đ” Claude Cowork is a masterclass in modern product management and Antigravity gets Skills
Plus: Google's new Personal Intelligence, Replit lets you ship mobile apps to the iOS store and how Ramp eliminated their traditional product backlog
Hi product people đ,
Welcome back. The second week of January is often when product teams finally ship what they were building before the holidays - and this week delivered some genuinely interesting releases.
Coming up: Claude gives us a masterclass in modern product management with Cowork, Antigravity adds Skills support, and Replit launches a mobile app builder that will not only build your app but publish it to the App Store.
Plus, new data on mobile app spending (downloads are down but revenue is up 21.6% - what that means for your monetization strategy) and fresh numbers on enterprise AI adoption from Harvard Business Review.
As always, feel free to drop any comments about anything featured this week below!
Rich
Watch on YouTube | Follow on Substack Notes
Key reads and resources for product teams
New from the Department of Product Substack this week:
Deep - How Spotify is using AI Agents to manage tech debt and more
With specific, real world examples from the likes of Spotify, DuoLingo, Atlassian, Instacart, DoorDash, Uber and others, if youâre interested in learning about how some of the worldâs leading companies are using AI Agents throughout their product development process, this Deep Dive should help.
New in the AI library - Competitive Landscape Agent Skill
This Skill packages proven frameworks (Porterâs Five Forces, Blue Ocean Strategy and positioning maps) into a practical workflow for assessing industry dynamics, profiling rivals, comparing pricing, and identifying defensible advantages. When you ask Claude to do things like âanalyze my competitorsâ or âhelp me with market positioning,â this Skill will use the frameworks and templates to guide your analysis. (Department of Product)
Industry perspectives - Why the ambitious version of your product roadmap just became the realistic one
Today, your agents can probably work reliably for ~30 minutes. But theyâll be able to perform a dayâs worth of work very soon â and a centuryâs worth of work eventually.What can you achieve when your plans are measured in centuries? (Pat Grady, Sequoia Capital)
Process - How Ramp eliminated their product backlog by building a new AI Agent
Rampâs CPO Geoff Charles says their new coding agent eliminated their traditional product backlog. This piece explains how they did it - and includes details on the open source tool theyâve built to help other product teams do the same. (Ramp)
Predictions - Why product managers are uniquely suited to thrive in this new world
In a thought-provoking take, a former Atlassian and OpenDoor exec argues that product managers are uniquely positioned to thrive in the AI era - and the industry hasnât caught on yet. (Nikunj Kothari, X)
Tool you can use - Openwork
An open source alternative to Claude Cowork that reads your files, creates documents, and automates repetitive knowledge work-all open-source with your AI models of choice. (Accomplish)
UX resource - Googleâs AI Assistant Design guidelines
New product releases and announcements
Claude revealed a new product called Cowork which takes many of the features already available in Claude Code and repositions them for a work context. Some of the use cases outlined by Anthropic in Cowork include things like creating files, crunching data, making prototypes and organizing files. In this example, you give Claude access to a folder on your computer and then Claude will organize the files into a spreadsheet. These use cases were already available in Claude but by repackaging the offering into a less intimidating tool that lives in the terminal, Claude Cowork is hoping that it can build upon the $1 billion dollar revenue run rate that it achieved in December.
Hereâs a quick overview of it in action:
The new product is a fine example of Anthropic doing what top product teams do in 2026. They shipped a product (Claude Code), realised that non-engineers found value in it and used that data to inform an entirely experimental new sub-product with Claude Cowork that takes the utility of Code but repackages it into something that is more suitable to its new user segment.
It was apparently hacked together in just 10 days and may ultimately fail, but AI is making these types of quickfire product experiments much easier to run. Hereâs what one of Claudeâs engineers had to say about the development philosophy behind it:
How to use Claude Code for non-engineering use cases
Cowork is currently only available for users on the Max tier but Claude Code already does an excellent job of helping non-engineers at work. Hereâs some of the use cases you can explore:
PRD and Jira Ticket Writing Agents - create custom AI agents that automatically generate product requirement documents and Jira tickets using pre-configured templates and the /agent command
SEO Audits - use Claude Code with Playwright to visit websites, analyze SEO performance, conduct keyword research, and even create landing pages based on that research
File organization and editing - organize local files using natural language commands (e.g., sorting screenshots by date), edit spreadsheets, convert file formats (CSV to JSON), and bulk edit documents
Second Brain/Knowledge Management - create a personal knowledge system by maintaining âbrain dumpâ folders where Claude Code analyzes your notes, creates mind maps, identifies patterns, and organizes thoughts into categories
Check out the full guide with practical prompts you can use here.
