AI for freelancers: proposals, client emails, and invoicing

The thing nobody tells you about freelancing is that the actual work (the design, the writing, the code, the photography) is maybe 50% of the job. The other 50% is proposals, invoicing, client emails, scope negotiations, scheduling, bookkeeping, and all the other business tasks you never signed up for.

Most freelancers I know got into freelancing because they’re great at their craft. Then they discovered they also need to be a salesperson, an accountant, a project manager, and a customer service rep. The prompts below are the ones I reach for to claw back the hours those roles cost.

Writing proposals that win work

A good proposal takes time to write. You need to understand the client’s problem, explain your approach, justify your pricing, and sound professional without sounding generic. Most freelancers either spend too long on proposals or send rushed ones that don’t win work.

Here’s a prompt for your next proposal:

“I’m a freelance web developer writing a proposal for a local restaurant that wants a new website. They currently have a basic WordPress site that’s slow and not mobile-friendly. They want online ordering, a menu that’s easy to update, and reservation integration. My approach would be a custom site on Next.js with Square for ordering. The project would take about 6 weeks and I’d charge $8,500. Write a proposal that explains my approach in plain language (the client isn’t technical), justifies the investment, and includes a timeline broken into phases.”

What you’ll get is a solid first draft that you can personalize. Add specific references to their current site, mention something you noticed about their business, and adjust the language to match how you actually talk. The AI handles the structure and phrasing. You add the personality and expertise.

For the follow-up when a potential client goes quiet:

“I sent a web development proposal to a restaurant owner 8 days ago and haven’t heard back. Write a short, friendly follow-up email. I don’t want to seem pushy, but I do want to know if they’re still interested so I can plan my schedule. Keep it under 80 words.”

Short, professional, not desperate. That’s hard to write when you’re anxious about whether you’ll land the project. AI removes the emotional weight from these communications.

Client communication for difficult situations

The easiest client emails are the ones where everything is going well. The hard ones are scope creep conversations, timeline delays, and rate negotiations. These are exactly the situations where having a drafting partner helps most.

When a client asks for work that wasn’t in the original agreement:

“My client just emailed asking me to add a blog section to their website. This wasn’t in our original scope. The project is a fixed-price contract for $8,500. Adding a blog would probably take another 15-20 hours. Write a reply that: (1) thanks them for the idea, (2) explains this is outside our original scope, (3) offers to add it as a separate phase with a cost estimate, and (4) keeps the tone positive and collaborative. I don’t want to seem difficult.”

This kind of email is agonizing to write from scratch because you’re trying to balance firmness with friendliness. Having a draft to work from makes it much easier.

For delivering bad news (project delays, unexpected issues):

“I’m behind schedule on a logo design project because the client took 2 weeks to send feedback on the first round of concepts (they originally agreed to 3-day turnaround on feedback). This pushed my timeline back, and I have another project starting next week. Write an email explaining the delay, gently noting that the feedback timeline affected the schedule, and proposing a new delivery date of March 28 instead of March 21. Be direct but diplomatic.”

For regular status updates (which many freelancers neglect, losing client trust in the process):

“Write a brief weekly status update email for my client. This week I completed the homepage design, started on the product pages, and resolved the mobile navigation issue we discussed. Next week I’ll finish product pages and start on the checkout flow. No blockers. Keep it professional but not stiff. Under 100 words.”

Sending consistent status updates is one of the simplest ways to keep clients happy, and AI makes it take about two minutes per client. For more on using AI for professional writing, see Writing and communication.

Invoicing and expense tracking

Freelancers who don’t stay on top of invoicing don’t get paid on time. It’s that simple. AI helps with the parts of invoicing that tend to slide.

For writing clear invoice descriptions instead of vague line items:

“Write professional invoice line items for the following work I did this month for a marketing agency client: I designed 3 social media ad sets (12 individual ads), revised their brand guidelines document, created email header templates for their 4 product lines, and attended 2 strategy meetings. Break these into separate line items with clear descriptions.”

For chasing late payments (a task every freelancer dreads):

“A client owes me $3,200 for work delivered 3 weeks ago. Payment terms were net-15 and it’s now 6 days overdue. This is a good client I want to keep working with. Write a payment reminder email that’s polite but clear. Mention the invoice number is #INV-047 and include the amount. Don’t be aggressive, but don’t be so soft that it’s easy to ignore.”

If you need a second reminder:

“Same client, it’s now 2 weeks past due. They haven’t responded to my first reminder. Write a firmer follow-up. Still professional, but this time mention that I may need to pause upcoming work until the outstanding balance is resolved.”

