# From Job Titles to Archetypes: How AI Dissolves Classic Software Roles

> Author: Chris Jon Graf (AI Strategist & CEO)
> Updated: 2026-07-10
> URL: https://ai-outsourcing.ch/insights/from-job-titles-to-archetypes-how-ai-dissolves-classic-software-roles

## Summary

Engineering, Product, Design, and Data Science are merging into a new kind of role in software development. Boris Cherny, Creator of Claude Code at Anthropic, observes five archetypes replacing classic titles: Prototyper, Builder, Sweeper, Grower, Maintainer. The role follows the product phase, not the CV. For Swiss companies, this means: recruiting logic, vendor selection, and internal structures must be recalibrated—without jeopardizing governance and compliance.

## What Boris Cherny observes at Anthropic: Titles vanish, archetypes remain

Boris Cherny, Creator and Head of Claude Code at Anthropic, made a public observation in summer 2026 that irritates many C-level decision-makers: Classic functions—Engineering, Product, Design, Data Science—are dissolving on his team. At Anthropic, everyone carries the same title: 'Member of Technical Staff'. The PM codes, the designer codes, even Finance codes. Cherny himself ships 20 to 30 pull requests per day and runs five parallel Claude Code instances simultaneously.

What remains when job titles vanish? Cherny identifies five archetypes no longer tied to disciplines but to product phases and working modes: Prototyper (generates many ideas, 80% are discarded), Builder (turns prototypes into production-grade products in days), Sweeper (simplifies code, removes features, optimizes performance), Grower (iterates on product for Product-Market Fit), and Maintainer (ensures scaling, security, reliability). These roles exist in parallel, not sequentially—and they rotate depending on phase.

**70%** — productivity increase per engineer at Anthropic since Claude Code runs internally

## Why the merger is happening now: AI as functional layer

Rich Archbold formulated it precisely in February 2026: 'Once AI commoditises the functional layer, once the engineer, the PM, and the designer are all just people directing the same agents, the function stops being the thing that distinguishes them.' McKinsey quantifies the impact: AI reduces product development time by up to 50%. BCG sees engineering costs drop by 15 to 25%. Accenture reports up to 2x productivity gains with AI-driven design.

At Anthropic, 80 to 90% of some teams' code is written with Claude Code. The model reviews every pull request before a human sees it and catches most bugs in the first pass. Figma reports for 2026 that 91% of designers use AI tools weekly and save an average of 4 hours per week on routine tasks—that's 40% of their work time.

> **LinkedIn renames its APM program**
>
> LinkedIn transformed its Associate Product Manager Program in 2026 into a 'Product Builder Program' and trains hires simultaneously across Product, Design, and Engineering. The signal is clear: disciplinary boundaries are artifacts of the past.

## The five archetypes in detail: Who does what, when?

### Prototyper: The creative discarder

The Prototyper generates many ideas quickly—and discards most of them. At Anthropic, only 20% of prototypes survive. The small spinner icon you see during Claude Code's work went through 50 to 100 iterations until the right design survived. Agent teams produce hundreds of versions before shipment. In the pre-Product-Market-Fit phase, this archetype dominates: exploration over perfection.

### Builder: From prototype to production in days

The Builder translates a validated idea into production-grade infrastructure or product—and does it fast. Cowork, a complete product for non-engineers, came together at Anthropic in about 10 days. Deloitte confirms the pace: 20 to 30% faster time-to-market with AI-enabled design systems. Builders dominate the phase between prototype validation and first release.

### Sweeper: The disciplined simplifier

Anthropic unshipped 80% of what they build. Features with under 5% adoption are deliberately removed. The Sweeper cleans up: simplifies code, reduces complexity, optimizes performance. In growing and mature products, this archetype is indispensable—it prevents technical debt and keeps the system navigable.

### Grower: The iteration engine for Product-Market Fit

The Grower takes a built product and iterates it toward Product-Market Fit. They work closely with users, test hypotheses, A/B test features. Gartner expects that by 2027, 70% of new products will integrate AI in the design process—Growers use these tools to translate signals from data into product decisions. This archetype dominates the phase after launch, before scaling.

### Maintainer: The guardian of scale

When a product matures, it needs the Maintainer: they make the system secure, reliable, fast, and efficient under growing load. At Anthropic, Claude Code automatically reviews every PR—a Maintainer function that used to tie up weeks of manpower. In regulated markets like Switzerland, this archetype remains indispensable even under AI deployment: governance, compliance, and audit trails are not technical but organizational obligations.

