Carta's AI Customer Research Strategy: How the $7B Equity Platform Listens to 40,000 Companies

TL;DR#

Carta, the $7 billion equity-management platform that administers cap tables for more than 40,000 private companies and 2 million security holders, runs customer research across four wildly different personas — founders, employees, investors, and law firms — that each use a different vocabulary for the same transaction. CEO Henry Ward has publicly framed Carta's strategy as "the layer between every private company and its capital," which means a single survey instrument has never been able to capture the platform's full voice-of-customer. Carta's modern research stack treats every cap-table event, 409A valuation, and CartaX transaction as a conversation rather than a form — letting founders talk about dilution in dilution language, employees ask about strike prices in employee language, and GPs probe diligence questions in fund language. This post breaks down how a multi-sided fintech platform uses AI customer interviews to do customer research at scale, what 409A and private-market data add to the conversational layer, and what other multi-sided SaaS platforms should copy. Across the 25-post batch we publish on AI customer research strategy, Carta sits next to peers like the Stripe customer research playbook for 4 million businesses, the Ramp onboarding strategy for finance teams, and the Notion research operating model for a $10B product company — and the through-line is the same: AI-first research starts with a conversation, not a form.

What is Carta doing with AI customer research?#

Carta is replacing static feedback forms across founder onboarding, employee equity education, investor diligence, and law-firm workflows with AI-moderated conversational research that adapts to each persona's vocabulary in real time. Carta's product surface area — cap-table management, 409A valuations, fund administration, the CartaX private-markets venue, and the equity portal millions of employees see — generates a continuous stream of moments where the company needs to learn why a customer just did something. A founder reading a SAFE conversion modal has a fundamentally different question from a Series C controller closing a quarter or a portfolio analyst at a fund running secondaries diligence. A single survey form built for any one of them flattens the other three. AI customer interviews let Carta ask the right follow-up question to the right persona in the same research instrument — a pattern that mirrors what Brex's customer research strategy across 30,000 startup customers and Mercury's onboarding research for 200,000 founders describe in adjacent fintech segments.

Carta has publicly disclosed that its platform manages equity for over 2 million security holders and more than $2.5 trillion in equity value — a customer base whose conversational density makes form-based research a structural mismatch. The harder problem isn't volume; it's that one transaction (say, an early-employee exercise) is simultaneously seen by the founder, the employee, the cap-table admin at a law firm, and the investor on the other side. Each of those four people has a legitimate reason to give feedback in their own words.

Why Carta's customer base breaks single-survey research#

Carta's customer base breaks single-survey research because four distinct personas with four incompatible vocabularies share three sides of every cap-table transaction — and no NPS form can capture all twelve combinations. The platform sits at the intersection of corporate finance, employment law, securities regulation, and venture capital. A "satisfaction" question that means one thing to a founder ("did this fundraise close cleanly?") means something completely different to an employee ("did I understand my offer?") and something different again to a fund LP ("did I get a defensible mark on this position?"). When a research team tries to fit all four mental models into one Likert scale, the signal collapses into noise.

This is exactly the failure mode Gartner has documented in VoC programs at multi-stakeholder fintech platforms. According to Gartner's research on voice-of-customer programs, more than 95% of organizations collect customer feedback, but fewer than half use it to drive action — and the gap is widest in B2B2C platforms where the buyer, the user, and the beneficiary are different people. Carta is the textbook B2B2C example. The buyer is often a founder or CFO. The daily user is often a controller or paralegal. The beneficiary is the employee or investor. Forms force all three to answer the same questions.

The four personas, briefly:

• Founders — speak in fundraise language: SAFEs, post-money caps, dilution, runway, secondaries. They want speed and want to understand trade-offs in plain English.
• Employees — speak in offer-letter language: strike price, vesting cliff, exercise window, tax withholding, ISO vs. NSO. Most have never owned equity before and need definitions, not jargon.
• Investors — speak in fund language: NAV, marks, distributions, capital calls, 409A defensibility, secondary liquidity. They want auditable data, not opinions.
• Law firms — speak in legal-ops language: stockholder consents, board approvals, 83(b) elections, Rule 701 limits, exemption tracking. They need workflow tools, not marketing surveys.

Three sides of every transaction means the same equity event triggers research-worthy moments for at least three of those personas simultaneously. A primary fundraise involves the founder, the new investor, and the law firm. An employee exercise involves the employee, the founder (cap-table impact), and the broker/agent. A secondary on CartaX involves the seller, the buyer, and the issuer. Conversational research is the only modality that can ask each side its own question in its own vocabulary in a single research instrument — the same insight that drives the 2026 AI Research Stack Report covering 100 SaaS teams that replaced survey tools.

