Design Retainer for AI Startups: When It Fits

A design retainer is a monthly arrangement where a senior designer works inside your team every week on whatever matters most right now. For a seed-stage AI startup selling to enterprise, that usually means the product flow buyers see in demos, the sales demo itself, the website, and the decks, all moving together while the product is still changing.

You are not buying a list of screens. You are buying weekly judgment from someone who knows your buyer, your product and your roadmap, without running a hiring process.

I run a design retainer at Grauberg for AI startups selling software to enterprise. Before that I was the founding designer at Almetra AI for two years, from the first prototype to a $19M Series A. This post explains how a retainer works, when it fits, when it does not, and how to tell after 90 days whether it is paying off.

The short answer

A design retainer fits when three things are true:

1. You just raised and have a team of roughly 3 to 6 people with no designer.
2. The product brief changes every sprint because pilots keep teaching you something.
3. Enterprise buyers are judging you on trust: the demo, the product, the site and the deck have to tell one story.

If the brief is stable and you need one asset shipped, buy a project. If design already fills a full week every week after product-market fit, hire. The full comparison is in founding designer vs agency, and the timing question is in when to hire your first designer at a startup.

What a design retainer actually covers

At seed, a retainer is closer to having a founding designer than to buying an agency service. In a normal month with a seed AI team, my work spans:

• Product flows that block revenue. The screen a buyer sees in a live demo, onboarding for the first pilot users, and the states nobody designed yet: empty, loading, error, partial data, permissions.
• The sales demo. A clickable path that shows the outcome, including features that are not built yet, so sales stops improvising. I wrote up how in how to design a sales demo for enterprise SaaS.
• A lightweight design system. A small set of components engineers can ship from, so two features built in the same week feel like one product. See design system for startups.
• GTM surfaces. Website sections, the enterprise sales deck, one-pagers and pitch slides that match what the product really does.
• Research. Sitting in customer calls, watching people use the product, and turning that into the next change.

The point is not that one person does everything perfectly. The point is that one person holds the whole picture, so the website does not promise something the demo cannot show, and the demo does not show something the product will never do.

The assets a retainer installs over time are laid out in Design OS.

How a retainer works week to week

Here is the rhythm that works, because the format matters more than the hours:

Monday: priorities. A short call with the founder. What is blocking revenue or learning this week? A pilot starts Thursday, a buyer asked for a security overview, onboarding is losing people at step two. We pick one or two things.

Midweek: design and review in the open. Work happens in shared files and in your Slack, not in a black box. Engineers see flows before they are polished, so we catch “that is three weeks to build” early.

Customer contact. I join a sales or user call when it helps. At Almetra, some of the most important product decisions came from interviewing manufacturing teams and watching line leads on the floor, not from internal reviews.

End of week: something ships. A flow handed to engineering, a demo path the founder can use on Monday, a page that goes live. If nothing ships for two weeks in a row, the retainer is not working and we should talk about why.

This only works if the founder makes decisions quickly. A retainer amplifies a team that moves. It does not fix a team that cannot decide.

When a retainer fits, and when it does not

Good fit

• You raised recently and can fund senior design for at least three months.
• Enterprise pilots keep changing the product, so a fixed scope would be outdated within weeks.
• Engineers are making UX decisions by default because nobody owns them.
• The founder is the design bottleneck, designing at night instead of selling.
• Demos stall on the same objections: “Is this ready?”, “How would my team use this?”, “Can I trust this number?”

Bad fit

• You need one thing with a clear finish line. A website redesign or a pitch deck for a raise next month is a project. At Grauberg those run from €1.5k to €20k depending on the asset.
• There is no budget owner or 90-day goal. If nobody can say what should be different in three months, do not buy design capacity yet. Tighten your go-to-market first.
• You already have a strong in-house designer. Then you need overflow help on a specific project, not a second owner.
• Design is a full-time load and the roadmap is stable. That is the signal to hire.

I would rather say no to a retainer in the bad-fit list. Paying monthly for a designer who has nothing important to work on helps nobody.

What it costs, and how to think about it

I will not invent market surveys. Here is what I can say plainly:

• My Grauberg retainer is €6k per month, “your founding designer, without the hire”, until product-market fit, then a planned handoff to your first full-time design hire.
• One-off projects run from €1.5k to €20k: website redesign, dashboard, onboarding, design system, pitch or sales deck, sales demo.

The better question is what you compare it to. A full-time senior product designer means months of hiring, salary, equity and a bet that you know what the role should be. An agency means a fixed scope that may be outdated by the time it ships. A retainer means senior capacity this month, with the option to stop, scale or hire once the picture is clearer.

The wrong comparison is “retainer vs doing it ourselves for free.” Founder design time is not free. It is time not spent selling.

How to judge a retainer after 90 days

Do not judge a retainer by the number of screens. Judge it by what changed for the business. After three months, you should be able to say yes to most of these:

1. Demos are easier. Sales has a demo path that matches the buyer’s job and does not depend on the founder improvising.
2. Pilots start cleaner. New users get through onboarding without a call, and messy-data states do not break the first week.
3. Engineering ships faster. Engineers pull from shared components instead of inventing UI each sprint, and fewer flows get rebuilt.
4. The story is consistent. Website, deck and product describe the same thing in the same words.
5. The founder is out of Figma. Design questions have an owner and get answered in days, not weeks.
6. You know what to hire. You can describe what your first full-time designer should own and inherit.

If most of those are still no, something is off: the priorities, the decision speed, or the fit. Say it out loud and fix it, or stop.

What a retainer looked like at Almetra AI

At Almetra AI I started on the first prototype, before there was much product at all. Over two years the work moved through every surface a seed AI company needs:

• Interviews with manufacturing companies and visits to production lines to understand what line leads actually need.
• A redesign from dashboards that looked smart to plain-language insights a busy line lead could act on.
• A design system with the states engineers needed in production, so they could ship from it directly.
• The website and decks, kept consistent with the product while the company raised its $19M Series A.

None of that was one project with a fixed brief. It was weekly judgment on a product that kept learning. That is exactly the situation a retainer is built for. The product lessons are written up in designing AI software for manufacturing and enterprise-ready UI for AI startups.

How to start without overcommitting

If you are unsure, do not sign a long retainer on day one. A safer sequence:

1. Start with the asset that is blocking revenue. Often the sales demo, the primary demo screen, or a website that makes you look unserious to enterprise buyers.
2. Use that project as a working trial. You see how the designer thinks, how fast they move, and whether they understand your buyer.
3. Move to a retainer if the brief keeps growing. If every week brings a new design question, that is the signal.
4. Plan the handoff from the start. The goal of a founding designer retainer is to get you to product-market fit with coherent product and sales surfaces, then hand off to a full-time hire.

Before you pick anyone, use the scorecard in how to choose a design partner for an AI startup. Ask for work on products like yours, ask who actually does the design, and ask what they would change in your demo first.

A quick self-check

Answer yes or no:

1. Did we raise in the last six months, with no designer on the team?
2. Will the product change meaningfully in the next 30 days?
3. Are enterprise demos or pilots stalling on clarity or trust?
4. Is someone on the founding team designing at night?
5. Can we name what should be different in 90 days?

Four or five yeses: a design retainer is probably the right model. Two or three: start with one scoped project. One or zero: you probably do not need design capacity yet.