Before a design partner starts, have one owner who can decide, access to the product and its data, what you know, the goal and the practical set-up. Before a design agency starts work on your product, have five things ready: one person who can make decisions, working access to the product and its data, what you already
Before a design agency starts work on your product, have five things ready: one person who can make decisions, working access to the product and its data, what you already know about users and the business, the outcome the work must move, and the practical set-up of tools, security and meeting rhythm. Everything else can be gathered with your partner in the first two weeks. A partner who starts designing before those five are in place is guessing, and you pay for the guesses twice: once when they are made, and again when they are unpicked.
The checklist below is the one we work to with every client. It is your side of the work. The designer's side, what they must learn about your business before designing anything, is a separate guide.
The checklist at a glance
| Area | Ready before day one | Gathered with your partner in the first two weeks |
|---|---|---|
| People | One owner with decision rights; named reviewers; a deputy for when the owner is away | Introductions to support, sales, engineering and, where useful, customers |
| Access | Product logins for each user role, analytics, design files, staging or code, support tickets | Any access that needs a security review, requested on day one |
| Knowledge | Existing research, the PRD or brief, known complaints, the business model in a paragraph, and a first list of what you believe but have not checked | That list, tested and extended with your partner |
| Goals | The outcome the work must move, who agreed it, and how the change will be measured | The baseline, if you do not have one yet |
| Practicalities | NDA, security questionnaire, rules on AI tools and data, a shared channel, the meeting rhythm | Review dates booked through the first milestone |
If the left column is complete, the engagement can start on evidence. If two or more rows are empty, read the last section first.
People: one owner who can decide
The single most useful thing a client can bring is one internal owner with the authority to decide. Not a committee, and not a project coordinator who passes questions upward. Atlassian's DACI play gives the role a name: the Approver, the person who "has the final say" on a decision. A design engagement needs the same thing on the client side, because every week of work ends in choices between directions, and a choice that waits for a meeting of five people stalls the work behind it.
Have ready:
- The owner. One name, with the right to approve a direction without escalating. In a startup this is a founder or the head of product; in a larger company, the product manager who owns the area.
- A deputy. Someone who can decide when the owner is travelling, so reviews do not slip by a week.
- Named reviewers. A short list of the people whose comments count, each told what they are reviewing: engineering for feasibility, support for real user problems, legal or compliance for the rules. Comments from everyone else go through the owner.
- Who owns which conversation. Decide in advance who invites customers to research, who answers technical questions, and who speaks to leadership about the work. Unclear ownership of these conversations is how a partner ends up waiting for an introduction that nobody thought was theirs to make.
We keep the product direction with the client: we bring options, the trade-offs between them and a clear recommendation, and the owner decides. In ongoing work with us, the owner's counterpart is a senior designer who leads the work.
Access: the product as it really is
A partner cannot design well for a product they have only seen in a demo. Give access to the product as your users experience it, and to the evidence of how they use it.
- The product itself. A login for each user role (an administrator, an everyday user, a new account), on an environment with realistic data. An empty account hides the screens that matter most: the busy table, the error state, the account with three years of history.
- Analytics. Read-only access to your product analytics and, where relevant, session recordings. Even an imperfect set-up shows where people drop out of a flow.
- Support tickets and sales notes. The raw text of what customers complain about and ask for. This is the cheapest user research you already own.
- Design files and the design system. The current files, the component library and any brand rules, with permission to comment.
- Staging or code. Read access to a staging environment, and to the codebase where the partner will work in code or must design within real technical limits.
Ask for anything that needs a security review on day one, not when the partner first needs it. An access request left in an approval queue can turn a partner's first week into a week of reading slides.
Knowledge: what you already know, including what you doubt
You do not need finished research before a partner starts. Finding out who the users are, what they are trying to do and where they struggle is the job of the early phase; the GOV.UK Service Manual describes discovery as understanding the problem before committing to build, and says plainly that you should not start building during it. What you need is to hand over what already exists, so the partner starts from your evidence rather than from zero.
Have ready:
- Past research. Interview notes, survey results, usability tests, even old or inconclusive ones. GOV.UK's guidance on user research in discovery lists reviewing previous research and examining existing data, such as analytics and support logs, among the activities for that phase.
- The business model in a paragraph. Who pays, who uses, and which actions in the product the business depends on. A redesign that improves the experience but weakens the moment a customer upgrades is not an improvement.
- Constraints. Technical limits, regulation, contracts with large customers, platform rules, the release you cannot move.
- The PRD or brief, labelled. A product requirements document helps when it describes the problem, the users and the constraints. Where it already specifies screens, mark which parts are decided and which are open. An unlabelled PRD full of solutions turns design into decoration of choices nobody tested.
- A prototype, if you have one. If your team has built something with AI tools, share it with a note on what it was meant to test. It is a strong input, and a poor specification.
- What you believe but have not checked. The most valuable page a client can write is a short list of assumptions the team holds without evidence. It tells the partner where to look first.
