products and weigh their options.

The part I keep thinking about is what happens when the agent starts influencing which options the customer even sees to begin with.

Read on for my conversation with Storiale about where that leads next.

Signal Capture

Signals from the frontlines of AI adoption

Synchrony SVP: AI Agents Are Changing How Consumers Choose to Pay

Ask an AI shopping agent for the best way to pay for a refrigerator and the answer may depend on far more than price.

“What is best? Best for you might mean lowest monthly payment. Best for me might be the thing that gets it to me fastest and I can pay cash,” Storiale told Newsweek.

An agent making a recommendation has to understand both the available payment options and the person it represents. Storiale said consumers are already using generative AI conversationally to search for products and ask questions such as how they can afford a purchase. A retailer or financial institution may no longer control the first description a consumer sees of its offer.

Payment and financing options need to be understandable to the AI systems comparing them. Financing terms, rewards, loyalty benefits and disclosures have to remain clear when an agent interprets an offer instead of a person reading it directly.

Storiale identified two risks when an offer is confusing. The agent could explain the product incorrectly or decide that an offer is too complicated and recommend something else.

Required disclosures add another layer.

“You want to make sure that everything that was disclosed transparently upfront continues to be disclosed transparently in this experience,” he said.

An agent’s past experience with a payment or financing option could also influence whether it recommends that option again. A system with memory could learn which payment methods a consumer prefers, which retailers they use and whether previous choices worked for them.

“If you burn that bridge today, I don’t have any data points to tell you how long it takes to get back, but I imagine it’s not immediate,” Storiale said.

Giving an agent permission to complete a transaction creates a problem traditional checkout does not have: The consumer may no longer be the one directly initiating it.

“We need to know who’s representing the customer and what the customer said that agent could do for them,” Storiale said. “What was the intent?”

Storiale called identity and intent two of the more complicated pieces because traditional commerce has not required companies to establish them in the same way. A person standing at a register provides a clear signal that they intend to buy something. An agent could act later, after a consumer has already given it instructions.

Consumers will have to decide where an agent can act independently and where it needs to come back for approval.

“What you’re trading is you get time back and you get mental capacity back,” Storiale said. “And in exchange, you’re delegating some of your decision-making power.”

The boundaries could vary by price or by purchase. An agent might be allowed to buy an item once its price falls below a certain amount, while another transaction could require final approval.

Synchrony is using its shopping-agent work to compare what consumers say they want with what they actually do.

“Shopping maybe for me is a chore, I’m not a person who likes shopping, I would delegate it all to an agent,” Storiale said. “And then you give them that access, and anxiety kicks in.”

Storiale said Synchrony wants to understand when that discomfort appears and what causes it, then use those lessons as it develops products and works with its retail partners.

Financial institutions can start learning now how AI systems present their products, Storiale said. He recommends testing what appears when consumers ask about affordability, payment options or financing and checking whether the information returned is accurate.

“If you’re saying, ‘I’ll figure it out later,’ later is going to be when it’s at scale,” he said. “And that means that you’re going to have a much larger hill to climb.”

Upcoming Webinars

Overcoming Barriers to AI Transformation in Legacy Industries

Newsweek

AI investment is accelerating, and long-established companies are finding ways to bring the technology into organizations shaped by years of customer relationships, complex operations, critical systems and regulatory responsibilities. The challenge is turning promising pilots into lasting business results without disrupting the strengths that have made those companies successful.

In an upcoming Newsweek AI webinar presented by Cognizant, Gabriel Snyder, Newsweek’s executive editor, enterprise, will moderate a discussion titled “Overcoming Barriers to AI Transformation in Legacy Industries.” The conversation will delve into how established organizations are connecting AI to existing systems and workflows, clarifying ownership and governance and aligning technology investments with people, processes and measurable business outcomes.

Matt Sanchez, chief operating officer at Yahoo, will join the conversation, with additional speakers to be announced.

Join the live discussion on October 21. Register for free.


The Inflection Point: How Agentic Operations Compound Enterprise Value

The next phase of enterprise automation is moving beyond tools that assist employees and toward systems that can execute parts of a business process themselves. That raises bigger questions for leaders about where people remain accountable, how operating models need to change and what happens to the economics of work when machines take on more of the execution.

On September 17, Dr. Ranjit Tinaikar, host of Newsweek’s “AI Impact Forum,” will sit down with Balkrishan “BK” Kalra, president and CEO of Genpact, for “The Inflection Point: How Agentic Operations Compound Enterprise Value.”

Kalra will bring Genpact’s perspective on what it calls “Agentic Operations,” including lessons from its own implementation and research into why some enterprise investments produce value while others stall.

They’ll get into the implications for jobs and roles, responsible AI controls, operating models and commercial structures as companies give autonomous systems a larger role in how work gets done.

Join the live discussion on September 17. Register for Free.

Prompt Injection

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

Harshal Dedhia | Vice President of AI, Demandbase

...

“(One recent realization is) AI hasn't just accelerated software development, it's fundamentally shifted where the most important work happens. Tools like Claude Code, Cursor, and Codex have dramatically reduced the time it takes to turn a product idea into something real. As implementation is no longer the constraint, the bottleneck shifts to either side of development: upstream to defining what to build, who it's for, the outcomes it should achieve, and the right design approach, and downstream to testing, validating, and proving its value.

