Think of AI as The Equalizer. Like the movie and the show, it gives the small guy backup he couldn’t otherwise get. Two years ago, a growth study that pulled five years of your 990s, sized your market against census data, benchmarked eight competitors on admission and membership pricing, and mined visitor reviews for patterns was a five-figure consulting engagement. Most museums simply went without. That study now takes less than an hour and costs whatever you’re already paying for an AI subscription.
At the end of this article are four prompts that produce exactly that. Paste each one into an AI tool, fill in four fields about your organization, and you’ll get back a fourteen-section report: a five-year financial comparison, a market and audience profile, two peer benchmarking groups, a visitor-sentiment analysis, a set of prioritized growth opportunities with estimated upside, a 90-day action plan, and a list of the internal data you’d need to validate any of it.
It’s written for museums. If you run a different kind of nonprofit, you’ll want to adjust the vocabulary, but the structure holds. I would recommend asking the AI tool to rewrite them for you.
Use a Tool That Can Actually Search
This matters more than anything else in this article.
The prompts instructs the AI to find your Form 990 filings, look up census figures, and check competitor pricing. If you run it on a model that can’t browse the web, it will not tell you that. It will produce a beautifully formatted report with confident, specific, completely invented numbers. That’s the most common way people get burned by AI, and it’s entirely avoidable.
Use a research or web-search mode. In ChatGPT, Claude, Gemini, or Perplexity, that means turning on search or using the deep research option. If you’re unsure whether your tool is searching, check the output for source links. Real research comes back with citations and, just as importantly, with honest gaps — the prompts instructs the AI to state what it couldn’t find, and a properly working tool will.
You’ll also need your EIN. If you don’t know it offhand, look up your organization at ProPublica’s Nonprofit Explorer; it’s free and takes thirty seconds.
Read It Like a Skeptic
The output will look authoritative. Some of it will be wrong. Both things are true at once, and learning to hold them together is the actual skill.
We ran these prompts on several small children’s museums to see what they would do. The results were genuinely impressive. Every figure traced back to a real filing. The arithmetic was correct — revenue components summed to totals, assets minus liabilities matched reported net assets, year-over-year changes reconciled to the stated surpluses. That’s not nothing.
Then we read them the way a consultant reads them. In one report, we found this:
The summary table didn’t match the analysis. Three of the opportunity estimates in the final matrix were different from the calculations in the body of the same report. Not wildly different, but different enough that a board treasurer would catch it, and then question everything else.
The baseline year was the wrong year. The report concluded revenue was down 25% from 2019. But 2019 was the museum’s founding year, inflated by startup fundraising. Measured on earned revenue, the museum was up more than 80%. Same data, opposite story.
The biggest stated risk might not be a risk. The report flagged liabilities exceeding assets as the museum’s central problem. Plausible. But for a museum, a chunk of that liability line is often deferred revenue — unexpired memberships, party deposits, prepaid field trips. If so, the scary number is partly evidence of forward sales. The AI had no way to know which, and didn’t raise the question.
It didn’t know the neighborhood. The report treated a nearby town as a normal secondary market, with no awareness that a wildfire had displaced most of its population years earlier and reshaped the entire regional picture. Any local board member would spot that in one sentence, and would discount the whole document because of it.
None of this means the report was bad. It means a first draft is a first draft. Challenge the findings, ask the AI to show its work, push back on the numbers that matter most, and verify anything you’d repeat in a board meeting. The output gets meaningfully better every time you do.
What Comes After the Report
Near the end you’ll find a section called “Internal Data Needed.” That’s the honest boundary of what public information can tell you. Attendance by visitor type, membership renewal rates, contribution margin by program — none of it is public, and all of it is what turns a hypothesis into a decision.
If you have a free AI subscription, don’t upload that data. If you have a paid plan with training turned off and your governance allows it, uploading internal data unlocks a substantially deeper analysis.
And this is still the shallow end. The same technology that writes this report can reconcile your membership records against your donor database every month, draft the narrative sections of a grant report from your program data, or triage inbound school inquiries into your booking calendar. Those are the unglamorous, recurring tasks that quietly consume a development director’s week.
If You Want Help
Most nonprofit leaders we talk to have two constraints: time, and a reasonable uncertainty about where this technology actually earns its keep. That’s a fair place to be.
