The User Research Strategist

9 research agents that run a whole stage of your study from one command, built from real studies I ran

For years I knew exactly how I wanted every stage of a study to go, and I was still the one doing all of it by hand. I'd open a blank chat, paste in the transcript, walk Claude through synthesis one instruction at a time and rebuild the report from scratch, study after study, even when I'd done the same setup a hundred times.

So I packaged how I run each stage into an agent. The UXR Claude Agent Suite is nine of them for Claude Cowork, from intake and scoping through synthesis, reporting and impact, and each one hands you back a finished first draft of that stage while the judgment stays with you.

Get the 9 agents for $99 Meet the 9 agents

One-time payment and instant download. Runs in Claude Cowork on any paid plan.

Why I built these

I know what a good interview guide looks like, why an insight needs a consequence on top of an observation, and how to structure a report so people act on it. I was still the one carrying every step, and the parts I dreaded most, like the synthesis that ate a whole afternoon and the report I kept putting off, were the parts I was redoing by hand every single time.

My first attempts at handing this to AI were absolute garbage. I'd get a flat, generic synthesis back and think, why didn't I just do this myself? It only started working once I built every step from previous work, real studies I ran, the insights they produced and the templates I use day to day.

Now I hand a stage over and review it with fresh eyes, instead of assembling it from nothing. A good agent takes the checklist, mundane parts of research off my plate so I have more time for the strategic stuff, and it still always gives me a first draft, never the final draft.

See the agents in action

Here's a walkthrough of the agents running real stages of a study inside Claude Cowork.

Running it by hand

You open a fresh chat for every study, rewrite your prompts, paste in the transcript and walk through synthesis one instruction at a time, carrying the work from one step to the next yourself. It gets you there, but the slow, repetitive assembling lands on you every time, usually at the end of a long week.

Handing it to an agent

You type /synthesize and the Synthesis agent clusters your transcripts, checks for contradictions, writes the insights and runs each one through the So-What check before it hands the set back. Your time goes to deciding whether the insights are right, which is the part of the work that needs you.

What's inside each agent

If agents are new to you, each one is a plugin with five parts that work together, and you install the whole thing in one step.

  • 1
    A lead skillThe part that runs the stage from your main chat. It starts each specialist in the right order, brings their questions back to you, and pulls the finished work together at the end.
  • 2
    Specialist sub-agentsEach stage is split into steps, and each step has its own specialist, like a method strategist and a protocol auditor for Scoping, so the deep work happens in a narrow lane instead of one chat trying to do everything.
  • 3
    SkillsThe know-how, one per task. Each skill is a single method written down, like how to write a screener or how to structure an insight, and the specialists reach for the right one at the right moment.
  • 4
    A slash commandThe shortcut that starts the stage. You type something like /synthesize and the whole stage kicks off, instead of you walking through it step by step.
  • 5
    Recommended connectorsLinks to tools you already use, like your survey platform or research repository, so the agent can pull in your real work and send the finished piece back out.

A skill

One job, done well. Ask for a screener or an insight and Claude follows the method and gives you that one piece.

An agent

A coordinator that runs multiple steps toward a goal, calling in the right specialist and skill at each one, and hands you a finished first draft of the whole stage.

How it works

1

Install the plugins in Cowork

Upload the nine plugins in Claude Cowork, which takes about five minutes the first time. There's no code and no API keys, and everything is ready to go once they're in.

2

Type a command or describe the work

Type something like /scope-study or /synthesize, or tell Claude in plain words what you're working on, and the right agent picks it up.

3

Review what comes back

The agent runs each step and hands the work back to you. You push back on anything that's off, and the judgment stays yours from start to finish.

What researchers have told me

“I've been playing around with your agents and skills. I really value your expertise in this space.”
Director of user researchRideshare company
“I've been playing a little bit with your sparring stuff. It's really good. I love it.”
ResearcherWorkshop attendee
“It totally makes sense, especially the distinction you shared of using the skills and these plugins as a first draft, so there's a level of interrogation, the sparring that is still happening.”
UX researcherWorkshop attendee

A lot of researchers would rather keep their AI use private, so I've shared roles instead of names.