Goals: the outcome the work must move
"Redesign the dashboard" is a task. "Cut the time a finance team takes to close the month" is an outcome, and only an outcome lets a designer choose between two good directions. Before day one, agree internally what the work must change, and make sure the people who will judge the result agree with it. A goal that the founders, sales and engineering would each describe differently is not yet a goal.
Have ready:
- The outcome, in a sentence, with who agreed it.
- How it will be measured: the metric, the event in analytics, or the support category that should shrink. If there is no baseline yet, say so; capturing one belongs in the first two weeks.
- What is out of scope, so the partner does not spend the first fortnight on a problem you have already ruled out.
GOV.UK makes the same demand of its own teams: by the end of discovery they should know how they will measure success. A partner can help define the measure. They cannot invent the outcome for you.
Practicalities: tools, security and rhythm
These are small items that cost weeks when they are left until they are needed.
- The NDA and your security or supplier questionnaire, sent with or before the proposal, not after the start date.
- Your rules on AI tools and client data. If your organisation limits which AI tools may see your product, data or code, say so before any file moves. We agree how such rules apply before work starts, and our page on how we use AI on client projects sets out our side.
- Tools. Invitations to your design tool, project tracker and shared channel, so the work happens where your team already looks.
- The meeting rhythm. A standing review with the owner, a shorter check-in between, and the hours you share across time zones. Book the first reviews before the kickoff.
The first two weeks
Whatever the size of the engagement, the first two weeks decide how the rest goes. Atlassian's project kickoff play states the reason directly: "Unclear requirements lead to costly rework." Here is what those two weeks should contain, and what your team does in each part.
| Part | What the partner does | What your team does |
|---|---|---|
| Kickoff | Confirms the outcome, the owner, the reviewers, the constraints and the access, and checks each login works | The owner attends and decides; anyone with a blocked access request chases it the same day |
| Immersion | Uses the product in every role, reads analytics and support tickets, reviews past research, and talks to support, sales and engineering, and to customers where agreed | Makes the introductions, answers questions quickly, and points out the quirks that have reasons behind them |
| Playback | Writes back what they understood: the problem as they now see it, the assumptions worth testing, the risks, the open questions and the plan for the first decisions | The owner corrects the understanding and makes the first decisions on direction |
The playback is the test of the first two weeks. If your partner can describe your users, your constraints and your goal back to you better than your own onboarding document does, the engagement is on evidence. If the first thing you see is polished screens, ask what they are based on.
What a good partner brings and asks for
Readiness is not only the client's job. The result of a design engagement belongs to both sides: the client holds the decisions and the access, and the partner holds the quality of the questions, the evidence and the design. A good partner should not wait politely for inputs that never arrive.
At Tcules, we ask for the items on this checklist before work starts and chase what is missing, because each missing item turns into a question the design cannot answer yet. Whichever partner you work with, expect them to:
- Ask for access and people early, by name, rather than wait to be offered them.
- Show questions before solutions, and say which assumptions they are making.
- Write the playback down, so your team can correct it in one reading.
- Keep the work visible between reviews, with early directions you can still change.
- Say when an input is missing and what it will cost the work, instead of filling the gap with a guess.
If you are still choosing who to work with, our guide to choosing a product design agency covers the step before this one. And if the engagement will run for a long time, keeping what the partner learns inside your own team matters as much as the start; we cover that in how to outsource UX design without losing product knowledge.
Not ready yet?
Some teams read this list and find two or more rows empty. That is useful to know before a contract, not after. The signs are clear: nobody can approve a direction without a meeting of several people, access needs approvals nobody has started, or the founders and the product lead would describe the goal differently.
You have three sensible options.
- Close the gap first. Naming an owner, agreeing the outcome and raising access requests are internal work. None of it needs a partner, and none of it should be paid for as design time.
- Start with one bounded problem. A pilot on a single workflow or screen shows, on something small, where decisions stall and where access is slow, before those gaps can slow a larger project. We price a pilot up front, as one fixed figure agreed before the work begins: you get real work on that problem, and the gaps it shows come with it.
- Check what you already have. If the open question is whether a prototype built with AI tools is fit to build on, a prototype readiness review answers that before anyone designs on top of it.
How we shape engagements around where a team is, from a first pilot to ongoing work, is on our How we work page.
About the authors

Rakesh Chitroda
Co-founder
LinkedInMore from RakeshRakesh Chitroda is a co-founder of Tcules and writes for teams choosing an outside partner for product work. Rakesh's guides cover how to choose a product design agency, what a product design agency does, when to hire a design system consultant, how to choose a UX audit agency or a front-end development agency, and how to compare software modernisation companies. Other pieces deal with cost and working arrangements: what an MVP costs to design and build, hiring a product designer for a SaaS product, and outsourcing UX design without losing product knowledge.
Anshu Jain
Business Development Specialist
More from AnshuAnshu Jain is a Business Development Specialist at Tcules, the digital product design studio in Ahmedabad. At Tcules, Anshu works on how SaaS teams choose a design partner, and co-writes about questions such as what a product design agency should do, how to hire a product designer, and what to ask a UX audit agency before signing. The pieces below speak for the studio, and this page grows as Anshu writes.