We’re not AI vendors. We’re two people who know this sector and happen to know this technology, in that order.
Start with a free, pressure-free, person-to-person discovery call: sagientpartners.com/contact
How to use these prompts
Each prompt runs in its own conversation and produces its own document, including opportunities and a reference list of internal data the museum could gather. None of the prompts asks you for internal or confidential data.
- Replace the bracketed placeholders with the museum’s name, EIN, city and state, and website.
- Run Prompts 1–3 in separate conversations, in any order or at the same time, and save each report.
- Start a new conversation, upload the three reports, and run Prompt 4 to produce the final report.
- Each report numbers its own sources (F for financial, M for market, C for competitors and reviews), and Prompt 4 adds N for any new sources, so the final report can combine them without renumbering.
- If the tool can’t create Word files, add “in Markdown” to the final line (Deliver this in a single formatted document in Markdown) of each prompt and convert the output afterward.
To download an MS word version of this document, including the prompts, click here
I have successfully run these prompts sequentially in a single conversation, which eliminated the need to upload the documents generated by Prompts 1–3. That may or may not work in your environment.
Prompt 1: Organization and financial analysis
Output: research coverage, organizational profile, five-year financials, financial opportunities, internal data to gather, and a summary for synthesis.
You are a nonprofit growth strategist with expertise in museums, cultural institutions, financial analysis, audience development, and fundraising.
I am a senior director exploring how our museum could increase revenue, attendance, reach, and community impact through general admissions, memberships, educational and public programs, philanthropic donations, and institutional funding, sponsorships, and partnerships.
This is one of four standalone research reports. This report covers the organization and its finances only. Produce a complete, standalone document.
Museum: [MUSEUM NAME]
EIN: [EIN]
Location: [CITY, STATE]
Website: [WEBSITE, IF KNOWN]
Rules:
– Use only publicly available information.
– Do not request, assume, or ask the user for the museum’s internal or confidential data. Where internal data would strengthen the analysis, list it as something the museum could gather; do not ask for it.
– State the date the research was conducted. For information that can change, such as leadership, report the most current information as of that date.
– Give a direct link and a publication or filing date for every material source. Number sources F1, F2, F3, and so on.
– Label every claim as Fact, Calculation, Estimate, Interpretation, or Recommendation.
– Do not invent missing figures. When reliable information is unavailable, say it could not be determined from public sources.
– Do not present estimates as forecasts or imply that public evidence proves causation.
– If you cannot complete every section, finish the current section fully, then list the sections that remain incomplete.
Sources to use: IRS Form 990 filings (the five most recent fiscal years available), ProPublica Nonprofit Explorer, Candid/GuideStar, Charity Navigator, IRS records, the museum’s website and annual reports, audited financial statements if public, and news coverage, press releases, and grant announcements.
Sections:
1. Research coverage: which sources were accessible, which were unavailable, which fiscal years were found, and how the gaps limit the analysis.
2. Organizational profile: confirm the legal name, EIN, fiscal-year dates, and tax-exempt status using authoritative sources. Note name changes, fiscal-year changes, mergers, related organizations, supporting foundations, capital campaigns, leadership changes, facility or lease arrangements, and other structural factors. Do not combine the museum’s finances with another organization’s unless reliable sources establish that doing so is appropriate.
3. Five-year financial comparison. Build a year-by-year table showing, where available:
– total revenue; contributions and grants; program-service revenue; investment and other revenue
– total expenses; program-service, management and general, and fundraising expenses
– annual surplus or deficit
– total assets, total liabilities, net assets
– major executive compensation
– number of employees and volunteers
Show reported figures separately from calculated metrics, and explain the formula for every material calculation.
4. Financial findings: long-term trends, unusual changes, one-time events, missing filings, reporting inconsistencies, revenue-concentration risks, expense, liquidity, or balance-sheet concerns, and other financial risks. Note that Form 990 data can be delayed and may not reflect the current operating position.
5. Opportunities suggested by the financial evidence (for example: revenue mix, earned-revenue growth, fundraising, cost structure, grants). For each, give:
– the supporting evidence
– the proposed direction
– a directional upside range with its assumptions, or an explanation of why it can’t be responsibly quantified
– the main risks
6. Internal data the museum could gather to strengthen this analysis. Examples: audited statements, functional expense detail, revenue by line, donor and grant history, cash reserves, staffing, and lease terms. For each item, say what it would clarify. This is a reference list for leadership, not a request.
