ChatGPT reasons. Dakota makes it specific.
Fundraising teams are already asking ChatGPT real questions: who should we call, who is active right now, and meeting prep. On its own, ChatGPT reasons well, but it relies largely on what it can find on the open web. Connect ChatGPT to Dakota Marketplace, and the same question can be answered from Dakota’s curated LP, GP, Investment, Performance, Transaction data and more.
Based on four fundraising prompts run on October 23, 2026. Results will vary by prompt and over time.
How We Tested
Each prompt is written the way a fundraiser would type it. The same prompt was run in ChatGPT with web search only, and again with the Dakota Marketplace connector enabled. With the connector, ChatGPT combines Dakota data and public web sources into one answer, and items from the web are tagged WEB or shown in a From the web column. Results on the following pages are condensed from the live responses. Contact last names are blacked out, and emails and phone numbers are omitted here.
Prompt 1 | Building a Target List
An accurate list of allocators who invest in your asset class
Prompt 1: “I’m raising a $400M lower middle market buyout fund focused on healthcare services. Who are the 10 best-fit allocators in the US to target first, and why? Include their AUM, what they’ve invested in recently, and who I should contact.”
ChatGPT with web search only | The actual response
I can’t give you a verified, ranked list of ten allocators from public sources. What I can see are recent lower middle market healthcare fund closes, which describe who backed them in general terms:
- Aphias Capital closed a $1.05bn debut fund in October 2026, backed by public pension plans, health systems, insurers, endowments, foundations, family offices and funds of funds.
- Madison River Capital closed a $370m first fund backed by family offices and institutional investors.
- Seven Hills Capital closed a $125m debut fund backed by endowments, family offices and outsourced CIOs.
- Beekman Group III drew endowments, family offices, insurers and pensions.
Navy columns: Dakota Marketplace. Grey column: public web sources found during the run. Decision-maker last names are blacked out here.
Prompt 2 | Call List
Prospecting by region and roadshow planning
Prompt 2: “Build me a list of RIAs in the Northeast with over $1B in AUM that allocate to private credit, with decision-maker contacts.”
ChatGPT with web search only | The actual response
I found directory pages that rank RIAs by regulatory AUM, but they cover the whole country and do not say who allocates to private credit. Northeast firms that appear include:
- About $2bn: Washington Crossing Advisors (Morristown, NJ), Halyard Asset Management (White Plains, NY), Avity Investment Management (Greenwich, CT), Portolan Capital Management and Boston Family Office (Boston, MA), Unison (Providence, RI).
- About $1.3bn to $1.5bn: Private Capital Advisors (Darien, CT) and Catholic Investment Services (Boston, MA).
Some firms in the results, such as Lord Abbett Private Credit Advisor, are credit managers rather than allocators. I could not confirm private credit allocation for any RIA above, and decision-maker contacts required requesting access.
What is missing: No count of matches. No private credit signal. No decision-makers.
The Largest, With CIO-Level Contacts
Cerity Partners, Focus Partners Wealth, Rockefeller Capital Management and AlTi Global list private credit, direct lending or private debt among their stated interests in Dakota Marketplace. Northeast = NY, CT, MA, NJ, PA, RI, NH, VT, ME. Contact last names are blacked out here; live results name Dakota’s CIO-category contacts for each firm.
Why it matters: the count is the story. In our run, the baseline listed firms nationwide. It could not say how many Northeast RIAs match, how many are over $25bn, or which of them allocate to private credit.
Prompt 3 | Before the Meeting
Meeting Prep
Prompt 3: “I have a meeting Thursday with Chicago Teachers’ Pension Fund. Give me a one-page briefing on their private markets allocations, recent commitments, and who I’ll likely be talking to.”
ChatGPT with web search only | The actual response
Here is what I could assemble from public sources. Some of the most relevant articles are paywalled.
- Assets: quoted at $9.6bn, $11.8bn and about $13bn depending on the article and its date.
- Early 2025: staff weighed $30m to Basis Real Estate Fund III, a $180m private equity pacing target for 2025 (including $150m to the Adams Street separate account program), and a consultant RFP for 2027 to 2031 as Callan’s contract ends June 30, 2026.
- July 2025: a search for an MWBDE real estate manager, roughly $30m to $50m.
- September 2026: a monthly pension brief references an investment consultant search update.
I could not find the current consultant outcome, recent private equity commitments, or the names of the staff you will meet.
What is missing: Current consultant status. Named contacts. Consistent, current sources.
WEB = public web source. Everything else is from Dakota Marketplace.
Even better with your own connectors. Add your own sources, such as your CRM or firmographic data, and ChatGPT can combine Dakota Marketplace, your data and the web in a single briefing. Dakota Marketplace plus your firmographic data plus the web is utopia.
Prompt 4 | Signal-Based Outreach
Investment Appetites and Activity
Prompt 4: “Which allocators have shown new activity in private equity in the last 90 days that I should reach out to this week?”
ChatGPT with web search only | The actual response
Public sources show some activity, but no single list for the last 90 days. What I found:
- Maine PERS approved up to $890m across six private markets vehicles in September, including $35m to Odyssey Investment Partners Fund VII.
- New York State Common Retirement Fund added $641m to European private equity, reported August 28 and covering June commitments.
- San Francisco ERS approved 2027 pacing of $1.2bn for private equity.
- Several other articles were paywalled or dated 2023 to 2025.
What is missing: A complete list. Commitment sizes by fund for most plans. Contacts. Anyone who is not buying.
Why it matters: Dakota and the web in one list, including who is not buying. WEB = public web source; all other rows are from Dakota Marketplace.
Put Dakota Marketplace to work inside ChatGPT
The prompts on the previous pages were typed in plain English. There is no special syntax to learn: ask ChatGPT the question you would ask a colleague, and Dakota Marketplace supplies the allocators, contacts and signals behind the answer.
-
Ask in plain English: Target lists, call lists, meeting briefs and signal checks, in the same chat you already use.
-
Get named, current answers: Allocators, decision-makers and dated events drawn from Dakota Marketplace data.
-
Act the same day: Move from research to outreach without stitching together tabs and exports.