## What this means concretely for Swiss companies

### Recruiting: From CV to portfolio-phase match

Classic job postings ('Senior Product Manager, 5+ years experience') lose relevance. The question is no longer: 'What title did you hold?' but: 'Which archetype role did you fill in which product phase?' An effective Prototyper is gold in the pre-PMF phase but may fail in a maintenance phase. Vice versa equally. This means: interviews must query archetype skills, references must provide phase context, onboarding must enable archetype rotation.

- Don't ask about job titles, ask about prototyping velocity, shipping frequency, refactoring discipline
- Evaluate portfolios not by tool lists but by phases: What did you build in which product phase?
- Structure teams by product phase, not by discipline: Pre-PMF needs more Prototypers/Builders, scale needs more Sweepers/Maintainers

### Vendor selection: Forward-deployed engineering reimagined

Our article 'Forward-Deployed Engineering: When AI Outsourcing Becomes the New Normal' described how Microsoft Frontier ($2.5B, 6000 engineers) and Amazon ($1B) deploy external engineering capacity into client infrastructure. With archetype logic, vendor selection changes: Are you booking a 'Senior Engineer' or a 'Builder for 8-week phase between prototype and MVP'? Vendors offering archetype-based team composition deliver higher velocity—provided they can read the product phase.

> **Archetype questions for vendor conversations**
>
> Ask your outsourcing partner: 'Which archetype roles does your team cover? Can you compose a team differently for pre-PMF than for scale? How do you rotate roles during phase transitions?' If the answer is 'We have Senior Engineers and Product Managers', the vendor is mentally still in the old model.

### Governance and compliance: Archetypes ≠ anarchy

A common misconception: 'When titles vanish, responsibility vanishes.' The opposite is true. Art. 716a OR demands non-transferable board duties—regardless of whether your CTO is called 'CTO' or 'Member of Technical Staff'. The EU AI Act (August 2026) defines deployer obligations for high-risk systems—our article 'EU AI Act August 2026: The Deployer Obligation Catalog for Swiss Companies' shows the details. FINMA Guidance 08/2024 and revised DPA demand documented responsibilities.

Archetype organization does not mean title abolition on the org chart but functional flexibility in daily work. Formally, roles and signatures remain. Operationally, people rotate through archetypes depending on product phase. This requires: clear RACI matrices (who in archetype X is responsible for decision Y?), documented archetype rotation (who was active in which role when?), and audit trails that hang not on job titles but on responsibilities.

## Archetype mapping: Who needs what in which phase?

- Pre-PMF (exploration): 50% Prototyper, 30% Builder, 20% Sweeper—high iteration, much discarding, fast pivots
- Growing PMF (traction): 30% Builder, 25% Grower, 25% Sweeper, 20% Maintainer—stabilization without rigidity
- Mature/Scale (optimization): 40% Maintainer, 30% Sweeper, 20% Grower, 10% Builder—efficiency, security, continuous improvement

This mapping is not static. A successful product goes through multiple cycles: After scale often comes a new prototyping cycle for features or spin-offs. People can and should switch between archetypes—but not all people are equally strong in all archetypes. Accepting this is the first step to more efficient team composition.

## The three biggest mistakes in transition

1. Misusing archetypes as new rigid titles: 'From now on you're our Sweeper' is just as wrong as 'You're our Senior Engineer'. Archetypes are roles you assume, not identities you carry.
2. Prematurely dismantling governance structures: Swiss companies are subject to Art. 716a OR, FINMA requirements, EU AI Act deployer obligations. Archetype flexibility must not undermine these.
3. Treating all software roles equally: These archetypes apply to software-PRODUCT development (Engineering, Product, Design, Data Science). Finance, HR, Operations, Legal follow different logics. Our article 'AI and HR Strategy: Why AI-Driven Headcount Reduction Delivers No ROI' shows why blind layoffs after AI introduction fail.

## What you can do concretely now

Start with a pilot team in a clearly defined product phase. If you have a new product in pre-PMF phase, assemble a small team of two Prototypers, one Builder, and one Sweeper—without rigid titles, with clear archetype roles and weekly retrospectives. Measure velocity (how many validated prototypes per week?), waste (how much is consciously discarded?), and time-to-production (how fast does a validated prototype become a running feature?).

Document what works and what doesn't. After 8 to 12 weeks you have enough data to roll out archetype logic to more teams—or adjust. Important: Governance structures remain but are calibrated to archetype rotation. A 'Member of Technical Staff' can be documented in the audit trail as 'Maintainer in weeks 12–18, Grower in weeks 19–24'—that's sufficient for revised DPA, FINMA, and EU AI Act.

