1: Plan your workspace and taxonomy

This guide focuses on how to organize your workspace, how to control who can see and do what, and how to classify your data so it stays useful — and AI-ready — for the long haul. A quick reassurance before we start: how you structure your workspace is entirely up to you. There’s no right or wrong way to do this. It depends on how your team is organized and what you’re trying to achieve in Dovetail. And nothing here is permanent. What you decide today doesn’t have to be your structure forever — it’s easy to rename, move, and rearrange things as you learn what works. The goal is simply to get a sensible baseline in place early, because building the system as you go is far easier than retroactively organizing a wild west of data later. There are two levels to think about:

• The macro level — how you organize your workspace as a whole: folders, projects, roles, and access.
• The micro level — how you classify the actual data you bring in: fields and tags.

We’ll take them in that order.

​

Part 1: The macro level — organizing your workspace

​

Folders and projects: your two containers

Think of organization in Dovetail as a set of containers. A folder is your largest container. It holds projects and other folders, and it’s how you organize the workspace as a whole — just like in any other system. A project is a container of data: a set of video calls, survey responses, support conversations, or whatever you’re bringing in. A project is flexible, and it can be either open or closed:

• A closed project is the most familiar type. It has a clear beginning and end. For example: “We’re interviewing these 50 people, asking these questions, finding these answers, and writing a report.” When you’re done, you tie it up and move on.
• An open project has no clear start or finish — it’s a place to continuously collect data. For example, a Customer Success team might keep all customer calls in one ongoing project to spot trends over time and across team members.

Both types can live in the same workspace. Don’t overthink which is which — just know a project is flexible enough to support either pattern.

​

Deciding your folder structure: three levers

When you design your top-level folder structure, you’re usually balancing three considerations. Different teams weight them differently, so the question is which ones matter most to you.

1. Wayfinding. Make it easy for people to find what they’re looking for, and easy for you to know where new information should go.
2. Access control. If you ever need to restrict who can see or edit what — say, keeping one group’s work visible only to that group — it helps to have your parent folder structure follow those boundaries.
3. AI and querying. Dovetail’s AI tools work across folders. If your data is organized into folders and subfolders, you can query a whole folder at once and start surfacing answers across many projects.

A couple of common ways teams cut their top level: by research or data type, or by team or business unit. There’s no single correct answer. When you’re looking at how other companies structure things, the most useful guide usually isn’t the industry — it’s the underlying shape of the organization. For example, “one product with distinct teams” is a very different structure from “distinct business units serving different customers.” Find the description that sounds most like you and borrow from it.

​

Roles: who can do what

Dovetail has three core roles, plus an admin flag that can sit on top of any of them.

• Manager and Contributor are your “doers.” These are the people who can upload data, analyze it, and create new things in Dovetail. The only difference is that a manager has a few extra capabilities. There’s no limit on how many of each you can have, and no need to ration them.
• Viewers can look at everything and chat with the data, but they cannot edit anything or create new objects. The only thing a viewer can leave behind is a comment — like an “@mention, look at this” note. Comments don’t affect search or AI; they’re purely a human signal. You can have as many viewers as you like.
• Admin can be added to any of the above roles. This covers genuinely administrative tasks — deactivating and removing users, managing authentication settings, turning on beta features — not “god mode” over all content.

To see or change admins, go to More settings → Users. Admins show a small admin flag; to grant it to someone else, open the three-dot menu next to their name and choose Grant admin access.

​

Access controls: fine-tuning what people can do

Roles set the ceiling — the most a person can ever do anywhere in the workspace. A viewer is always at most a viewer; a contributor has at most a contributor’s abilities. Access controls let you turn that dial down for a specific folder. You’ll typically choose between:

• Full access (the default) — people act according to their role. Contributors can edit and invite others, viewers can view, and so on.
• Limited access — for example, contributors can still edit in this folder but can’t invite new people or change access. You can also drop a contributor down to view-only inside a particular folder — useful when you want one team to see another team’s work without being able to change it.
• No access — the folder becomes completely invisible except to people given explicit access. It won’t appear in search or in AI results. This is a full lockdown, usually reserved for sensitive PII or confidential work.

Between roles and access controls, you can filter permissions up and down to fit almost any situation.

​

User groups: access control at scale

If managing access person by person starts to get tedious, create a user group — a named set of people you can grant access to all at once. Instead of adding five individuals to a folder, you add the group. Groups also make team changes painless. Rather than remembering every folder someone belongs to, you add them to the right group and they inherit its permissions; move them to a different group and the permissions follow. It’s a tool in your toolkit, not a requirement — reach for it only if you need stricter or more frequent access management.

​

Part 2: The micro level — classifying your data

Once your structure is in place, the next job is classifying the data you bring into projects so it stays findable and useful. You have two tools for this: fields and tags. They do very different jobs.

