Claude Design and Claude Science (where they fit)

Someone asks Claude for “a deck.” Someone else asks for “help with my single-cell analysis.” Both open the same family of products and land in very different tools. If you mash those jobs into ordinary chat, you get okay paragraphs and a lot of manual glue work. If you open the right specialty surface, you get visual drafts or research artifacts that match the job.

This is Part 4 of the Claude product map series. We already placed chat, Claude Code (software), and Claude Cowork (office and knowledge files). Now we place two more: Claude Design for visual work, and Claude Science for research workbenches. Neither replaces everyday chat. Neither is a peer-review stamp or a free pass on brand, methods, or safety. If Claude chat still feels new, keep Learn Claude from scratch nearby; this map assumes you can already use chat without panic.

What you’ll learn

Where Claude Design fits: decks, mockups, prototypes, brand systems, export and handoff
Where Claude Science fits: tools, packages, compute, auditable artifacts for scientific work
Who can open each surface (beta paths on Pro, Max, Team, Enterprise as of mid-2026)
A simple chooser: Design vs Science vs chat vs Code vs Cowork
Honest limits (not peer review, not automatic production software, not “pretty equals true”)
What to practice before you chase every beta logo

The map so far, plus two niches

Hold the core rule from Parts 2 and 3:

Chat = conversation and drafting with you in the loop on every paste.
Claude Code = software in repos, terminals, and IDEs.
Claude Cowork = non-coding knowledge work on folders and tools you choose.

Design and Science sit beside that core, not above it. They are specialized workbenches for visual production and scientific research workflows. Most office workers will touch Design occasionally. A smaller set of researchers and lab-adjacent roles will live in Science. Plenty of people need neither yet.

Side by side: Claude Design for visual drafts decks mockups and exports versus Claude Science for research tools comp…

Claude Design: visual work with Claude

Claude Design is Anthropic’s product for collaborating on polished visual work: designs, prototypes, slides, one-pagers, marketing collateral, and related artifacts. Anthropic Labs announced it as a research-preview style product powered by a strong vision-capable model path, aimed at people who need more than a text outline of a deck.

In plain English: you describe (or import) what you need, Claude builds a first visual version, you refine with comments, direct edits, and conversation, then you export or hand the work to another tool. Designers can explore more directions before lunch. Non-designers can get something shareable without opening five apps and fighting slide masters from 2014.

What Design is good for

Pitch decks and presentations: outline to on-brand slides, then export (for example PPTX or PDF paths the product supports).
Product mockups and wireframes: feature flows PMs can share, then refine with design partners or hand toward implementation.
Interactive prototypes: static ideas turned into something clickable for feedback without a full engineering sprint.
Marketing collateral: landing-page style layouts, campaign visuals, social-ready assets as a first pass.
Brand-consistent exploration: when a design system or brand kit is available, outputs can stay closer to real house style.

Brand systems and import paths

A useful Design feature family is brand and system awareness: onboarding that reads design files or codebases (where you grant access), multiple systems for teams that have more than one brand, and imports from prompts, documents (DOCX, PPTX, XLSX style inputs), images, or captures from existing sites. Exact import menus move; the product idea stays stable. Start from something real (brand PDF, existing deck, product URL) when you care about consistency.

Export and handoff

Design is only half useful if the work dies in a proprietary preview. Anthropic’s announcement path includes exports such as PDF, PPTX, standalone HTML, Canva-oriented flows, shareable internal URLs, and a handoff bundle toward Claude Code when a prototype should become product software. That handoff is the bridge between “looks right in a design tool” and “lives in the repo.” Design is not a replacement for Code’s engineering loop; it is often the sketch that Code implements.

Who can use Design (as of writing)

Claude Design has been offered in research preview / beta style access for Pro, Max, Team, and Enterprise subscribers, using subscription limits (with extra usage options where the plan supports them). Enterprise often has admin toggles (off by default in some org setups). Always re-check claude.com/product/design and admin settings the week you care. Free users should not plan a whole visual workflow around a paid beta surface.

What Design is not

Not automatic brand law. Pretty slides can still be off-message or off-legal.
Not a substitute for accessibility review, print production, or final art direction on high-stakes campaigns.
Not Claude Code. Shipping production UI still needs engineering judgment, tests, and review.
Not Cowork’s folder agent. If the job is “clean this directory of CSVs into a workbook,” Cowork is closer; Design is about visual composition.

