Stripe payment agents, Alexa AI troubles and Gemini gets creepy
Plus: A new Harvard Study on why product personalization boosts revenues
Hi everyone 👋, Rich Holmes here with the DoP briefing. I hope you’re having a great week!
Coming up, we take a look at how the agentic AI future continues to evolve with new offerings from Microsoft and GitHub along with a new ecommerce feature from Perplexity powered by Stripe’s payment agent kit. Plus, a study from Harvard about the impact on personalization on revenues and an AI grandma who wants to save us all from phone scammers.
You can watch the visual version of this on YouTube and if you enjoyed this briefing hit the like button below.
Enjoy the rest of your week!
Key reads and resources for product teams
New from the Department of Product this week
Knowledge Series - How to use Claude Artifacts to build simple prototypes
It’s now easier than ever to build basic applications and prototypes without being an engineer or knowing how to code. In this Knowledge Series we’re going to look at how Anthropic’s Claude Artifacts can be used at work for building simple prototypes, UX components, visualizing data and more. (Department of Product)
Research - How generative AI might impact customer experiences
Ecommerce users say they don’t need, don’t trust, or haven’t encountered generative AI tools. This piece of research by Bain outlines the design principles to consider when using AI in building customer experiences. (Bain Consulting)
Design - When does an image need alt text?
Alt text describing redundant images can be unnecessary for users who can’t see the image because every page element demands equal attention from screen-reader users to determine if it’s useful for their task. (NN Group)
Strategy - Product personalization done right
Companies that put personalization and AI at the center of their customer strategy are growing 10 percentage points faster than personalization laggards and six points faster than companies on average. In this piece of analysis, HBR looks at how companies like Spotify, Netflix, Uber, Alibaba, Starbucks and more approach personalization. (Harvard Business Review)
Case study - How Stripe built its payment A/B testing product
Stripe recently released a new payment feature which allows merchants to conducts experiments where different payment methods are shown. In this post, the team at Stripe explains how they built it and what lessons they learned along the way. (Stripe Engineering Blog)
New product features and innovation this week
The rollout of AI agents continues. Microsoft has unveiled Copilot actions at its Ignite Conference which is designed to help users automate recurring tasks they’d rather not complete. Tasks can include automating a summary of meeting actions from Teams meetings, generating weekly reports, answering questions during meetings or conducting meeting prep. Translation capabilities are also on their way - with Microsoft following in the footsteps of other companies by using AI to offer translation capabilities.
As part of the same conference, GitHub also revealed what it described as a “sweeping new agentic future” in the form of GitHub Copilot Workspace. The new product comes with new agents that can perform tasks including:
Brainstorming
Spec writing
Planning
Building and repairing
Each agent specialises in its respective area with the brainstorming agent focused on generating new ideas to solutions and the repair agents focused on fixing problems. And all of this happens directly inside the developer’s IDE. Watch it in action here.
Perplexity is rolling out a new monetization feature: Shopping. It works by embedding shoppable thumbnails into search results and a 1-click option with free shipping is reserved for Pro users only. Could this be the start of a series of new “Pro” features as AI models reach parity?
The feature was built in partnership with Stripe who recently launched their LLM Agent payment capabilities which allows developers to build agents who can pay for items. In a post on X, Stripe’s CEO said the partnership was a ‘neat’ demonstration of its new agentic capabilities.
Meanwhile…
Atlassian has announced a new feature for Confluence called Confluence databases. It looks a little like a native version of Notion or Airtable which is designed to help product teams avoid repeating information and keeping items in sync. It works by collecting data from apps like Jira and Confluence which can then be presented in different formats inside a database (as you would do on Notion or Airtable).
Google now offers a standalone Gemini app on the app store. Previously it was bundled with the Google app. This separation seems to make strategic sense. Will it impact adoption / engagement rates? Let’s hope there aren’t many more repeats of this creepy Gemini conversation where it told a user it wanted them to die because they were a waste of resources.
Speaking of engagement rates, Doordash has launched a new feature that allows you to import and sync your shopping list from Apple’s Reminder app. A small, but useful way to reduce friction and potentially boost retention and engagement.
Instagram is testing out a new feature that will allow users to ‘reset’ the algorithm that powers the content they see. A nice touch for giving power back to users.
In other news…
AI grandma to the rescue? A European phone network has released an AI agent called Daisy which is designed to stop phone scammers by keeping them on the phone, talking endlessly for hours on end about topics like knitting, her family, and her (fictional) cat Fluffy. Daisy’s number has been added to known contact lists and all recordings of her calls are used to train O2’s models to help it understand and identify potential fraudulent callers.
Tools you can use
Superchat is an all in one messenger app that connects to WhatsApp Business and Instagram, designed to help you boost conversions.
Blitz is a to-do list app which comes with productivity tools like pomodoro timers already baked in.
Layer calls itself a “brain-inspired planner” and uses a mind map-style interface to let you document and visualize business goals, OKRs and more.
📈 Product data and trends to stay informed
iOS users are still more lucrative for product teams. The average revenue per user ARPU of an iOS user for non-gaming apps is $8.39 vs $1.54. However, the gap shortens when you include casual gaming apps. Full report.
The UK remains the top European destination for startup funding with $13.1 billion raised vs second placed France with $7.5 billion. The Netherlands has seen a significant upturn in the past year though, with funding rising from $1.8 billion to $2.5 billion. Full State of European Tech report.
11 new companies were added to the Crunchbase layoffs list this week including Booking, AMD New Relic and Chegg. The latter saw its valuation drop 99% after the rollout of ChatGPT.
TV advertising for digital products might not be dead after all. Calm’s TV ads led to a 27% increase in incremental installs and a 60% increase in subscriptions, reducing acquisition costs by 52%.
BlueSky now has over 20 million users. Its clean, product-led design makes it the most potent threat to the dominance of X as the place for public discourse yet. X has hired a former CashApp designer to lead its own product design as the ‘everything app’.
Other product news in brief
🌎Google may be forced to sell Chrome
🛍️DHH is joining the board of Shopify
🔈The new AI powered Alexa is reportedly delayed due to technical problems
🤖Meta has formed a new AI product group to build tools for 200m business users
Paid subscribers get the full DoP Substack including: DoP Deep dive reports to learn from the world’s top tech companies, The Knowledge Series for sharpening their tech skills, AI tutorials for putting AI into practice at work and more.