For quarterly expense review:

“Here are my freelance business expenses from Q1. Categorize each one as: software/tools, office supplies, marketing, professional development, travel, meals (client), or other. Then calculate the total for each category. [paste expenses]”

This gives you a quick snapshot for tax planning without needing to log into your accounting software for a rough estimate.

Time management and scheduling

When you work for yourself, nobody manages your time. That’s both the appeal and the challenge. AI can help you plan your days more realistically than the optimistic to-do list most freelancers write each morning.

At the start of each week:

“Here are my active freelance projects and deadlines: (1) Restaurant website, due April 15, about 30 hours of work remaining. (2) Logo design for a startup, due April 8, about 10 hours remaining. (3) Monthly social media templates for a retail client, due April 1, about 5 hours remaining. I work about 6 billable hours per day, Monday through Friday. Today is March 24. Build me a weekly schedule that prioritizes by deadline and accounts for the fact that I’m usually less productive on Mondays and most creative in the mornings.”

You’ll get a realistic daily breakdown. The AI accounts for deadlines, energy levels, and available hours in a way that’s hard to do in your head when you’re juggling multiple projects.

For handling the “should I take this project?” decision:

“A new client wants me to design a brand identity package. It would pay $4,000 and take about 40 hours. My current utilization for the next 6 weeks is about 80%. If I take this project, I’d be at roughly 110% capacity. I could push my social media client’s next deliverable by a week (they’re usually flexible). Should I take the project? What are the risks of being overbooked, and what would I need to negotiate with the new or existing client to make it work?”

AI won’t make the decision for you, but it’ll lay out the trade-offs clearly so you can decide with your eyes open. One caveat worth being honest about: ask AI to evaluate “should I take this project?” and it will tend toward yes-with-asterisks, because the prompt itself frames the project as an opportunity. If you’re already at 110% capacity, the right answer is usually no, and you don’t need an LLM’s permission to say it.

Portfolio and case study writing

Your portfolio is how you get the next client. But writing about your own work is surprisingly hard. You know what you did, but articulating the business impact and your process in a compelling way is a different skill.

“Help me write a portfolio case study for a website redesign project. Here are the details: Client was a regional insurance agency with a 10-year-old WordPress site. They were getting about 200 monthly visitors and zero online quote requests. I redesigned the site on Webflow, focused on mobile experience and clear calls to action. After launch, their traffic went to 800 monthly visitors and they’re getting 15-20 quote requests per month. Write this as a short case study (250-300 words) with sections for: the challenge, my approach, and the results. Make it sound confident but not braggy.”

For shorter portfolio descriptions:

“Write a 2-sentence project description for my portfolio. The project: I designed the packaging and brand identity for a craft hot sauce company called Ember & Vine. They went from farmers market sales only to placement in 45 grocery stores within 6 months of the rebrand.”

Having well-written case studies is the difference between a portfolio that gets you contacted and one that gets skimmed and forgotten. AI helps you write them without spending an entire weekend on it.

Setting up systems, not one-off prompts

The real power of AI for freelancers isn’t in individual prompts. It’s in building a system where AI handles the recurring tasks that eat your time.

Create a folder (a notes file is fine) of your go-to prompts for the work that recurs every week. Mine looks roughly like this:

• A weekly status update template, with one variation per active client (their tone, their priorities)
• Invoice-line-item phrasing for the project types I do most often (a redesign, a logo package, a one-off consulting call)
• A proposal skeleton per service offering, with the cost and timeline phrasing I keep refining
• Three follow-up-email variants: quiet prospect, late payment, post-project check-in
• A weekly scheduling prompt that takes my open projects + capacity and produces a realistic plan

Start with whichever task you dread most. For most freelancers, that’s either proposals or difficult client emails. Build the habit with one task, then expand. If you’re managing other aspects of a small business, AI for small business owners covers marketing, bookkeeping, and broader customer communication. For specific support around the meeting half of client work (pre-call research, agendas, follow-up emails), see AI for meeting prep. And for the broader productivity layer, organization and productivity goes deeper on the daily-rhythm side of working solo.

One more thing worth saying out loud: AI-written client emails sometimes sound exactly like AI-written client emails. The phrasing gets too rounded, the warmth too symmetrical, the closing too neutral. Sophisticated clients notice. The fix isn’t to stop using AI for first drafts, it’s to actually edit them. Run the AI’s draft through your own voice. Cut the corporate softening. Add a half-sentence aside that only you would write.

The point of all this

Freelancing is supposed to be the work you got into it for: the design, the code, the writing, the photography. The admin half is the reason so many talented freelancers burn out and go back to staff jobs. AI doesn’t make the admin half disappear. It does make the dread disappear, which is most of what was wrong with it. That’s the trade you’re making, and it’s a good one.