> **Caution with regulated systems**
>
> If your product falls under EU AI Act high-risk categories (see our article on the deployer obligation catalog), you must document who made which safety-relevant decision when. Archetype rotation is permitted but not anonymized: persons, timeframes, and responsibilities must remain traceable.

## The AI Design Market grows at 34% CAGR—archetypes are the structure behind it

The AI Design Market grew from $741M (2024) to a projected $13.94B (2034)—an annual growth rate (CAGR) of 34.11%. This explosion is not just tool-driven but structural: Teams that use AI tools effectively are those that flexibly map roles to product phases. Anthropic, LinkedIn, and a growing number of European scale-ups show: Archetypes are not Silicon Valley gimmickry but competitive advantage.

For Swiss decision-makers, this means: Those who ignore archetype logic lose velocity against competitors who use it. Those who blindly copy it without governance calibration risk compliance violations. The middle path—archetype flexibility with documented responsibility—is demanding but rewarding. The question is not whether, but how quickly you begin.

## FAQ

### What are the five archetypes replacing classic software roles?

Boris Cherny identifies Prototyper (generates many ideas, discards 80%), Builder (makes prototypes production-ready in days), Sweeper (simplifies code, removes features), Grower (iterates for Product-Market Fit), and Maintainer (ensures scaling and reliability). These roles follow product phases, not CV titles.

### Does this mean job titles like 'Product Manager' or 'Engineer' completely disappear?

Not necessarily. Formally, titles often remain (for governance, contracts, org charts). Operationally, however, people work in archetype roles that change depending on product phase. At Anthropic, everyone carries 'Member of Technical Staff' but works in different archetypes as needed.

### How does this affect recruiting in Swiss companies?

Recruiting must shift from 'What title did you hold?' to 'Which archetype role did you fill in which product phase?' Interviews should query prototyping velocity, shipping frequency, and refactoring discipline. Portfolios should be evaluated by product phases, not tool lists.

### Does archetype flexibility endanger governance and compliance in Switzerland?

No, when properly implemented. Art. 716a OR, FINMA Guidance, revised DPA, and EU AI Act demand documented responsibilities—independent of title. Archetype rotation must be captured in RACI matrices and audit trails: who was responsible in which role when? Flexibility yes, anonymity no.

### Does archetype logic apply to all company roles?

No. These archetypes apply to software-PRODUCT development (Engineering, Product, Design, Data Science). Finance, HR, Operations, Legal follow different logics. Not every role in the company merges through AI—only those working directly on the software product.

### How do I start with archetype teams in my company?

Begin with a pilot team in a clearly defined product phase (e.g., pre-PMF). Assemble a small team without rigid titles (e.g., two Prototypers, one Builder, one Sweeper), measure velocity and time-to-production over 8–12 weeks. Document learnings and then roll out—with existing governance structures calibrated to archetype rotation.

## Sources

- [Boris Cherny (Creator of Claude Code) On How His Career Grew](https://www.developing.dev/p/boris-cherny-creator-of-claude-code)
- [Anthropic's Claude Code Team Has 5 Roles and Zero Job Titles](https://aakashgupta.medium.com/anthropics-claude-code-team-has-5-roles-and-zero-job-titles-bf4860a389fc)
- [The 5 job archetypes of the future, according to Claude Code's creator](https://benzatine.com/news-room/the-future-of-work-five-emerging-job-archetypes-in-the-age-of-ai)
- [How AI Tools Are Transforming Product Design in 2026](https://caddcentre.com/blog/how-ai-tools-are-transforming-product-design-in-2026/)
- [Future of Product Management Beyond 2026](https://blog.productmanagementsociety.com/future-of-product-management-beyond-2026/)
- [The Rise of the Design Engineer: How AI is Collapsing the Digital Creation Stack](https://ourculturemag.com/2026/07/07/the-rise-of-the-design-engineer-how-ai-is-collapsing-the-digital-creation-stack/)
- [The AI-Powered Product Engineer: A New Engineering Archetype](https://medium.com/@rich_archbold/the-ai-powered-product-engineer-a-new-engineering-archetype-3119ed9716e4)
- [Designing the Future with AI: Trends & Insights for 2026](https://www.parallelhq.com/blog/future-of-design-with-ai)
- [Anthropic's Boris Cherny says there are days he manages tens of thousands of AI agents](https://fortune.com/2026/06/08/anthropics-boris-cherny-creator-of-claude-code-says-there-are-days-he-manages-tens-of-thousands-of-ai-agents-at-once/)