7. Summary for synthesis:
– the 8 to 12 most decision-relevant findings, each with its source number
– a key-figures table
– open questions
– the complete numbered source list with links and dates
Deliver this as a single formatted document.
Prompt 2: Market and audience analysis
Output: market definitions, market profile, underserved groups, potential partners, market opportunities, internal data to gather, and a summary for synthesis.
You are a nonprofit growth strategist with expertise in museums, cultural institutions, audience development, and market research.
I am a senior director exploring how our museum could increase revenue, attendance, reach, and community impact through general admissions, memberships, educational and public programs, philanthropic donations, and institutional funding, sponsorships, and partnerships.
This is one of four standalone research reports. This report covers the museum’s market and potential audiences only. Produce a complete, standalone document.
Museum: [MUSEUM NAME]
EIN: [EIN]
Location: [CITY, STATE]
Website: [WEBSITE, IF KNOWN]
Rules:
– Use only publicly available information.
– Do not request, assume, or ask the user for the museum’s internal or confidential data. Where internal data would strengthen the analysis, list it as something the museum could gather; do not ask for it.
– State the date the research was conducted, and use the most recent data vintage available for each statistic.
– Give a direct link and a publication date or data year for every material source. Number sources M1, M2, M3, and so on.
– Label every claim as Fact, Calculation, Estimate, Interpretation, or Recommendation. Show formulas for material calculations.
– Do not invent missing figures. When reliable information is unavailable, say it could not be determined from public sources.
– Do not present estimates as forecasts or imply that public evidence proves causation.
– Limit the analysis to the primary and secondary markets plus the visitor market; do not profile more than five counties or equivalent areas in detail.
– If you cannot complete every section, finish the current section fully, then list the sections that remain incomplete.
Sources to use: the museum’s website, Census data (QuickFacts, American Community Survey), state and regional economic data, local tourism reports, school enrollment data, and news coverage.
Sections:
1. Research coverage: which sources were accessible, which were unavailable, and how the gaps limit the analysis.
2. Market definition: define reasonable primary, secondary, and visitor markets. Explain the geographic boundaries, the evidence supporting them, and the assumptions used.
3. Market profile. For each market, where available:
– population and household trends
– number and age distribution of children
– household income and poverty
– school enrollment
– tourism and visitor patterns
– geographic accessibility (roads, transit, parking)
– relevant cultural, educational, and leisure spending
4. Underserved groups: demographic or socioeconomic groups that may be underserved, and any access programs the museum already runs.
5. Potential institutional partners: schools, libraries, municipalities, corporations, foundations, tourism organizations, and healthcare organizations, with the rationale for each type.
6. Opportunities suggested by the market evidence (for example: audience segments, geographic expansion, access programs, school partnerships, corporate and tourism partnerships). For each, give:
– the supporting evidence
– the unmet need and the target audience
– the proposed direction
– a directional upside range with its assumptions, or an explanation of why it can’t be responsibly quantified
– the main risks
7. Internal data the museum could gather to strengthen this analysis. Examples: visitor ZIP-code origins, visitor types, school-visit history, and access-program participation. For each item, say what it would clarify. This is a reference list for leadership, not a request.
8. Summary for synthesis:
– the 8 to 12 most decision-relevant findings, each with its source number
– a key-figures table
– open questions
– the complete numbered source list with links and dates
Deliver this as a single formatted document.
Prompt 3: Competitors and visitor reviews
Output: the museum’s current offer, two competitor peer groups, visitor sentiment, competitive and visitor opportunities, internal data to gather, and a summary for synthesis.
You are a nonprofit growth strategist with expertise in museums, cultural institutions, competitive research, and visitor experience.
I am a senior director exploring how our museum could increase revenue, attendance, reach, and community impact through general admissions, memberships, educational and public programs, philanthropic donations, and institutional funding, sponsorships, and partnerships.
This is one of four standalone research reports. This report covers competitors and public visitor reviews only. Produce a complete, standalone document.
Museum: [MUSEUM NAME]
EIN: [EIN]
Location: [CITY, STATE]
Website: [WEBSITE, IF KNOWN]
Rules:
– Use only publicly available information.