That changes how we think about product development. Instead of asking how we can ship more features, we're looking to AI and our teams to help us make better product decisions up front and validate outcomes more rigorously afterward. Design, implementation, and validation all remain critical, but as the pace of innovation accelerates, solving the right problems becomes more important than solving them faster. That makes product engineers, and product experts who deeply understand the user, increasingly valuable. Otherwise, organizations risk building AI-generated products that are faster to create but fail to solve meaningful customer problems, wasting resources and eroding trust in the product.”

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

Run Log

AI use case of the week

Pie Measures AI by Calls, Bookings and Sales

Syed Ali | Co-Founder and CEO, Pie

...

Business owners told Syed Ali they wanted calls, bookings, walk-ins and revenue, along with clear evidence that their marketing spending produced them.

Ali, co-founder and chief executive officer of Pie, a company that uses AI to help small businesses attract and respond to customers, said managing all the places where those customers might discover a business was another matter. Local discovery now spans Google, Maps, Yelp, AI search and other digital channels, while many business owners lack the time or expertise to manage each one. Some had also paid outside marketing firms without gaining a clear connection between clicks or impressions and sales.

Pie uses AI at several points between someone looking for a local service and contacting the business. Its Growth product uses AI to build targeted ad campaigns around high-intent keywords, then adjusts spending as those campaigns run. AI Search focuses on visibility in AI-powered search results. Front Desk can answer phone calls, take bookings and respond to customer questions.

The system uses information such as a business’s services, location and hours, along with performance data, to decide where marketing activity should run and adjust it over time. Ali said Pie was built to handle more of that work without requiring owners to manage it themselves.

“The goal was never to give owners another dashboard,” Ali said.

Ali said Pie reached roughly 2,000 small-business customers while operating in stealth, largely through referrals and platform partnerships. He said the company has driven more than 150,000 phone calls to client businesses and that customers typically see year-over-year sales increases of 15 to 20 percent.

The work stretches from helping a business appear when someone looks for a local service to answering when that person calls. The numbers Ali provided put the focus on business results: calls generated for clients and the sales growth that Pie says its customers typically see.

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

Context Window

■ Tech workers at OpenAI, Anthropic, Meta and Google described long hours as their companies accelerate AI development, while UC Berkeley research found employees using AI worked faster, took on more tasks and extended work into more hours of the day. [BBC]

■ Nvidia signed agreements with Apollo, Blackstone, BlackRock, Brookfield, Goldman Sachs and KKR to establish financing platforms aimed at mobilizing more than $500 billion in third-party capital for AI data centers and compute infrastructure. [CNBC]

■ IBM and OpenAI announced a strategic partnership to embed GPT-5.6, Codex and ChatGPT Work into IBM Consulting Advantage, with joint industry solutions and a new OpenAI practice focused on deploying AI across enterprise workflows and cybersecurity. [IBM]

■ Congressional offices are increasingly using AI for research, drafting and communications, while House and Senate rules remain limited, poorly understood and largely left to individual offices to interpret. [The Washington Post]

■ U.S. intelligence agencies are building frameworks for networks of AI agents that can coordinate across intelligence and operational tasks, with DIA, NGA and FBI officials emphasizing governance, monitoring and human control as they develop the underlying infrastructure. [Breaking Defense]

Transfer Protocol

Tracking executive moves across the AI landscape

Vishal Vallabha, most recently founding partner and chief technology officer at NexGen.ai and a strategic AI adviser to Powerfleet, is joining Powerfleet this week as chief AI officer after previously holding technology and AI leadership roles at Freeman Company and Lumen Technologies.

Chandhu Nair, previously senior vice president of stores, data, AI and innovation at Lowe’s, will join Target on Aug. 24 as senior vice president and the retailer’s first chief AI officer, bringing greater focus and coordination to AI across the enterprise.

Preeti Shukla, an enterprise SaaS and applied AI leader with previous leadership roles at Workday and Oracle, has been appointed chief product and AI officer at Brightfin, where she will lead product strategy and artificial intelligence across the platform and own the next phases of the company’s AI roadmap.

Chetan Gupta, who spent nearly a decade leading global AI research at Hitachi, has been appointed chief AI officer at Rackspace Technology, where he will lead AI strategy, research, innovation, governance and adoption across the company and its customers.

Jake Bouvrie, previously head of research and development, digital and AI at Nestlé, has been appointed Insurify’s first chief AI officer, where he will lead AI strategy focused on conversational insurance shopping, marketplace optimization and personalization.

Know someone on the move in AI? Send job change info to a.mills@newsweek.com

Magic Moment

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

“We have years of proprietary data from artists who have used tools to mark up professional TV shows with manually inserted brands using VFX. These specialists have an eye for quality well beyond a lay person, and they can sometimes judge with sub-pixel accuracy. But you don’t always have such a specialist on hand for every move you make, especially if you are using AI to automate the brand placement process in the first place.

So it occurred to us that, given our years of proprietary data, we could train an AI judge to quickly and cheaply respond to AI brand insertions with an eye to identifying potential flaws or room for improvement. The unexpected approach here is that we could use one agent to automate a process and another agent to judge and improve it, with the caveat that those systems need to be trained and built on different data assets with different goals in order to be useful.”

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