The 9 agents

One for every stage of the work, in the order a project moves.

01The Planning agentIntake & Planning/intake-request

What it does

Turns a vague stakeholder request into a sharp plan tied to a real decision, and checks it before it reaches anyone.

What runs inside it

An intake gatherer, a question sharpener, an artifact builder and an intake auditor work the request in order, then it builds the brief, business case or roadmap the situation calls for and logs it against a named decision.

What you get back

An audited research brief or business case, tied to a real decision and checked against what the stakeholder meant, ready to share.

02The Scoping agentScoping/scope-study

What it does

Picks the method, builds the protocol and checks it for bias before it goes near a stakeholder or a recruiter.

What runs inside it

A method strategist, a qual guide builder, a structured method builder and a protocol auditor run in order, then it drafts the starting sample plan, previews the design for your stakeholder and logs the approved study.

What you get back

A field-ready, bias-checked protocol, a starting sample plan and a stakeholder sign-off.

03The Recruitment agentRecruitment/plan-recruitment

What it does

Works out who to recruit, how many and what to pay them, and checks the whole plan is ethical before a single invite goes out.

What runs inside it

A sample designer, a participant comms manager, an incentive planner and an ethics guardian run in order, then it logs the recruitment funnel against the sample plan's target.

What you get back

A screener, participant comms, incentive and consent form, cleared by an ethics check, with the funnel tracked from screened through completed.

04The Synthesis agentAnalysis & Synthesis/synthesize

What it does

Takes your transcripts and survey data and comes back with insights that hold up to a hard look, instead of a pile of quotes.

What runs inside it

A qual analyst and a quant analyst organize the raw data side by side, an affinity diagrammer clusters it and checks for contradictions, and an insight auditor writes and so-what checks every insight. It builds a persona or journey map when the evidence earns one.

What you get back

A contradiction-checked insight set ready for your repository, translated for whichever audience needs to read it.

05The Reporting agentReporting & Activation/build-report

What it does

Shapes insights into a readout that leads with the answer, plans how you'll present it, and tracks what happened after.

What runs inside it

A report writer, a shareout planner, an activation strategist and a readout auditor run in order, then it handles the newsletter cadence and logs what got adopted, deferred or rejected.

What you get back

A pyramid-structured report with a live-session plan and named owners, plus a record of what stakeholders did with it.

06The Impact agentInfluence & Impact/track-impact

What it does

Shows what your research changed in the language leadership funds, and keeps the practice itself measurable and well run.

What runs inside it

A repository architect, an impact metrics analyst, a process improver and a governance auditor run whichever piece you need, then it pulls the numbers and findings into an impact story for leadership.

What you get back

An ROI and impact story built from real numbers, plus a clear call on any democratization request before it scales.

07The Career agentCareer/assess-career

What it does

Places you against the next level, builds the materials that get you there and preps you for the interview itself.

What runs inside it

A self-assessor, a materials builder, a materials auditor and an interview-prep coach run in order, then it logs the next actions you agreed on against your target level.

What you get back

A straight level assessment, a resume and case study checked for activity dressed up as impact, and interview prep with a mock round.

08The Secondary Research agentSecondary & Competitive Research/competitive-scan

What it does

Answers what can be known from public sources before anyone commits two weeks to a primary study.

What runs inside it

A source scout and a competitive analyst gather public material, an evidence synthesizer rates it for credibility, and a claims auditor catches overstated or out-of-date claims, then it writes the stakeholder brief.

What you get back

A sourced, credibility-rated answer to a competitive or market question, and a clear call on whether a primary study is still needed.

09The Consulting agentConsulting Ops/scope-engagement

What it does

Runs the business side of independent research consulting, from a first client call through a signed engagement to a clean close.

What runs inside it

A discovery intake specialist and a proposal builder turn a client conversation into a scoped, priced proposal, an engagement auditor checks the price against the real scope, and a client comms manager handles updates and scope changes.

What you get back

A scoped proposal and SOW checked for vague language and underpricing, plus a running log of the engagement and a renewal conversation planned ahead.