Antigravity now supports Agent Skills and Gemini gets personal; is Googleâs data moat too powerful?
Googleâs answer to Claude Code, Antigravity, now officially supports what was previously called Claude Skills (now renamed to Agent Skills since theyâre an open standard). Their official Docs are worth a read since they include more information about the anatomy of a Skill and how you might want to use them in day to day workflows.
Google also revealed a new feature for Gemini this week called âPersonal Intelligenceâ. Itâs a little surprising that it took Google this long to release a product like this but Personal Intelligence allows users to connect apps like Gmail, Photos, YouTube and Search. As context becomes increasingly important for AI Assistants, Google is sitting on a data moat that ChatGPT, Perplexity and others simply cannot compete with. While these new features are opt-in only, in a sense, adding these new tools could make Gemini too powerful. For now, this is available in the US only and under the EUâs DMA, Google is a designated âgatekeeperâ and cannot prefer its own services which could mean it is delayed before rolling out elsewhere.
Hereâs Googleâs VP of Gemini explaining how it works and how it protects your privacy
And this is what Google product lead Nicole Brichtova is using it for.
Manus launched a partnership with data insights company Similarweb that will see it become a proprietary data provider through MCP. In practical terms, this means that youâll be able to get real time access to competitor data from Similarweb via Manus.
Replit wants you to make production-ready mobile apps, not just prototypes
Replit has launched Replit Mobile Apps - a new tool for building end to end mobile apps quickly. It integrates the front end to a database,authentication and third party apps APIs like OpenAI, Apple Maps, Twilio and more. When an App is ready, it can be pushed to the App Store.
Replit is positioning these apps as productionâready rather than prototypes. Still, itâs arguable that vibeâcoded mobile apps are unlikely to meet the needs of larger, more complex product organizations. For smaller teams that donât yet have a mobile app, though, this looks like a neat way to get prototypes or experimental apps into usersâ hands quickly. As most mobile product teams will attest, the real work often begins once an app ships to production.
Theyâre also hosting an event later today to showcase it which might be quite interesting.
đ Product data and trends to stay informed
Global consumer spending on mobile apps reached $155.8 billion in 2025, up 21.6% from 2024. But, downloads decreased by 2.7% to 106.9 billion, down from 109.8 billion in 2024 and 113.6 billion in 2023. With downloads declining but revenue up 21.6%, the data suggests that acquisition volume is becoming less important than monetization effectiveness.
Non-game spending grew 33.9% and now exceeds gaming revenue. For product teams in productivity, health, finance, entertainment, or other non-game categories, this validates that users are willing to pay significantly for utility and ongoing value.
A new survey published in Harvard Business Review found that 99% of leaders say that investments in data and AI are a top priority with 38% of companies confirming that theyâve appointed a Chief AI Officer.
Other nuggets that might be of interest to product teams include:
39% report AI is now in production at scale, up from 5% two years ago
54% have implemented AI in limited production, up from 24% two years ago
94% are beyond the pure experimentation stage, up from 29% two years ago
Despite gains from Anthropic, OpenAI remains the leading model provider of choice in the latest Ramp Index - and theyâve increased their share to almost 37% for the first time.
According to a new report, the smart glasses market is set to quadruple in 2026 with sales volume rising from 6 million units in 2025 to 20 million units in 2026. Metaâs new layoffs tell a similar story; theyâre cutting 10% of Metaverse staff and instead reallocating spend to their smart glass division.
49% of jobs now see Claude used for at least a quarter of their tasks, up from 36% in January 2025 according to the latest Anthropic Economic Index Report.
Paid subscribers get the full DoP Substack including: The Knowledge Series for sharpening your tech / AI skills, the AI Prompt and Skills library and DoP Deep dive reports for in-depth analysis to learn lessons from the worldâs top tech companies.