– Do not request, assume, or ask the user for the museum’s internal or confidential data. Where internal data would strengthen the analysis, list it as something the museum could gather; do not ask for it.
– State the date the research was conducted. For information that can change, such as admission prices, hours, programs, and ratings, report the most current information as of that date.
– Give a direct link and a publication date for every material source. Number sources C1, C2, C3, and so on.
– Label every claim as Fact, Calculation, Estimate, Interpretation, or Recommendation.
– Do not invent missing figures. When reliable information is unavailable, say it could not be determined from public sources.
– Do not present estimates as forecasts or imply that public evidence proves causation.
– Avoid false precision, and use the same definitions and comparison periods wherever possible.
– Scope limits: profile up to 6 organizations per peer group; review up to 3 platforms per competitor (all available platforms for the museum itself).
– If you cannot complete every section, finish the current section fully, then list the sections that remain incomplete.
Sections:
1. Research coverage: which sources were accessible, which were unavailable, and how the gaps limit the analysis.
2. The museum’s current offer: admission and membership prices and benefits, hours, facility, exhibits, programs, school and camp offerings, parties and rentals, donor and sponsorship programs, and access initiatives, as published.
3. Geographic competitors: up to 6 museums, science centers, zoos, aquariums, cultural attractions, educational venues, entertainment businesses, or other destinations competing for similar audiences in the museum’s market.
4. Comparable museums: up to 6 museums with a similar mission or operating size in regions with comparable population, income, visitor demographics, tourism patterns, or proximity to a major metropolitan area. An institution may appear in both groups.
For both peer groups, explain why each organization was selected, favoring analytical relevance over count. Compare what is publicly available on:
– mission, positioning, and target audiences
– admission prices, and membership prices and benefits
– hours, accessibility, facility size, and attendance
– exhibits and visitor experiences
– educational and public programs; school, group, and camp programs
– birthday parties, events, and facility rentals
– donor and sponsorship programs
– community-access initiatives
– distinctive experiences or competitive advantages
– revenue and revenue mix, where public
5. Visitor reviews. Analyze reviews of the museum and its principal competitors on Google, Yelp, Tripadvisor, Facebook, GreatNonprofits, local publications, and community forums. Where information allows:
– compare overall ratings, approximate review volume, and recency
– compare recent reviews with older ones to identify changing perceptions
– identify recurring positive, negative, and mixed themes
– examine comments on exhibits, programs, staff, cleanliness, accessibility, prices, parking, food, age appropriateness, crowding, educational value, and overall value
– identify experiences that encourage repeat visits, memberships, recommendations, or donations
– identify frustrations that could reduce attendance, renewals, recommendations, or loyalty
– note frequently requested exhibits, amenities, programs, and improvements
– compare sentiment with competitors
– note language visitors commonly use that could inform positioning and marketing
Summarize in a table with these columns: review theme; positive/negative/mixed; approximate frequency or strength; source platforms; comparison with competitors; operational or growth implication; confidence level.
Treat reviews as directional evidence, not a representative survey. Give more weight to patterns across multiple recent reviews and platforms. Account for selection bias, small samples, isolated complaints, duplicates, extreme opinions, possible manipulation, and operational changes that make older reviews less relevant. Do not reproduce reviewer names or identifying details, and do not infer demographics unless reviewers disclose them.
6. Opportunities suggested by the competitive and review evidence (for example: pricing, membership, exhibits, programs, school visits, camps, parties, rentals, positioning, visitor experience). For each, give:
– the supporting evidence, and the review themes that support or contradict it
– the unmet visitor need and the target audience
– the proposed direction
– a directional upside range with its assumptions, or an explanation of why it can’t be responsibly quantified
– the main risks
7. Internal data the museum could gather to strengthen this analysis. Examples: ticketing and attendance data, membership conversion and renewal, visitor surveys and complaints, program capacity and pricing, and marketing performance. For each item, say what it would clarify. This is a reference list for leadership, not a request.
8. Summary for synthesis:
– the 8 to 12 most decision-relevant findings, each with its source number
– a key-figures table
– open questions
– the complete numbered source list with links and dates
Deliver this as a single formatted document.
Prompt 4: Growth strategy and final report
Output: the complete 14-section report built from the three attached reports, with prioritized opportunities, a 90-day plan, the opportunity matrix, and the five first actions.
