The Product Model #308: AI Is Rewriting How Teams Work

This week’s updates: Shopify’s engineering culture, PostHog’s self-disruption, AI research guardrails, human-centred design and the coordination cost of AI.

This Week’s Updates

Enabling the Team

Inside Shopify’s high-intensity engineering culture by Bill Doerrfeld
Shopify cuts meetings, pairs engineers, ships weekly and keeps humans accountable for AI-generated code without extending hours.

Good Culture is the Biggest Productivity Hack, Not AI by Gregor Ojstersek
AI amplifies existing culture; clear ownership, priorities, trust and outcome focus are prerequisites for useful productivity gains.

Product Direction

You need to find product-market fit again (sorry) by Cleo Lant
PostHog shows how incumbents can fund small attacker teams, measure pre-PMF bets differently and let usage overrule opinion.

Useful isn’t enough: Why our B2B AI product failed by Govind G Nair
A failed AML AI product shows why solving a real problem is not enough when buyer incentives, packaging, channels and org design misalign.

Continuous Research

Scale Your AI Research Program, Without Losing Strategic Value by Paige Bennett and Stevie Vanderwiel
A fourfold jump in research output did not create strategic value; the fix is charters, workflow boundaries and decision-quality metrics.

One AI Output Is an Example, Not an Evaluation by Raluca Budiu
AI evaluation needs representative inputs, repeated runs and confidence intervals; one good output proves possibility, not reliability.

Continuous Design

Behind the design: Assisted Culling in Adobe Lightroom by Shipali Shetty, Ankur Murarka, Kwamina Arthur and Christopher Sun
Lightroom’s team shares five principles for AI-assisted culling: transparent signals, conservative rejects, local processing and user control.

Am I still designing? Yes, but on a bigger canvas by Ilona Jaudzemyte
An Automattic designer built a WooCommerce image plugin and calibrated an AI judge around product fidelity and merchant trust.

Continuous Development

Staff+ engineers must rebuild team culture by James Socol
AI-enabled solo work weakens shared goals; engineering leaders can rebuild team culture through charters and modelled collaboration.

AI usage patterns in software teams by Tim Qi
AI raises software output but not time saved, as coordination grows, roles blur and more people become active builders.

This Week's Poll

Where should your team move human judgement earlier as AI speeds delivery?

Choose the area where earlier human judgement would make the biggest difference.

Login or Subscribe to participate in polls.

Last week’s poll: Reviewing AI-generated design and code led with 11 of 31 votes (35.5%), followed by validating research and user evidence with 10 votes (32.3%). Setting priorities and decision rights and managing dependencies and security each received five votes (16.1%).

Seen an interesting article online? Share it with us, and we might feature it in our next issue!
Click here to share an article

New Leaders You Will Meet At UXDX SF

Featuring Leaders From
Knapsack, Sephora, NTT Data, Optimal & Docusign

UXDX San Francisco is getting closer, and more leaders have now joined the agenda. The latest additions include Sneha Narahalli from Sephora, Jod Kaftan from NTT DATA, Marina Lin and Amit Sathe from Docusign, Chris Strahl from Knapsack, and Alex Burke from Optimal.

With almost all sessions now announced, the programme is taking shape around how AI is changing product decisions, team roles and delivery.
Explore the growing agenda, choose the sessions most relevant to your team, and use NEWSLETTERSF26 for 10% off tickets here: https://uxdx.com/sanfran/2026/tickets/

Explore the UXDX San Francisco lineup
Use ‘NEWSLETTERSF26’ at checkout to get 10% off.

FREE COMMUNITY EVENTS 

UPCOMING IN-PERSON

6 Oct: Amsterdam
7 Oct: Austin
8 Oct: Dublin
8 Oct: San Francisco

27 Oct: Munich

UPCOMING ONLINE

No published online events at the moment.

Want a UXDX Community event in your city?

Video Of The Week

Bridging Human Insight and AI: How Pinterest Builds Visual Experiences That Connect

In this UXDX USA 2026 talk, Larkin Brown, Senior Director of Product Research at Pinterest, shares how Pinterest is using generative AI to support discovery, shopping, visual search and inspiration while keeping real user needs at the centre. Drawing on examples like Pinterest Assistant and visual search, she explains how research, design and machine learning teams work together to evaluate LLM outputs, use real user queries, and turn foundational research into prompts, principles and product experiences.

Watch the full session to learn how Pinterest is keeping humans in the loop while building the next generation of AI-powered discovery:

Want to go into how careers and leadership are shifting as AI compresses the ladder? My ebook Managing Your Career In The Age Of AI explores how to build judgment, relationships, and influence in a world that keeps trying to automate the surface of the work.