Who it's for

It's a good fit if

  • You already use Claude Cowork, or you're ready to, and you want it running whole stages instead of one prompt at a time
  • You run the full research process and you're tired of assembling every step by hand
  • You're a team of one or a small team and could use a few teammates who know how good research is done
  • You want the method applied for you, and you still want to be the one making the calls

It's probably not for you if

  • You're on the free Claude plan since the agents run in Cowork on a paid plan
  • You're still learning what user research is and want a course on the basics first
  • You want something that does your thinking for you since everything it hands back is a first draft

Who made this

I'm Nikki Anderson. I've been a user researcher for 13 years, leading research at Zalando and Olapic and teaching it at Pratt, and I write The User Research Strategist. I've been using AI in my research for over three years, and a lot of that started with outputs I didn't trust.

Every workflow in these agents came out of real studies I ran, real readouts that changed a decision, and plenty I got wrong before I got it right, written down so you get the benefit of those years without repeating my mistakes.

13years in user research
9agents built
100+skills inside them
3+years using AI in research

UXR Claude Agent Suite

9 research agents for Claude Cowork

$99one-time payment

That works out to $11 per agent.

  • Nine plugins, one for every stage of a study
  • Specialist sub-agents and an auditor inside every agent
  • 100+ skills built from real studies and worked examples
  • Slash commands that start each stage
  • Recommended connectors for the tools you already use
  • A plain-language setup guide, about 5 minutes to install

You'll need Claude Cowork in the Claude desktop app, on Pro, Max, Team or Enterprise.

It's a digital download, so I don't offer refunds. If something isn't working for you, email me at nikki@userresearchacademy.com and we'll work something out.

Secure checkout through Stripe. If the button doesn't load, use this checkout link.

Questions people ask me before buying

Nine plugins for Claude Cowork. Each one bundles a lead skill that runs the stage, the specialist sub-agents for that stage, the skills they use, a slash command to start it and a list of recommended connectors. You upload them once and they're there every time you open Claude.

A skill is one job, like writing a screener or structuring an insight. An agent is a coordinator that runs multiple steps toward a goal, calling in the right specialist and the right skill at each one, so you hand off a whole stage instead of one piece of it.

An agent is only as good as the skills inside it, which is why every one of these is built from real studies I ran.

Cowork is the part of the Claude desktop app where plugins run, and the agents need it. It's on every paid plan, so Pro, Max, Team or Enterprise. If you already pay for Claude, you have everything you need.

Yes. There's no code, no API keys and nothing to configure, and if you've installed an app from a download, you already have every skill this asks of you. You don't need to learn the commands either since you can describe what you're working on in plain words and the right agent picks it up.

Some of them overlap, and the agents have extra skills I built just for running a whole stage, like the lead skills and the auditors. If you'd like both, my AI bundle has the agents, UXR Claude Skills and the Prompt Library together.

Yes. Install all nine, or only the stages you work in right now, and add the others whenever a project calls for them.

I treat everything an agent hands back as a first draft, never the final draft. Every agent has an auditor built in, like the protocol auditor in Scoping, the ethics guardian in Recruitment and the insight auditor in Synthesis, so the work gets checked before it reaches you. You still review it, push back on anything that's off, and make the calls.

Cowork only reaches your files, apps and tools when you give it permission, and the files the agents make save into a folder on your own computer. I recommend setting Cowork to ask before it acts, removing participant names and personal details before you hand over transcripts, and checking your company's policy on using AI with research data.

Yes. Every skill inside the plugins is a plain-text file you can open and read, so you can add your own templates and examples. If you'd like help setting this up across a whole team, that's what my Claude Training for Teams covers.

It's a digital download, so I don't offer refunds. If something isn't working for you, email me at nikki@userresearchacademy.com and we'll work something out.

Your $99 purchase is the nine agents as they are on the day you buy. If you'd like every new agent and update as I add them, that comes with my AI bundle, along with UXR Claude Skills and the Prompt Library.

Get the 9 agents

Nine agents for Claude Cowork, about five minutes to set up, and a one-time payment of $99.

Secure checkout through Stripe. If the button doesn't load, use this checkout link.