Working with clients: transparency, contracts and pricing

0:00 Working with clients

Imagine this. You deliver a beautiful campaign. The client loves it. Two weeks later, their legal team asks a simple question. Which of these images were made with AI, and do we own them? You didn't mention AI anywhere in your proposal. You don't have a record. And suddenly a happy client is a nervous one. In this lecture, you'll learn how to work with clients on AI-assisted design: transparency practices, contract points to discuss with a lawyer, pricing that reflects value, and how to present AI-assisted work so clients see thinking, not a slot machine. By the end, you'll have an AI statement for your proposals and answers to the questions every client eventually asks.

0:50 Why it matters

Why does this matter? Because clients hire you for results and for trust. Undisclosed AI use can damage a relationship when it's discovered, especially if it causes a legal, quality or reputational problem. But transparent AI practice can be a selling point. You explore more options faster, and you handle risk professionally. Think of a mechanic who shows you the old part they replaced and explains why. You trust them more, not less. The same is true here. Clients don't mind AI. They mind surprises.

1:27 Transparency practices

Let's start with transparency practices. First, state your AI policy in proposals or on your website: which tools you may use, for what, and your safeguards. Second, ask about the client's policy before you start. Some organizations prohibit or restrict generative AI, especially in finance, healthcare and government. Third, flag AI use per deliverable with an AI usage summary in the delivery package. And fourth, be honest about limitations: ownership uncertainty, the need for real product photos, and the time that review takes. Here's the key idea. Transparency isn't a confession. It's a professional standard, like listing your stock image licenses.

2:11 Contract topics (with legal advice)

Now contracts. I'm not a lawyer, and you should have yours review your template once. But here are the topics to cover. Permission: the client agrees AI tools may be used, or sets restrictions. Tools and data: no uploading confidential material to tools without proper data controls. Ownership: you assign whatever rights you hold, and note that purely AI-generated elements may have limited copyright protection. Warranties: you warrant no knowing infringement and compliance with tool terms, and the client warrants rights in anything they supply. Likeness: no real person's likeness without documented consent. Disclosure: who is responsible for labeling at publication. Plus revisions and reasonable liability limits.

2:58 Pricing AI-assisted work

Let's talk about pricing, because AI changes production time, but not necessarily the value you deliver. If you bill hourly, AI speed cuts your income unless your rates rise. Value or project pricing captures the outcome, so efficiency benefits you. Package pricing, like twenty social assets a month, is predictable, as long as you define revision limits. Usage or licensing pricing reflects where and how long assets are used, which is common in advertising and illustration. And be clear about tool costs for big volumes. Avoid two traps. Don't underprice because the AI made it fast. Clients pay for concept, direction, quality, consistency and risk management. And don't pretend everything was hand-made when it wasn't.

3:48 Worked example 1: a yoga studio client

Worked example one, simple. A freelance designer in Toronto gets a new small-business client, a yoga studio. In her proposal, she includes a short AI statement. During kickoff, she asks if the studio has any AI policy. They don't, but the owner says, please don't make up photos of our classes. Perfect. That becomes a rule in the brief: class photos are real, AI is for backgrounds and illustrated quotes. At delivery, she includes a one-page AI usage summary listing which assets used AI. The owner later tells her that page is why she felt comfortable sharing the work with her investors.

4:32 Worked example 2: a compliance-minded retailer

Worked example two, a business scenario with illustrative details. An agency in Abu Dhabi is onboarding a retail client with a strict compliance team. At kickoff, they share their AI statement, ask for the client's policy, and learn that no customer data or unreleased products may go into public AI tools. So the agency agrees to use enterprise plans with data controls, and to shoot unreleased products in-house. The contract assigns rights, notes the limited protection of purely AI-generated elements, and makes the client responsible for final publication labels, with the agency recommending wording. They present three directions with reasoning, not a pile of forty generations. The client signs a twelve-month retainer, partly because the agency's AI practices matched their compliance needs.

5:25 Watch me do it: presenting AI-assisted work

Watch me do it. I'm presenting AI-assisted work to a client. I don't open with a grid of forty images. I open with the strategy recap: the audience, the message, the feeling we're after. Then I show three directions, each on one slide, with a sentence on how it serves the strategy. For direction B, I say it clearly: the background scenes are AI-generated around your real product photography, and every image went through our QA. Then I show the layouts in context, on a phone mockup and a story frame. At the end, I show one slide with what happens next: refinements, disclosure recommendations, and the delivery package. Calm, clear, and nobody feels tricked.

6:15 Client FAQ

Let's handle common client concerns. Is this cheaper because it's AI? AI speeds some production steps, but you're paying for concept, direction, quality control and risk management, and we price the outcome. Do we own it? We assign all the rights we hold, but purely AI-generated elements may have limited protection, which is why key brand assets are human-made. Will people know it's AI? Realistic AI content is labeled where platforms or laws require it, and we'll recommend wording. And can you use our CEO's face? Only with her written consent and clear disclosure. Short, honest answers like these build more trust than any portfolio piece.

7:01 Common mistakes

Common mistakes. Not mentioning AI at all until someone asks. Uploading a client's confidential material to a tool that trains on user data. Promising ownership you can't guarantee. Pricing by the hour and then quietly losing income as you get faster. Presenting dozens of raw generations and calling it choice. Letting the client assume every image is a real photo. And not asking about their AI policy, then discovering it after delivery. Every one of these is avoidable with a statement, a question, a summary page and a reviewed contract.

7:40 Recap and try this now

Let's recap. Clients don't mind AI. They mind surprises. State your AI policy, ask about theirs, include an AI usage summary with every delivery, and be honest about limits like ownership. Cover permission, data, ownership, warranties, likeness and disclosure in your contracts, with legal review. Price for value, not hours, and present directions with reasoning. Try this now. Write your own AI statement for proposals, one hundred to one hundred fifty words, using the template in the lesson. Then list the AI-related contract points you'll discuss with your lawyer or template provider this month.