Claude Science: a workbench for scientists

Claude Science is Anthropic’s AI workbench for scientific research work. The June 2026 announcement frames it as an app that integrates tools and packages researchers already use, produces auditable artifacts, and provides flexible access to computing resources. Think less “chat about biology” and more “environment where literature, analysis, figures, and compute jobs live with history you can inspect.”

Scientific work is fragmented by default: PubMed one tab, Jupyter another, R somewhere else, a cluster terminal, file formats that each want a bespoke pipeline. Science aims to pull stages of that work into one research environment where agents can help plan, run, and document steps, with reviewer-style checks on citations and calculations as part of the product story.

What Science is good for

Multi-step research workflows that mix literature, data, and analysis code.
Domain tooling with preconfigured skills and connectors (genomics, proteomics, structural biology, cheminformatics, and related areas called out in Anthropic’s materials).
Auditable outputs: figures and manuscripts paired with the code, environment notes, and history that produced them so validation and reproduction are possible later.
Compute management: plans that can run on a laptop, remote Linux, HPC login paths, or on-demand compute partners, with human approval before new resources are touched (as described in product materials).
Rich scientific views: structures, tracks, and other domain visuals rendered with traceability back to how they were made.

Who can use Science (as of writing)

Claude Science launched in beta for Pro, Max, Team, and Enterprise users, with early desktop paths emphasized for macOS and Linux. Team and Enterprise often need an admin to enable the product. There are research-lab oriented Team offerings and program notes that change; verify on claude.com/product/claude-science and Anthropic’s announcement page before you design a lab policy around it.

What Science is not

This is the line to tattoo on the lab whiteboard:

Not a replacement for peer review. Journals, committees, and independent replication still sit outside the product.
Not automatic truth. Auditable artifacts make checking easier; they do not make wrong methods right.
Not general office Cowork. Expense folders and status decks are Cowork territory. Science is research tooling and scientific compute.
Not a free pass on data governance. Sensitive human data, PHI, and proprietary sequences still need institutional policy, consent, and secure infrastructure choices.
Not “Claude Code with a lab coat” only. Overlap exists (code runs, environments), but Science’s product bet is domain skills, scientific artifacts, and research compute, not general software shipping.

Side-by-side: Design vs Science

	Claude Design	Claude Science
Primary job	Visual drafts: decks, mockups, prototypes, collateral	Research workbench: analysis, literature, figures with history
Typical user	PMs, marketers, founders, designers, AEs	Scientists, lab analysts, research software adjacent roles
Signature outputs	Slides, layouts, interactive prototypes, brand-consistent visuals	Analyses, scientific figures, manuscripts, compute job trails
Export / handoff	PPTX, PDF, HTML, Canva-style paths, handoff toward Claude Code	Auditable artifacts with code and environment context; compute on your infrastructure
Access (mid-2026)	Beta / research preview on Pro, Max, Team, Enterprise	Beta on Pro, Max, Team, Enterprise (macOS/Linux emphasis)
Hard limit	Pretty ≠ approved brand or accessible final art	Not peer review; not automatic scientific validity

Chooser: which surface for this job?

Chooser cards: need a visual or deck go to Design, need science tools and compute go to Science, need words and think…

Use this decision path before you open five product doors.

Need words and thinking only? Stay in chat. Outline the deck, critique the abstract, rewrite the email.
Need software shipped in a repo? Claude Code.
Need multi-step office files and folders? Claude Cowork.
Need a visual or deck that looks like design work? Claude Design.
Need scientific tools, packages, compute, and auditable research artifacts? Claude Science.

Borderline cases exist. A scientist preparing a conference poster might draft analysis in Science, then compose layout in Design, then paste talking points in chat. A founder might sketch UI in Design and implement in Code. A finance lead might stay in Cowork for the workbook and never touch Science. The map is a chooser, not a loyalty program.

Worked scenarios

Scenario A: founder pitch next week

You have a messy outline in a doc, three competitor screenshots, and a logo. Chat can help you sharpen the narrative. Design can turn that narrative into a first visual deck with export to PPTX for offline edits. Code is irrelevant unless you are also shipping a demo app. Cowork might help if the supporting research lives as a pile of PDFs in a folder you want summarized into a brief first. Sequence that works for many people: chat for story, Design for slides, human for investor-specific claims.

Scenario B: PM feature flow

You need a clickable path through a new settings page. Design is the right first tool for mockups and interactive prototypes. When engineering picks it up, a handoff toward Claude Code (or a normal ticket with attached frames) beats pasting screenshots into a chat thread and hoping. Chat still helps write the PRD. Cowork might organize research interviews if those files live in a folder.

