Most AI for photographers advice starts with generating images. That's the least useful place to begin. Clients book you because you were there, with a camera, at their wedding, in their kitchen or in front of their product.
The hours that vanish are the ones around the shoot: the enquiry that arrives at 9pm, the proposal, the questionnaire nobody fills in, the culling of 2,400 frames, the gallery email and the blog post you meant to write in March.
What AI should and shouldn't do for a photographer
- AI drafts the enquiry reply and asks qualifying questions. You decide who you want to work with.
- AI turns the client's answers into a shot list. You decide what's possible on the day.
- AI flags soft frames and closed eyes. You choose the final gallery.
- AI drafts the blog post from your shoot notes. You supply the real details.
- AI never signs, writes or interprets a contract or model release. That stays human, and legal questions go to a lawyer.
Workflow 1: enquiry replies that qualify
The first reply decides whether you get the call. Give Claude or ChatGPT your packages, your lead times, your travel area and five of your best past replies. Ask it to draft replies that confirm the date, answer the question asked and ask two or three things you need to know: location, guest count, the use of the images.
You check: availability in your calendar, and that the price matches your current card.
Workflow 2: packages and proposals
Paste your packages and add-ons once. For each enquiry, AI assembles a proposal from the details given and lists what the client hasn't told you, rather than inventing it.
You check: hours of coverage, travel, deliverables and usage rights. Terms and licensing wording come from your contract template, pasted in, never generated fresh.
Workflow 3: shoot-prep questionnaires and shot lists
Send a short questionnaire after booking: key people, family dynamics, must-have moments, venue restrictions, times. AI turns the answers into a one-page run sheet and a shot list grouped by location and time.
You check: with the client, before the day. Anything about family situations or sensitive details needs a human read, and a quick message to confirm.
Workflow 4: culling, editing and gallery delivery
This is where photography-specific AI tools sit, and where they genuinely differ from a general assistant.
- Adobe Lightroom. Adobe's Assisted Culling page describes AI that detects issues like focus and closed eyes and sorts images into selects and rejects for review. Check which version of Lightroom you have before you plan around it, and remember you can override its decisions.
- Aftershoot. Aftershoot describes AI culling, editing in your own style, retouching and client galleries. Its site says images are processed on your computer, but read the current terms and the data-sharing option on import before you run client work through it.
- Generative tools inside editors. Lightroom's Generative Remove is one example. Removing a bin from a background is one thing. Altering faces, bodies or the moment itself is another, and that needs the client's knowledge.
You check: every cull, every time. Run the tool on an old shoot first and compare with what you would have picked.
For delivery, AI drafts the gallery email: how to download, how long the link lasts, how to order prints. You add one personal line about a specific moment from the day.
Workflow 5: blog posts, SEO and review requests
A post for each shoot helps local search, but only if it contains real information. Record a two-minute voice note after each job covering the venue, the light, the best moment and a detail the couple loved. AI turns the note into a draft with headings and image alt text.
You check: names and locations with the client's permission, and anything you wouldn't want on a public page.
For reviews, schedule a message a week after delivery that asks for one sentence about one moment. My post on repurposing one video into social posts shows the same idea for turning one piece of work into many.
Honesty about AI imagery and client trust
If you deliver photographs, the client expects photographs. A few plain rules protect that trust:
- Tell clients what AI does in your workflow: sorting, drafting, light retouching.
- Never present a generated or heavily composited image as something that happened.
- For product and brand work, agree up front whether generative fills or backgrounds are in scope.
- Keep a short written policy and send it with your contract.
Contracts, releases and photos of other people
AI can help you explain a contract in plain English to a nervous client, or list the questions to ask before you sign a venue agreement. It cannot be responsible for the document. Your contract, your model releases and your usage terms should come from a template a lawyer has checked.
Be careful with what you upload, too. Photos of children, guests and private homes belong to the people in them. Check the tool's terms and your client agreement before putting identifiable images into any online AI service.
The small stack worth learning
- Claude or ChatGPT: replies, proposals, shot lists, blog drafts
- Lightroom or Aftershoot: culling and first-pass editing, after a test
- Your gallery and booking tool: keep it, and let AI draft the emails that go through it
- Google Drive or Notion: one place for packages, templates and voice examples
Your first build: a package guide that qualifies leads
Before you automate replies, build one small thing that earns enquiries. A short "which package fits you?" quiz, where the answers pick a recommended package, collects exactly the details you ask about in Workflow 1. I explain the build step by step in how to build an AI lead magnet.
If you're unsure what to tackle first, use the scoring test in what to automate first in your business. For help building it alongside other business owners, see Wright Mode.