The Morning Download: Principal Financial’s AI Surge

Good morning. Principal Financial Group CEO Deanna Strable says AI adoption is soaring within the financial services company.

The 146-year-old company based in Des Moines has about 17,000 active AI users, or 85% of its 20,000 employees. That’s up from 800 active users, or 4% of the workforce, early last year.

The AI surge reflects a concerted effort by leadership. Strable and CIO Kathy Kay met with members of the WSJ Leadership Institute team and discussed their approach to integrating AI into the company. Today’s Leadership Outlook below features highlights of that conversation.

Principal Financial’s use of AI is deeply embedded into the company’s broader strategy, according to Strable, a company veteran who took over as CEO in January 2025.

The investment management and insurance company is evolving from its business-to-business model to incorporate business-to-consumer efforts as well, according to Strable. “Ultimately, how you interact from a technology perspective is very different. You have to think about 24/7 versus interacting during the day. Personalization becomes much more important. And so one of our big strategic initiatives is continuing to build that piece out,” Strable said. That’s where AI plays a significant role.

AI? What AI? Principal’s AI applications range from predictive machine learning used by asset managers to generative AI tools for administrative relief. For example, Principal developers utilize GitHub Copilot, while the broader workforce relies on tools like Microsoft Copilot and internal platforms. The company uses Moveworks to power an internal AI assistant named Penny, which handles a range of HR and IT questions.

AI has dramatically improved workflows in call centers and benefits departments by instantly summarizing dense policy documents, saving human agents from the need to manually sort through disparate, older systems. The company can build platforms internally on top of Amazon Bedrock, according to Kay. The company tends to be an AWS shop, Kay said.

A self-funding model. Implementing enterprise-wide AI is notoriously expensive, according to Strable, who said she leans heavily on her background as a former CFO to manage those costs effectively. In 2025, the company invested between $20 million and $25 million directly into AI tools, seat licenses, training, and staff expertise, without going to Wall Street to report higher expenses or to lower earnings forecasts, she said.

The company saved and reinvested money by utilizing internal AI rather than hiring external vendors, she said. Some savings fell straight to the bottom line, some were funneled into new business initiatives, and some paid for the $25 million AI expansion, according to Strable.

Today, at least half of Principal’s tech portfolio has an AI component, making AI a part of the standard operational budget, according to Kay.

Use it or lose it. Principal expects employees to use the tools at hand, and it won’t let expensive AI investments go to waste. “If you haven't used a model in 30 days, we'll take the license. You can then easily get it back,” Kay said. “So, there is an expectation that you use the tools that you've been trained with to be more productive.”

The company also tracks who its biggest AI users are, as well as the volume of code written by AI.

“We have some goals that would not be achievable if you're not using AI,” Kay said. In some instances, those who don’t use AI will “definitely stand out as being not as productive as others,” she said.

Strable said her primary goal is to maintain Principal as a growing company. By making the existing workforce massively more productive through AI, Principal can service a growing customer base and increase revenue without needing to expand headcount at the same rate,” she said. While some high-turnover areas like call centers might see "right-sizing" through attrition, the main goal is increasing throughput.

“It’s getting us to think differently about change management, and how you design jobs or put jobs together,” Strable said.

The profile of the ideal job candidate is evolving along with AI, according to Kay. Pure technical prowess is no longer the sole metric for success. The company increasingly hires for creativity and adaptability, candidates who may not be "super engineers" in a traditional sense, but who are highly creative when it comes to leveraging AI tools to write code and solve problems.

Consultants are surviving the AI threat. Artificial intelligence initially threatened to replace consultants. Now, it is giving them a boost—at least for a little while, says the Journal's Allison Pohle.

But it's not consulting as usual.  As the WSJLI reported, clients are less interested in paying for a large number of junior associates to collect and synthesize data. Instead, more agreements are tied to outcome-based pricing—where a firm is paid partly based on whether a project gets a specified result—rather than how many people they throw at it.

For tech leaders driving AI efforts, this moment is a chance to seize leverage in consultant relationships.

AI work has increased demand for consultancies, according to K2 Consulting Research, which tracks the industry. The firm’s research found global consulting grew 5.5% in 2025, double the rate of the prior year.

Amazon backs Anthropic on cloud. Amazon on Friday joined Microsoft and Google, telling customers it will continue offering Anthropic's Claude to cloud customers, except for Defense Department work, CNBC reports. The move follows the Pentagon labeling Anthropic a "supply chain risk." Amazon has invested $8 billion in the startup since 2023 and its cloud unit is Anthropic’s primary cloud and training partner.

Oracle and OpenAI scrap plans for data center expansion. The two companies have walked away from a 600-megawatt expansion deal for a  flagship 1.2GW data center in Abilene, Texas. Meta is in early talks to take over the space. The FT says the decision may reflect moves by OpenAI to diversify its infrastructure. 

OpenAI's robotics chief quits. Caitlin Kalinowski cited the company's Pentagon deal, warning that domestic surveillance and autonomous weapons deserved more deliberation, Bloomberg reports. OpenAI defended the agreement as responsible and said it “creates a workable path for responsible national security uses of AI while making clear our red lines."

Nvidia backs Nscale. AI data center startup Nscale raised $2 billion at a $14.6 billion valuation, with Nvidia among the participants, CNBC reports. The UK-based company, founded in 2024, has partnerships with Microsoft and OpenAI, among others. 

DOGE unleashed AI on the humanities. The New York Times reports how DOGE employees used ChatGPT prompts to identify and cancel over 1,400 National Endowment for the Humanities grants, flagging projects ranging from indigenous language archives to a history of American music to Holocaust documentaries. Lawsuits filed by major scholarly organizations argue the cuts violated the First Amendment and discriminated based on race, ethnicity, and gender. 

Game on. Nintendo is suing the US government seeking a refund for tariffs it calls unlawfully collected. Thousands of companies are pursuing roughly $170 billion in refunds after the Supreme Court ruled that Trump's tariffs were illegal, Bloomberg reports.

One week into President Trump’s war on Iran, the most severe shock to energy markets since the 1970s is cascading through the world economy. (WSJ)

Brent crude, the global oil benchmark, surged nearly 30% overnight before paring some gains. A G-7 meeting set for Monday, where finance ministers are set to discuss the possibility of releasing oil from reserves, quelled some concerns about global supplies. (WSJ)

A WSJ investigation tracked the U.S. citizens caught in the crosshairs of an aggressive government campaign to detain and demonize dissenters. (WSJ)

The WSJ CIO Journal Team is Steven Rosenbush, Isabelle Bousquette and Belle Lin.

The editor, Tom Loftus, can be reached at thomas.loftus@wsj.com.