At Principal Financial, multicloud is insurance

Welcome to Runtime! Today: why Principal Financial Group splits its workloads between AWS and MIcrosoft Azure, Google Cloud becomes the latest to offer customers generative AI indemnity protections, and this week in enterprise moves.

(Was this email forwarded to you? Sign up here to get Runtime each week.)

Planning for the future

For many years financial services companies dragged their heels along the road to cloud computing, worried about security and reliability. Principal Financial Group was lucky enough to have started its transition just before the pandemic made the need for modern digital services an existential crisis, and has only accelerated that process since Kathy Kay came on board.

Three or four years before the pandemic began, "we did not have experiences that made it easy to serve our customers in (a digital) way," said Kay, executive vice president and CIO for the Iowa-based conglomerate. "We knew we needed to go to the cloud, because a lot of the capabilities from a data and analytics perspective and being able to build more modern applications and solutions, we couldn't do on-prem anymore," she said in a recent interview.

• Principal started its migration to Microsoft Azure in part because it was a longtime Microsoft Office user, but is now pursuing a multicloud strategy.
• The company now has more workloads in AWS but is willing to put up with the complexity that comes along with multicloud because it wants to be able to pick and choose the services most attuned to its business needs from each cloud, Kay said.
• "The hyperscalers are all trying to build capabilities and some have stronger capabilities than others, but they are constantly improving. Being able to take advantage of that more flexibly has been our mindset," she said.

One understandable reason for the slow movement of financial companies to the cloud was the heightened sensitivity around the personal data customers had entrusted those companies to manage.

• But Principal, like many other consumer-facing companies, also needed to make sure it had a data strategy for finding new customers and retaining current customers after changes to privacy policies on mobile devices made it harder to get that data from third parties.
• "While we might have had (customer data) in the past, it was (dispersed) across the company" thanks to the siloed divisions, Kay said.
• The company evaluated three or four different CDPs before settling on Salesforce, which the company was already using in other ways, Kay said.
• "We were very clear with our teams; let's find the right thing, this is too important," she said. "Just because we have Salesforce doesn't mean you have to use Salesforce. Pick the right solution, and we will support it."

Principal already had a data analysis and machine learning team when the generative AI craze took hold around the beginning of the year, and rather than setting up a new centralized organization to study its potential impact the company decided to form what Kay called a "study group."

• "We had about 70 employees who were passionate about learning," she said, and those folks started generating ideas about how to use the technology both internally and in customer-facing applications.
• Principal also quickly purchased an enterprise license for GitHub's Copilot after more than 250 engineers expressed interest in using the technology to help them code.
• Kay is expecting to detail the company's progress and propose additional internal uses for generative AI at a board meeting next month.

Real the latest edition of "How We Built It" on Runtime.

I'm feeling lucky

Google Cloud joined Microsoft and IBM Thursday in offering legal protections to customers using its generative AI tools to produce content and code. Google also explicitly clarified that its general customer indemnification policy also applies to the training data it uses to build its generative AI products, after it said customers asked for that extra step.

Google's pledge applies to several of its new AI products, including Duet AI and Vertex AI. However, "this indemnity only applies if you didn’t try to intentionally create or use generated output to infringe the rights of others, and similarly, are using existing and emerging tools, for example to cite sources to help use generated output responsibly," Google said in a blog post.

That leaves AWS as the only one of the Big Three that hasn't committed one way or another to offering legal protections for generative AI services. Last month in Seattle, CEO Adam Selipsky declined to address that question directly.

Enterprise moves

Renaud Chaput is the new CTO of Mastodon, where he'll oversee the budding distributed social network's infrastructure strategy.

The Runtime roundup

Atlassian acquired Loom, a video messaging startup, for $975 million in cash with plans to keep it as a standalone brand alongside other productivity tools like Jira and Trello.

Nvidia plans to release new versions of its flagship H100 AI chip every year, instead of the two-year cadence for new products it used to follow.

The SEC will investigate Progress Software's MOVEit vulnerability disaster, which has impacted as many as 64 million people around the world and counting.

ArangoDB announced that it would be adopting the Business Source License for future editions of its database, following a similar decision by HashiCorp to limit commercial use of its previously open-source software.

Thanks for reading — see you Tuesday!