Lenfest AI Fellow spotlight: Kevin Delaney

October 19, 2026

This profile is part of the Lenfest AI Collaborative and Fellowship Program’s 2026 Impact Report.

As a Lenfest Institute AI Fellow, Kevin Delaney, editor-in-chief and co-CEO of The San Francisco Standard, is exploring what an AI‑first news interface might look like and how it can deepen reader engagement while supporting a sustainable subscription business. We spoke with Kevin about what he’s learned so far, what the industry gets wrong about AI, and where he believes journalism is headed next.

About his role and the problem he is trying to solve

Kevin, who previously co-founded business publication Quartz, sits at the intersection of newsroom leadership and product innovation. The Lenfest fellowship project gave him the mandate to explore a fundamental question: What would a truly AI‑first news experience look like for subscribers?

The hypothesis: If The Standard could build an app that uses AI to surface the right journalism to the right people at the right time, it would both serve readers better and support subscription growth.

The Standard’s business model is built around subscriptions and sponsorship. It did not have a mobile app before this project.

“We believe that if we could create a good mobile app, it would support the regular engagement that allows you to attract and retain subscribers, and so our focus on the project quickly became: Could we use the mobile app as a way to imagine an AI-first news interface that would support our business goal of growing subscriptions?” Kevin said.

The Standard already used AI in the newsroom for help with research, monitoring news and public information sources, audio and video editing, and other workflows. This project deliberately shifted the focus to the reader interface. It is considered a Mode 2 application of AI, emphasizing learning, experimentation outside of existing workflows, and a shift toward AI-native journalism that can be fine-tuned to user needs. That contrasts with Mode 1, which focuses on optimizing existing workflows and efficiency in core operational areas to drive near-term results.

“We’ve seen really high engagement among subscribers,” he said. “I’ve not been this excited about news product development in a long time.

The biggest surprise during the fellowship

Kevin’s biggest surprise has been just how quickly you can now build products with AI.

The Standard received the $150,000 Lenfest Institute grant in late January 2026 and launched the first version of the app to subscribers in early April, taking roughly three months to get the live product off the ground, in partnership with Bit Bit Press, an external product and AI CMS startup.

The second surprise is how developers are using prompts to rapidly develop products.

“AI prompts are the way that we create new features in the app,” Kevin said. “That’s super exciting, to be able to create new product features in minutes and refine them, whereas until recently, the ability to launch a new product feature was measured in weeks and months.”

For example, a single prompt can instruct the system to:

Identify a person frequently covered by The Standard
Create a card with their photo, name, short bio
Add a button to explore more coverage

That ability to “ship with prompts” makes experimentation much faster and more continuous than traditional feature development cycles.

One thing the industry gets wrong about AI

Kevin sees two opposite but equally mistaken instincts in the industry’s thinking about AI.

On one side, he believes people underestimate how much commodity news AI will be able to handle, such as the rewriting of press releases at wire services. “They’ve been doing this for years. AI effectively repackages them, and that trend can continue deeper into the commodity news area.”

On the other side, he argues that people also underestimate the resilience and value of distinctly human journalism, such as:

Opinion and analysis
Narrative features
Investigative reporting
Local news rooted in proximity and community

“I do think that there are many areas of journalism that are pretty well protected from AI. Looking ahead, human journalists continue to have an important role, and I see news organizations like ours wanting to hire even more of them.”

One lesson from his work that other newsrooms could apply tomorrow

If there’s one structural lesson Kevin would recommend, it’s to formally separate Mode 1 and Mode 2 work:

Mode 1: Daily operations, such as publishing the site, running the newsroom, and hitting today’s deadlines.
Mode 2: Longer‑term, strategic projects, focused on what the organization needs to look like a year or more out, especially in relation to AI.

“The ability to separate the two is really important in this moment, and to be able to think clearly about the opportunities that AI brings, about which parts of the business are positioned to thrive and which are more vulnerable,” he said. “I do believe in the power of the Mode 2 project.”

By treating the AI‑first app as a Mode 2 initiative built with an external partner, The Standard was able to protect it from daily operational firefighting while it was getting the app off the ground. (Aspects of the app today rely on the core Standard team, which will likely take increased responsibility for it over time.)

What he is paying attention to right now

Looking ahead, Kevin is particularly focused on the emerging world of AI agents talking to AI agents, and what that might mean for journalism as a business.

He’s watching:

How news content is being made available to agents
Whether there is a viable agent‑to‑agent marketplace for journalism
If models like micropayments or microtransactions could fairly compensate publishers when agents draw on their reporting

What he is most optimistic about

Kevin is most excited about AI’s ability to turn a city’s journalism into practical, personalized utility for residents.

By combining reporting with signals like location and interest profiles, the app can:

Surface news that’s hyper-locally relevant to each user
Match people to the topics and beats they care about (e.g., Golden State Warriors fans vs. non‑fans)
Help residents navigate different aspects of their city, such as schools, events, and restaurant recommendations from trusted critics

Many of these features were theoretically possible pre‑AI, but for a smaller organization, AI is a powerful enabler that dramatically

The Fast Four

A book, article, or newsletter you’d recommend?
Seeing Like a State by James C. Scott, which examines how individuals and communities interface with large, centralized institutions. “I think it has real relevance now relative to tech companies and how to preserve the specificities of your culture and lifestyle in the face of a homogenizing global force.”

What’s one thing you’re learning that has nothing to do with AI?
Learning how to play the accordion.

What local journalism challenge deserves more attention?
Subscription top‑of‑funnel.

One word that describes the next year of AI in news?
Deep experimentation.