Inside Carta's multi-persona conversational research#

Carta's multi-persona conversational research splits into three flagship motions — founder onboarding, employee equity literacy, and investor discovery — each running as a separate conversational track but feeding one consolidated voice-of-customer layer. Walking through each shows why AI moderation matters more than form templating.

Founder onboarding research#

Carta's founder onboarding research uses conversational interviews to understand why a founder picked Carta over the incumbent and where the platform falls short for first-time vs. repeat founders. A founder going from Notion-and-a-spreadsheet to a real cap-table tool has fundamentally different questions from a third-time founder who has already done a Series C — and the form fields a research team would write for one of those journeys are almost useless for the other. Conversational research lets the AI moderator ask, "what were you doing for your cap table before Carta?" and then probe on that answer in real time. A second-time founder might say "we were on a competitor and switched"; a first-time founder might say "we were on a spreadsheet and finally raised enough to need this." Both answers route to completely different follow-up sets. This pattern mirrors what we describe in the jobs-to-be-done interview template and what the How Top Founders Are Rethinking Customer Research playbook covers in depth.

Employee equity literacy research#

Carta's employee equity literacy research reaches the ~2 million security holders on the platform and discovers, in their own words, what they actually understand about their equity grant — separately from what HR thinks they explained. The single biggest research opportunity at any equity platform is the gap between how employers describe an equity offer and how employees experience it. Forms make this gap invisible because the questions are written by the employer side. Conversational research lets an employee say, "I have no idea what my strike price is supposed to mean," and the AI moderator picks up on the uncertainty marker ("no idea") and asks a clarifying follow-up rather than recording a "neutral" Likert score. The same approach powers the employee onboarding survey template and the employee experience interview template.

Investor discovery research#

Carta's investor discovery research interviews GPs, LPs, and fund administrators about how they actually use Carta for marks, capital calls, and secondary liquidity — three workflows that look similar in a survey and are radically different in practice. An LP wanting a clean quarterly statement and a GP wanting flexibility on a fair-value mark are answering different research questions even when the survey asks "rate your satisfaction with reporting." Conversational interviews split those personas dynamically by detecting which side of the fund relationship the respondent is on and routing follow-ups accordingly. For research teams designing this kind of multi-segment flow, the buyer persona interview template and the customer segmentation interview template are the structural starting points.

409A and private market data — the conversational angle#

409A valuations and private-market data are the inputs that make Carta's conversational research more accurate than a generic SaaS feedback program — because every question can be grounded in the respondent's actual equity context. A 409A valuation is the IRS-required fair-market-value calculation that determines an employee's strike price on a new option grant. Carta is the largest provider of 409As in the U.S., with the company publicly stating it has issued more than 38,000 409A valuations over its history. That gives Carta an unusual research advantage: it can run a conversational interview where the AI moderator already knows the respondent's industry, last-round valuation, and grant timing, and can ask context-aware questions like "your strike price changed last quarter — do you understand why?" rather than a generic "rate your satisfaction with equity reporting."

Private-market data — secondary transaction history, comparable company multiples, exit benchmarks — adds the same grounding to investor research. When Carta talks to a GP about portfolio marks, the AI moderator can reference the specific marks the fund has taken and ask conversational follow-ups in the GP's own language. This is the conversational equivalent of Stripe's customer research approach with 4 million businesses — context-rich AI interviews that no human researcher could scale. According to a McKinsey analysis of fintech customer-experience programs, context-aware research instruments deliver 2–3x the actionability of generic NPS programs, because the respondent doesn't have to translate generic questions into their own situation.

The CartaX angle is even more interesting. CartaX is Carta's private-markets venue for secondary transactions — essentially a regulated exchange for shares in private companies. Every CartaX transaction generates feedback signals from both sides (seller, buyer) plus the issuer, and traditional survey forms cannot meaningfully capture seller-side emotion ("am I leaving money on the table?") and buyer-side diligence ("did I get the data I needed?") in the same instrument. Conversational AI can, and the resulting data improves the design of the venue itself.