Scenario C: lab analysis pipeline

You need to run a multi-step omics analysis, produce figures with methods history, and keep large data on lab machines. Science is the product aimed at that shape. Chat alone will thrash you through copy-paste of partial scripts. Code can help if you are building a production pipeline product for other teams, but day-to-day research iteration is what Science is marketing toward. Design might enter only when you need a polished poster layout, not when you need a differential expression table.

Scenario D: weekly marketing metrics

CSVs in a folder, need a four-slide update every Monday. Cowork is the closer fit for wrangling files into a recurring deliverable. Design can restyle a deck template when brand refresh matters. Science is the wrong aisle unless the “metrics” are actually scientific assays.

Plans, betas, and usage reality

Both Design and Science sit on paid paths (Pro and above in the mid-2026 announcements), often labeled beta or research preview. That means:

UI and export options will move. Prefer official pages over screenshots in random threads.
Usage still comes from your plan pool (plus extra usage where enabled). Heavy visual generation and long research agent runs are not free side dishes.
Team and Enterprise admins may gate features. If Design is missing, check org settings before you blame your browser.
Platform notes matter for Science (macOS/Linux emphasis early). Windows lab boxes may need a different path until support widens.

Do not upgrade five seats to Max solely because a Science beta logo is shiny. Upgrade when a named weekly job fails on Free or light Pro usage, or when your lab policy requires a paid seat with admin controls.

Common mistakes

Mistake	Why it hurts	Better move
Using Design as the system of record for product specs	Visuals drift; engineering needs tickets and code	Design for exploration; Code and docs for truth
Shipping Science figures without methods review	Pretty plots with wrong assumptions	Trace code, check inputs, independent validation where stakes are high
Confusing Science with general data analytics for business BI	Wrong tool family for dashboard ops	Cowork or chat for office metrics; Science for research workflows
Skipping brand review on Design exports	Off-brand decks in customer meetings	Human brand pass; lock systems in Design when available
Assuming beta features are permanent	Process depends on a preview that changes	Document workarounds; re-check product pages quarterly
Ignoring admin and data policy	Sensitive files in the wrong surface	Least privilege; institutional rules first

How this fits the AMS Claude path

Learn Claude from scratch: everyday chat, Projects, privacy, judgment.
Claude product map (here): chat vs agentic surfaces, Code, Cowork, Design, Science, then models and API.
Deep tutorials for Code and Cowork when you live in those surfaces daily. Design and Science get lighter map treatment until you have a real weekly job that needs them.

Most readers should leave this part able to ignore Science with confidence and open Design only when a visual deliverable is the bottleneck. Specialists in research environments can reverse that emphasis without guilt.

Practice this week

Write down one real job on your plate (one sentence, one success artifact).
Map it with the chooser: chat, Code, Cowork, Design, or Science.
If it is Design-shaped and you have Pro+, try one low-stakes visual (internal deck draft), export it, and do a human brand pass.
If it is Science-shaped and you are in a research role with beta access, try a non-sensitive public dataset exercise first and inspect the artifact history.
If neither fits, stay in chat or Cowork and skip the beta tourism.

Quick recap

Claude Design: visual work (decks, mockups, prototypes, collateral), brand systems, export (PPTX/PDF and related paths), handoff toward Claude Code.
Claude Science: research workbench with tools, packages, compute, and auditable artifacts; beta on Pro/Max/Team/Enterprise; not peer review.
Chat still wins for pure words. Code for software. Cowork for office folders. Design for visuals. Science for scientific research loops.
Betas move; re-check official pages. Usage still meters. Humans still own brand and scientific validity.

Next in this map: Part 5 is the models chooser (Haiku, Sonnet, Opus, and how to switch without hype). Part 6 covers API and Console only if you build apps that call Claude for other users.

Sources

Product details change. Re-check these the week you enable a feature.

Anthropic: Introducing Claude Design by Anthropic Labs (visual work, brand systems, export, handoff to Claude Code, Pro/Max/Team/Enterprise access)
Claude Design product page
Anthropic: Claude Science, an AI workbench for scientists (tools, packages, compute, auditable artifacts, beta access)
Claude Science product page
Claude Cowork product page (for contrast with office-file agent work)
Claude Code product page (for contrast with software work and Design handoff)
Claude pricing (confirm paid plan inclusions)
Analytics Made Simple: Learn Claude from scratch
Analytics Made Simple: Learn hub