You are a nonprofit growth strategist with expertise in museums, cultural institutions, financial analysis, audience development, fundraising, and competitive research.
I am a senior director exploring how our museum could increase revenue, attendance, reach, and community impact through general admissions, memberships, educational and public programs, philanthropic donations, and institutional funding, sponsorships, and partnerships.
Museum: [MUSEUM NAME]
EIN: [EIN]
Location: [CITY, STATE]
Website: [WEBSITE, IF KNOWN]
Attached are three research reports on this museum, produced from public information: (1) organization and financial analysis, with sources numbered F1, F2, and so on; (2) market and audience analysis, with sources M1, M2; and (3) competitors and visitor reviews, with sources C1, C2. Use them as your evidence base to develop the growth strategy and assemble the final report. Produce a complete, standalone document.
Rules:
– Base the analysis on the attached reports. Only run new research to verify or fill a specific gap, and label anything new with the source prefix N (N1, N2, and so on).
– Do not request, assume, or ask the user for the museum’s internal or confidential data. Where internal data would strengthen the analysis, list it as something the museum could gather; do not ask for it.
– Cite the source numbers from the reports.
– Label every claim as Fact, Calculation, Estimate, Interpretation, or Recommendation.
– Do not invent figures. Do not present estimates as forecasts or imply that public evidence proves causation.
– Where the opportunities in the three reports overlap or conflict, reconcile them and explain the choice.
– If you cannot complete every section, finish the current section fully, then list the sections that remain incomplete.
Develop:
1. Prioritized growth opportunities. Cover admissions; membership; exhibits and visitor experience; programs and events; school and group visits; camps, parties, and facility rentals where relevant; donations and major gifts; grants and corporate sponsorships; strategic partnerships; visitor conversion and retention; market positioning and awareness; and community reach and mission impact.
For each opportunity, provide:
– the supporting evidence
– relevant review themes, including whether they support or contradict it
– the experience problem or unmet need it addresses
– the target audience
– the proposed initiative
– its likely revenue or mission benefit
– a directional upside range, with the assumptions behind it
– required capabilities or investments
– major risks and dependencies
– a practical first pilot
– success measures
If public information is insufficient to quantify an opportunity responsibly, explain why and name the internal data that would help.
2. Risk factors and unanswered questions.
3. A recommended 90-day action plan, split into days 1–30, 31–60, and 61–90, with owners and outputs.
4. Internal data the museum could gather to validate the conclusions: transaction and attendance data; visitor types and geographic origins; membership conversion, usage, and renewal; program-level revenue, attendance, capacity, and expenses; visitor surveys and complaints; donor and fundraising history; marketing performance; and facility and staffing capacity. Link each item to the opportunities it would test. This is a reference list for leadership, not a request.
5. A prioritized opportunity matrix rating each initiative High, Medium, or Low on:
– potential financial upside
– mission and community impact
– confidence in the supporting evidence
– required investment
– implementation difficulty
– time to measurable results
Explain each rating briefly, and define whether High means faster or slower for time to results.
6. Closing section:
– the five actions leadership should consider first
– the evidence supporting each
– the most important assumptions still to test
– the internal data that would most improve confidence
Assemble the final report in this order:
1. Executive summary (written last)
2. Research coverage, key limitations, and confidence assessment
3. Organizational profile (condensed from report 1)
4. Five-year financial comparison (condensed from report 1)
5. Market and audience assessment (condensed from report 2)
6. Geographic competitive landscape (condensed from report 3)
7. Comparable-museum benchmarking (condensed from report 3)
8. Public reviews and visitor-sentiment analysis (condensed from report 3)
9. Prioritized growth opportunities
10. Risk factors and unanswered questions
11. Recommended 90-day action plan
12. Internal data the museum could gather
13. Sources and methodology (combined source list, keeping the F, M, C, and N prefixes)
14. Opportunity matrix and the five first actions
Deliver this as a single formatted document.
That’s the whole thing. Run it, read it skeptically, and if the “Internal Data Needed” section is where you get stuck, that’s exactly what the discovery call is for.
Sagient Partners was founded by John Rees and Joe Garappolo — technology veterans with 75+ combined years of experience, four companies founded, and two exits — who for 25 years have built software for and consulted to nonprofits, including museums, zoos, and cultural attractions.