Build a Custom GPT for Your Studio's Pricing and Policies

What This Builds

The same pricing and policy questions come in from an inquiry email, an Instagram DM and a website chat widget, and the answer drifts depending on who types the reply that day or how tired you are by the fifth repeat of the week. A Custom GPT loaded with your current prices, packages and policies gives the same correct answer every time you or an assistant asks it, and it never forgets a policy the way a busy afternoon can. You still send every answer yourself. What changes is that you are not reconstructing the answer from memory or last year's price sheet each time.

Prerequisites

• Plus subscription ($20/month), required to build and use Custom GPTs: Sign up
• A current, accurate price sheet and written policies document for your studio, with any client names, addresses or booking details removed
• Comfortable using ChatGPT for basic drafting tasks (Level 3)
• Total ongoing cost for this build: $20/month, since the Custom GPT builder is included in that plan with no separate charge

The Concept

A Custom GPT works like handing a new front-desk hire a binder with every price, package and policy already written down, then telling them "answer questions using only what's in this binder." You write the binder once, as a document, and upload it. From then on, every conversation with this particular GPT starts already knowing your packages, your rescheduling policy and your print release terms, instead of you typing all of that into a fresh chat every single time a question comes up.

Build It Step by Step

Part 1: Prepare the Knowledge Files

1. Open a blank document and write out your studio's current pricing exactly as you would want a client to see it: session fees, package tiers, add-ons, travel policy, and anything you charge extra for.
2. In a second document, write out your policies in plain language: rescheduling and weather policy, deposit and payment terms, turnaround time for galleries, print and usage rights, and anything clients frequently ask about.
3. Read both documents once specifically looking for any client name, address, event date or other identifying detail that might have been copied in from an old quote or contract. Remove anything you find. These files should contain only your studio's general terms, never a specific client's information.

Part 2: Build the Custom GPT

1. In ChatGPT, look for the option to build or explore GPTs, usually reachable from the sidebar, and start a new one.
2. Use the Configure tab (skip the conversational builder if you prefer writing the instructions directly) and set the name to something like "[Studio Name] Pricing and Policy Assistant."
3. Paste in system instructions along these lines, adjusting the specifics to your studio:

You are the pricing and policy assistant for [Studio Name], a photography
studio. Answer questions about session pricing, packages, add-ons,
scheduling, deposits, rescheduling, turnaround time and print or usage
rights using only the information in your knowledge files.

If a question asks about something not covered in the knowledge files,
say so plainly and suggest the person ask the photographer directly,
rather than guessing or estimating an answer.

Keep answers warm but concise, in the voice of a small photography studio,
not a large corporate call center. Never invent a price, discount or
policy that is not in the knowledge files.

1. Upload the two documents from Part 1 as knowledge files.
2. Look for the capability settings and turn off web browsing and code execution, since this assistant only needs its own knowledge files, not the open internet.

Part 3: Test and Refine

1. Ask it five or six real questions you actually get from clients: "how much for a two hour session," "what's your rescheduling policy if it rains," "can I get a print release." Check each answer against your actual price sheet and policies.
2. If it answers a question it should have deferred, or misquotes a price, tighten the instructions in step 6 and re-test. A common fix is adding "always quote the exact figure from the knowledge files, never round or estimate."
3. Decide who else needs access. If an assistant answers inquiries for you, check the sharing setting in the GPT builder for options to share it with specific people rather than making it public, and choose the narrowest option that still lets your assistant use it.

Real Example: A Midnight Instagram Question

Setup: A wedding studio builds a Custom GPT loaded with its 2026 package pricing and a written policy document covering deposits, rescheduling and print releases.

Input: A prospective client DMs the studio's Instagram at 11:40pm asking whether the mid-tier package includes a second shooter and what the deposit is.

Output: The next morning, instead of digging through last year's price sheet to check, the photographer pastes the question into the Custom GPT and gets back the exact current answer, worded consistently with every other quote the studio has sent, ready to paste into a reply.

Time saved: Answering a one-off pricing question by memory or by searching an old document usually costs a few minutes of digging per question. Across dozens of similar questions a month, a consistent, instantly available answer removes that digging entirely, and it removes the risk of quoting last year's price by mistake.

What to Do When It Breaks

• It quotes an old price → your knowledge file itself is out of date. Update the source document and re-upload it. There is no separate step to refresh the GPT's memory beyond replacing the file.
• It answers a question that should have been deferred to you directly → the instructions in step 6 need to be more explicit about what falls outside its scope, such as custom quotes for unusual bookings.
• It sounds too formal or off-brand → adjust the tone guidance in the system instructions, and give it one or two example answers written in your actual voice to match.
• Multiple people using it get slightly different answers to the same question → check whether everyone is using the same GPT rather than separate copies each person built individually, since only one shared GPT stays in sync with your latest knowledge files.

Variations

• Simpler version: Skip the Custom GPT builder entirely and keep a single saved conversation or Project in ChatGPT with the same pricing and policy documents attached, if you are the only one who will ever use it.
• Extended version: Add a third knowledge file covering frequently asked logistics questions (parking, what to wear, weather contingency) so the same assistant handles session-prep questions too.

What to Do Next

• This week: Test it against every pricing and policy question you can remember getting in the last month.
• This month: Update the knowledge files the moment your prices change for the season, so the assistant never lags behind your real pricing.
• Advanced: Pair this with the inquiry-to-quote-to-signed-contract Zapier chain, using the same current pricing in both places so the automated quote and the Custom GPT's answers never drift apart.

Advanced guide for photographer professionals. These techniques use more sophisticated AI features that may require paid subscriptions.