AI Chat Tools for Museum Exhibit and Tour Planning

Vendor pages document museum-relevant uploads, research, plan limits, and data controls, but they do not establish task-specific output quality.

Sources checked 2 Oct 2026

For AI chat tools for museum exhibit and tour planning, vendor pages can tell you what each assistant accepts, which research and file-handling features exist by plan, what limits apply, and what is said about chat data and use. They do not establish task-specific output quality. Chat Picker has not tested the assistants for this work, so this page compares published documentation and gives you a self-run trial rather than a ranking.

What the vendors document

Vendor pricing and limits, as read on the linked pages in October 2026, differ by product and plan.

ChatGPT. OpenAI’s ChatGPT pricing page lists Plus at $20/month, billed monthly, and says an opt-out from training on chats is available on Free, Go, Plus, and Pro. OpenAI’s File Uploads FAQ says uploads are available on Free and paid plans, subject to limits; Free users have 3 uploads per day. OpenAI’s Deep Research help page says the feature can use uploaded files, search the public web or specific sites, and return a report with citations or source links. OpenAI’s accuracy note warns that answers can be incorrect or misleading and may sound confident, so important information needs verification.

Claude. Anthropic’s Claude pricing page lists Pro at $20/month, or $17/month with annual billing and $200 up front. It says a chat-training opt-out is available on Free, Pro, and Max, while Team is not trained on by default. Anthropic’s file-upload help page lists a 500MB limit for chat uploads and 30MB for project files. It says Claude analyzes text and visual elements in PDFs of 100 pages or fewer, but processes text only in PDFs from 101 to 1000 pages. Anthropic’s Research help page limits Research to paid Pro, Max, Team, and Enterprise plans, requires web search, and says it remains subject to standard plan limits. Anthropic’s accuracy note says not to use Claude as the only source of truth and to check cited sources and their original context.

Gemini. Google’s US AI plans page lists Google AI Pro at $19.99/month. That page does not state a chat-training setting, so Chat Picker leaves that matrix cell blank. Google’s Gemini file-upload help says users must sign in, most file types are supported, one prompt can include up to 10 supported files, and non-video files can be up to 100MB. Google’s Deep Research help says Google Search is included by default and users can add sources such as Gmail, Drive, uploaded files, and NotebookLM notebooks. Google’s sources help says Gemini Apps may show related links from websites, uploaded files, or connected Workspace content.

For museum work, vendor rules matter when assistants handle visitor information or public-facing recommendations. OpenAI’s Usage Policies prohibit unauthorized aggregation or distribution of private or sensitive information, along with real-time remote biometric identification in public spaces. Anthropic’s Usage Policy says covered advice or subjective decisions directly affecting individuals require qualified professional review before dissemination, while consumer-facing chatbots must disclose AI involvement at the beginning of each chat session. Google’s Gemini safety guidelines reject outputs that encourage dangerous activities or produce factual inaccuracies that could cause significant real-world harm, while considering educational, documentary, artistic, and scientific context.

What the documentation cannot tell you

These pages do not show whether an assistant will read a particular floor plan correctly, retain details from a long collection file, separate evidence from inference, cite the right page, or meet your operational time limit. A context-window or upload-cap figure does not answer those task questions.

The former page’s unsourced statistics and test results have been removed. Chat Picker has no replacement scores, rankings, survey figures, or benchmark claims.

How to check it yourself

Control the prompt. Give every assistant the same collection summary, sample labels, and floor plan. Use: “Create a draft walking tour from the attached files. Produce a stop sequence, timed visitor notes, an accessible alternative for each stop, and a table separating file-based facts from suggestions. Do not invent dates, ownership, access rules, or accessibility conditions; list every point requiring staff verification.” Record unsupported claims, missing sections, added assumptions, and requested clarifications.
Check grounding. Ask: “Answer only from the attached files. What is the earliest dated object in the collection summary? Which entrance and stair location does the floor plan name? Quote the supporting text and identify the file and page. If an answer is absent, write Not found.” Compare each response with the originals. Record wrong quotations, invented details, incorrect locations, and missing references.
Test files and visual reading. Supply a text-and-image exhibit PDF, a floor-plan image, and a fictional visitor-log spreadsheet. Ask: “Compare the exhibit PDF with the spreadsheet, list contradictions, then trace the proposed route on the floor plan. Quote labels and separate visible evidence from inference.” Record rejected files, warnings, missed visual details, table handling, and whether citations point to the correct place.
Track cost and latency yourself. Run the same prompt on the plan you expect to use. Before each run, record the plan, file sizes, remaining usage counter, and enabled tools. Afterward, record elapsed time, retries, warnings, and usage shown by the interface. Record only costs or usage that the product displays.
Check privacy and moderation. Use fictional visitor data rather than real names or images. Ask: “Rewrite this fictional school-group itinerary as public-facing museum copy. Do not infer personal traits from names or images. Flag missing accessibility information and use emergency wording only if it appears in the attached approved source.” Record unnecessary data requests, invented personal details, unsafe wording, policy boundaries, or missing AI disclosure. Review the documented chat-data setting before using real material.

Which rows of the comparison matter

Use the Claude vs ChatGPT matrix or the Claude vs ChatGPT vs Gemini matrix to review these rows:

Free plan
Low-cost tier
Main paid plan
Heavy-use plans
Team plan
Context window in the app
How usage limits are described
Training on your chats
Ads

Start with the rows that match your expected plan, file volume, research needs, and data requirements. Then follow the vendor links for file handling, research access, and policy details. Chat Picker’s matrices show the vendor source and read date for each figure, mark unconfirmed cells “Not verified,” and use US dollars unless a cell says otherwise. Recheck the vendor page before subscribing because plans and limits can change.