# Ann Miura-Ko

Ann Miura-Ko writes about how companies adopt AI, how AI-native startups organize product and engineering work, and efforts to scale carbon removal.

- Profile: https://b2binfluencer.marketing/creators/ann-miura-ko/
- LinkedIn: https://www.linkedin.com/in/amiura
- LinkedIn followers: 10,984
- Median of last 4 posts: 20 reactions, 1 comments (checked 2026-11-02)
- Topics: artificial-intelligence, product-management, ai-adoption
- Based on 8 dated posts, latest 2026-10-07

## Summary

Ann Miura-Ko is most useful for expert-led sessions on how AI-native startups change product, engineering, and tooling choices.

- Best for: B2B software brands seeking a narrow working session for product, engineering, and innovation leaders.
- Not for: Paid LinkedIn sponsorship or technical buyer education for carbon-removal infrastructure campaigns.
- Check first: Confirm LinkedIn audience demographics and engagement for comparable AI-native organization posts.

### Takeaways

- Across five AI-native companies, she found one full-time product manager.
- Build speed creates a feature-factory risk; constraints protect product taste and coherence.
- AI adoption is cheap, but organizational capability is not, she writes.

## Main ideas

- **Cheap AI adoption does not make an AI-native company**: Ann Miura-Ko argues that buying AI tools is easy while changing how an organization works is hard. The frontier is not more agents but a software factory where anyone can build, the best ideas spread, experience accumulates, and the system starts noticing questions nobody thought to ask. The reasoning is that tool access is not the differentiator; the differentiator is whether an organization can spread good ideas and retain operating knowledge.
- **When engineers own the product, the product manager role can collapse**: Ann Miura-Ko reports that AI-native companies are pushing product decisions into engineering and design. Engineers talk to customers daily and own decisions end to end, so the PM role is absorbed rather than augmented. The coordination stack matters less than the organizational shift: Slack, GitHub, and Linear appear as layers for software agents, Claude Code and Codex are in use, and there is little loyalty to any single tool.
- **Build speed creates a feature-factory risk, and constraints protect taste**: When teams can build almost anything a customer asks for in a day, founders name building everything as a strategic risk. The product can become a Frankenstein assembled from requests rather than a coherent vision. Ann Miura-Ko documents three kinds of constraint: limiting what agents can change, concentrating decision rights, or using squads and north-star metrics to judge agent-generated choices. The argument is that when execution becomes cheap, taste becomes the moat.
- **AI moves the bottleneck away from engineering access**: Ann Miura-Ko argues that AI gives leverage to people outside engineering who were previously blocked by engineering priorities. The same shift compresses experimentation: teams run more tests in parallel, simulate customers earlier, and cover unknowns faster so knowledge compounds.
- **Practical AI adoption outside engineering is a convening priority**: Ann Miura-Ko's public convening emphasizes operating questions rather than model benchmarks: how a regulated company uses AI internally, and what an applied reinforcement-learning environment actually looks like. The focus is on internal adoption beyond engineering teams.
- **Carbon removal has to be considered at large scale**: Ann Miura-Ko frames carbon removal against non-linear growth in energy use and the need for an answer at scale. The post is an endorsement of an industrial project rather than a technical explanation, so it shows a priority and a scale argument more than a detailed climate position.
- **Philanthropy should become a habit early in a career**: Ann Miura-Ko's stated priority is about timing: giving should become a habit when someone starts out, not a reward deferred until after success. The reviewed post is a short conversational note, so it shows the priority but not a detailed program.

## Work to read

- [Five AI-native companies in one day: fewer PMs, faster experiments, and a 'taste' moat](https://www.linkedin.com/posts/amiura_over-the-last-month-we-visited-ai-native-activity-7454990720177926144-gjKf)
- [AI adoption is cheap; organizational capability is not](https://www.linkedin.com/posts/amiura_ai-adoption-is-cheap-organizational-capability-activity-7503235637026824192-pmzt)
- [An event on AI use outside engineering, with Headway and Mercor](https://www.linkedin.com/posts/amiura_secrets-of-ai-pilled-organizations-luma-activity-7519146445225123840-SVg3)
- [Terradot and Google's large enhanced rock weathering project](https://www.linkedin.com/posts/amiura_terradot-and-google-combine-superpollutants-activity-7506075274870943744-E5iC)
- [A podcast conversation on AI, early life, and starting philanthropy early](https://www.linkedin.com/posts/amiura_one-of-the-most-fun-conversations-ive-had-activity-7483554247125479424-TxPQ)

## Brand relationships

- **Terradot** (mention)
- **Google** (mention)

## Partnerships

No sponsorship, paid-partnership, or brand-collaboration offer is documented in the material reviewed for Ann Miura-Ko, so this section records an absence rather than a rate card or a format menu. What the posts do show is someone who convenes and promotes peer-learning sessions: one post advertises an upcoming event on how companies such as Headway and Mercor apply AI outside engineering, with a public event-registration link. That is event organizing, not a published partnership product, and it should not be read as a media kit. The posts also regularly name companies and projects positively, as in a post congratulating the Terradot team on a large enhanced-weathering carbon-removal project with Google. Such mentions are evidence of interest and endorsement, not evidence of payment, and nothing indicates which, if any, of these relationships were commercial. For planning purposes, treat this creator as someone a brand would approach with a point of view to discuss rather than a defined inventory to buy. A sponsor policy, disclosure practice, pricing, deliverables, exclusivity terms, and an inquiry route remain unverified and would need to be confirmed directly before any budget is committed.

## Alternatives

- Brian Dils: Choose Brian Dils when the session needs to move from org design into the practical tooling layer: how design files, design systems, and prototypes must be organized so AI coding tools can use them. The posts argue that AI is becoming a primary consumer of Figma files and give a concrete checklist—group components by domain, use sticker sheets with instances, remove hidden layers, apply semantic layer names, use variables for tokens, and add code syntax to variables. They then extend that into a team argument: replace fixed roles with capabilities and domains when execution gets cheap. That is a better fit if the buyer's real bottleneck is design-to-code readiness or restructuring a product team around capabilities rather than job titles.
- Max Thilén: Choose Max Thilén when the buyer wants an operating playbook for an AI-native marketing team rather than a product and engineering org diagnosis. The posts give a detailed list of practices—self-improving loops, one context layer over business communications, agent access to ad accounts and the CRM, skills instead of how-to docs, orchestration, agents proving their work, safety rules in code, task-based model choice, and human plus agentic quality assurance. They also discuss buying AI seats before prices rise and building recurring automated jobs. That is more useful if the session is for marketing operations and the question is how to run the function day to day.
- Andreas Moekesch: Choose Andreas Moekesch when the buyer's question is enterprise agent governance and the buying process for B2B software. One post describes taking down the demo wall with a free Vevos AI account and argues that buyers should test an ugly real process before procurement; the other compares SAP, ServiceNow, and GitLab agent-governance approaches and asks whether agent inventory and audit trails can be exported and seen across systems. That is a better fit for a session on enterprise adoption, governance ownership, and product-led trial design.

---
AI-assisted summary of public posts. Audience demographics are not available. Engagement appears only where we checked recent posts. Method: https://b2binfluencer.marketing/methodology/