Best of LinkedIn: Digital Products & Services CW 38/39 2026

AI has made building faster and cheaper, and the edition's central finding is that the constraint has moved to judgment, discovery and organizational design. Product roles are converging on builders who can prototype, yet the evidence points to teams, accountability and clear ownership as the scarce inputs. Meanwhile, operating models, not tooling, separate the organizations that scale AI from those that simply move faster.

Date

September 30, 2026

Digital Products & Services

Thomas Allgeyer

Methodology: Every two weeks we collect most relevant posts on LinkedIn for selected topics and create an overall summary only based on these posts. If you´re interested in the single posts behind, you can find them here: https://linktr.ee/thomasallgeyer. Have a great read!

Listen to our podcast

If you prefer listening, check out our podcast summarizing the most relevant insights from Digital Products & Services CW 38/ 39

AI and the reshaping of product roles

Prototypes and customer previews are replacing detailed specs and review gates
The one person AI product team has not materialized, as judgment still needs teams
Product work is moving from an assembly line of risk gates to a jazz band
AI replaces activities such as user stories and backlog checks, not accountability
A 1931 Procter & Gamble memo on one accountable owner per brand still defines the PM job
Forward Deployed PM roles are spreading from Palantir to Scale AI, Salesforce, Glean and Cresta
A multimodal insurance agent was built in three to four hours with a coding agent
Guardrails, not walls, let product, design and engineering blur safely
Company wide AI use creates a coherence problem, making standards the real maturity test
Bottlenecks keep moving in an AI accelerated world, so visible systems beat prediction

‍

Product discovery, strategy and decision quality

Evals now pay off: Ramp went from 35% to 83% accuracy and Cursor cut cost by 41%
Shopify's AI workflow builder is 2.2x faster and 68% cheaper, while Harvey nearly doubled its quality score
Nearly half of 25 shared PM openings ask for experience writing evals
Minimum Viable Quality replaces the question of what accuracy is needed
Cohorts exposed retention falling from 51% in January to 31% in May behind a 42% average
Discovery must get faster and cheaper because decision making is now the constraint
Customers describe the problem in front of them, so discovery depends on extrapolation
A voice AI product reached around 50 Bay Area restaurants before proving a vitamin, not a painkiller
Vague targeting is blamed for 90% of digital product failures before launch
Product discovery now spans Amazon, Google, social, creators and AI shopping, all needing clean product data

‍

Product operations and organization design

A PM team that doubled in August codified its operating model into an agentic toolkit
AI cannot fix project based delivery, blurred ownership or unclear accountability
Product operations is an organizational design challenge, not a documentation exercise
Many leaders cannot say how much R&D spend maps to each strategic priority
Overinvolved CEOs are named as a costly bottleneck in product organizations
European PMs are often accountable for scope and timeline but not outcomes or decisions
The right PM profile changes between $2M, $10M and $30M ARR
A startup first PM role offers $110,000 and 0.25% equity against $135,000 and a 10% bonus in enterprise
Flatiron cut quarterly feedback processing from four people over four weeks to one person over one week
Companies replacing third party procurement systems with in house AI products must prove business value

‍

Launches and company news

Qantas opened its Product Innovation Centre in Adelaide with 50 team members already hired
Royal Mail relaunched its locations experience for millions of UK consumers on a new front end platform
Votito launched after about a year of development to improve product management decisions
Noodle Seed went live on Product Hunt, adding a conversational layer reachable from ChatGPT and Claude
AI powered Flows turn workshop sticky notes into prioritized backlog items
PLUS Retail is shifting its digital products from personalization toward context
Germany's 2027 e invoicing deadline belongs on ERP and accounting software roadmaps now
Design decides whether Digital Product Passports earn engagement beyond a QR code
67% of B2B buyers prefer a rep free experience, making founder led thought leadership a trust lever
Future Product Days, ProdCon26, POPSSF, Product Lab and TechTides Derry centered on AI and judgment

‍

Thanks to Sachin Rekhi, Christoph Steinlehner, Ravi Mehta, Helene Amster Kopel, Ohad Biron, Jyothi Nookula, Moe Ali, Sowmya Sampath, Victoria Sheer, Marianne Elizalde, Lenny Rachitsky, Marily Nika, Pankaj Zanke, Janina Franzkowiak, Teresa Torres, Shounak Mitra, Dickie Bush, Seth Hurd, Nikita Chepuri, Aakash Gupta, Brian Muskoff, Kelvin Dart, Barry O'Reilly, Genia Jacques, Mackenzie Hughes, David Pereira, James Jordan, Kia M., Kassim Alani, Roman Pichler, David Pizzey, David Gross, Gojko Adzic, Ali Hasan Butt, Dr Bart Jaworski, Arvid Nieuwsma, Oriol Bausà, Thomas L. Roedding, Iqra Khalid, Arne Kittler, Alastair Cameron and Joel Eilat, along with everyone else who contributed insights to this edition.

Find the full list of posts and voices on LinkedIn: https://www.linkedin.com/pulse/best-linkedin-cw-38-39-digital-products-services-thomas-allgeyer-4kg8e/

Subscribe to newsletter

Subscribe to receive the latest blog posts to your inbox every week.

Thank you! Your submission has been received!

Oops! Something went wrong while submitting the form.

Frequently asked questions

If AI makes building software cheap, what is the new bottleneck?

Judgment, discovery and organizational design. Prototypes and customer previews are replacing detailed specs, but the one person AI product team has not materialized. AI replaces activities like user stories and backlog checks, not accountability, and a 1931 Procter & Gamble memo on one owner per brand still defines the PM job.

Are AI evals worth the effort for product teams?

The data says yes. Ramp went from 35% to 83% accuracy with evals and Cursor cut cost by 41%. Shopify's AI workflow builder is 2.2x faster and 68% cheaper, and nearly half of 25 shared PM openings now ask for experience writing evals.

Why does product operations fail in many companies?

Because it is treated as documentation instead of organizational design. AI cannot fix project based delivery, blurred ownership or unclear accountability. Many leaders cannot say how much R&D spend maps to each strategic priority, and overinvolved CEOs are named as a costly bottleneck.