​

Fields: the metadata of your data

A field is metadata — the information that’s always true about a piece of data, regardless of what’s discussed in it. If you were interviewing someone, their fields might be: works at Dovetail, lives in California, owns a dog. No matter what you talk about, those stay true. Fields are powerful for two reasons:

1. They capture unspoken context. Nobody opens a call by stating their company, region, industry, and account size out loud. But if you want to filter later to only your European customers, or only people in finance and banking, you need that context recorded as a field.
2. They’re what makes your data AI-ready. Fields give your data longevity. If your CEO asks, “What are the top use cases for our finance customers?”, an “Industry” field lets you instantly filter to those calls — in Dovetail Chat or in the workspace itself — and answer in minutes. Without it, that question becomes slow and unreliable, because no one remembers the details of a call from three months ago.

​

Workspace fields vs. project fields

Fields come in two flavors, and the distinction matters:

• Workspace fields (sometimes called global fields, marked with a globe icon) are shared across every project. They’re your team’s common language for filtering. Because they’re consistent everywhere, you can ask Dovetail Chat something like “using only data where subscription type equals student…” and it will filter across projects. Workspace fields are what connect the dots across everything you do.
• Project fields exist to help you get a specific job done. A personal status tracker (“analyzed / still to do”), or a field that splits this project’s data into cohorts you’re still developing — these are for your own workflow and don’t need to mean anything to anyone else.

In short: project fields help you do your work; workspace fields let you do cross-project analysis.

​

How to choose your fields: the rule of five

Fields are where to start if you do nothing else, because they’re how you’ll connect and filter data long term. Two ways to keep the list disciplined:

• The rule of five. If you could only have five fields on every single piece of data, what would they be? Six is fine. Twenty is not — no one fills out twenty fields, so they’ll go unused. Keep the scope tight.
• The boss’s-boss test. Ask: “What does my boss, or my boss’s boss, ask me?” Then check whether your fields and folders together can get you to that answer. The things you segment by in nearly every analysis — persona, industry, country, region — are strong candidates for workspace fields.

​

Tags: marking the moments that matter

Where fields describe the whole piece of data, tags classify moments within it. It’s a two-step idea:

1. Highlight a moment — a quote, a pain point, a feature request, a competitor mention. The highlight says “this is important, I might want it later,” and turns that moment into its own object you can pull into analysis or a report.
2. Tag the highlight to say why it matters. The highlight flags it; the tag qualifies it. For example: highlight the answer to “why did people choose us?”, then tag it so it also lands in the highlight reel for your final report.

​

Workspace tags vs. project tags

• Workspace tags tend to be high-level building blocks — the things you reuse across projects and combine in different ways (think “pain point,” “feature request”). The goal is to reduce duplication: you don’t need 25 versions of a “pain point” tag, you need one. Workspace tags also solve the blank-page problem, because you can attach them to a project template so they’re ready the moment you add data.
• Project tags are more flexible and personal — specific things you’re looking for in this project, or quotes you want to route into a particular report.

Again: project tags help you get your work done; workspace tags connect feedback across projects.

​

Fields vs. tags, and what AI actually uses

This is the key distinction to remember:

• Fields feed the AI. They provide the context and longevity that make your data queryable over time. Treat them as a requirement.
• Highlights and tags are human tools. They have no impact on Dovetail’s AI, search, or analysis. They’re for you — use them if and where they’re useful to your workflow.
• Comments are also purely human. They don’t affect AI or search either.

If you’re prioritizing, invest in fields first.

​

Part 3: Making it scale

​

Templates: never start from a blank page

Once you’ve settled on your workspace fields and tags, attach them to a project template. Then every new project starts pre-loaded with your classification system already in place. Your contributors don’t have to configure anything — they just start adding data. This is what lets you scale Dovetail across dozens or hundreds of people without each one reinventing the setup.

​

Leveraging AI and Dovetail Chat

This is where the discipline pays off. Because your data carries consistent workspace fields, you can ask questions in plain language and get instant, cited answers — filtered precisely to the cohort you care about. “What are the top use cases for our finance customers?” becomes a filter plus a prompt, answered across every relevant project at once. Good fields are what turn a pile of calls into an always-on, queryable intelligence layer.

​

Getting started: a simple plan

You don’t need to do everything at once. A sensible order:

​

Taxonomy Workshop Guide

If pulling this together feels daunting, that’s normal — and there are resources to help. The Taxonomy Workshop Guide was created to help teams craft their workspace from the ground up. While you don’t need to host an entire workshop, reviewing the Guiding questions and Real-world examples sections is a good place to start for inspiration and guidance. Learn more →

​

What’s next

The next guide focuses on how to actually build this system — folders, fields, tags, and your project template. To get the most out of it, we recommend having a plan before you start building: what your folder structure may be, the workspace fields you’re considering, and the tags you’re considering.

​

Where to next