What this signals for SaaS platforms with multi-sided customer bases#

What this signals for SaaS platforms with multi-sided customer bases is that the era of one-survey-fits-all VoC programs is over — and that conversational AI is the only research modality that scales to four-plus personas without flattening any of them. If your product sits at the intersection of more than two stakeholder groups, you are already paying a hidden tax on every survey you ship: the question that's clear for buyers is confusing for users, and the question that's clear for users is meaningless to admins. Carta is one of the cleanest examples of this problem at scale, but the same pattern appears at any platform with a similar topology.

The signal for product, research, and CX leaders is threefold:

1. Treat each persona as its own research track. A founder, an employee, an investor, and a law firm are not the same customer — they should not share a survey form. Conversational AI interviews can run all four tracks under one research umbrella while keeping the question logic distinct. This is the same principle described in the playbook of best AI customer discovery platforms for founders.
2. Ground every conversation in the respondent's actual product context. Carta has 409As and cap tables; Stripe has transaction volume; Notion has workspace data. Whatever your platform's "data spine" is, the AI moderator should know it before it starts the interview. See the Glean enterprise search customer research approach for the enterprise-software version.
3. Replace your "rate your satisfaction" reflex with a "tell me what just happened" question. Forms ask for scores. Conversations ask for stories. Stories are the raw material of product decisions — see the case for AI moderated interviews and the mechanics of good AI interviewing in 2026 for the structural argument.

For research and CX leaders who want to operationalize this, Perspective AI is built for product teams and for CX teams running this kind of multi-persona research at scale. The Interviewer agent handles the conversational moderation; the Concierge agent replaces traditional intake forms with conversational onboarding. Together they cover the founder-side and employee-side flows Carta-style platforms need.

Frequently Asked Questions#

How does Carta use AI customer interviews at scale?#

Carta uses AI customer interviews to run separate conversational research tracks for founders, employees, investors, and law firms — each grounded in the respondent's actual cap-table and 409A context. Rather than sending one NPS survey to all four personas, the AI moderator adapts vocabulary, follow-up questions, and definitional support to the persona answering, then consolidates the signal into one voice-of-customer layer. The approach scales to millions of security holders without flattening any persona's vocabulary into a shared dropdown.

Why don't traditional surveys work for equity-management platforms like Carta?#

Traditional surveys don't work for equity-management platforms because the same transaction is simultaneously experienced by founders, employees, investors, and law firms — and each group uses incompatible vocabulary for it. A "rate your satisfaction with equity reporting" question is interpreted four different ways by four personas, so the resulting Likert score is noise rather than signal. Conversational AI interviews fix this by detecting which persona is responding and routing follow-up questions accordingly.

What is a 409A valuation and how does it intersect with customer research?#

A 409A valuation is the IRS-required fair-market-value assessment of a private company's common stock used to set employee strike prices on new option grants. It intersects with customer research because the 409A and the resulting strike price are the source of most employee confusion about equity — making them a high-value research surface. Equity platforms that can ground a conversational interview in a respondent's actual 409A context capture far richer feedback than generic equity-satisfaction surveys.

Who runs customer research at Carta?#

Customer research at Carta is distributed across product managers, customer experience leaders, designers, and the company's own market-research function — with each team responsible for the persona segments closest to its workflow. Henry Ward, Carta's CEO, has publicly framed customer-listening as a CEO-level responsibility because the platform's multi-sided nature means insights from one persona constantly inform product decisions affecting another. AI customer interviews allow this distributed research model to operate without losing consistency across persona tracks.

How does Carta's customer research compare to other fintech platforms?#

Carta's customer research is unusually multi-sided compared to peer fintech platforms because the same event touches founders, employees, investors, and law firms simultaneously. Stripe's research is more buyer-centric (the developer/finance buyer), Brex's and Mercury's are more founder-centric, and Ramp's is more finance-team-centric — Carta has to talk to all four sides of an equity transaction at once. This is why Carta's research stack leans more heavily on conversational, context-aware AI moderation than on traditional segmented NPS.

What can SaaS platforms learn from Carta's approach?#

SaaS platforms with multi-sided customer bases can learn three things from Carta's approach: treat each persona as its own research track, ground every conversation in the respondent's actual product context, and replace "rate your satisfaction" forms with "tell me what just happened" conversations. The pattern applies to any platform where buyer, user, and beneficiary are different people — marketplaces, B2B2C platforms, vertical SaaS with admin-and-end-user splits, and any product with a procurement layer on top of a daily-user layer.

Conclusion: the equity-management blueprint for AI customer interviews#

Carta's approach to AI customer interviews is the cleanest demonstration available of why multi-sided SaaS platforms cannot run customer research through a single survey form. With 40,000+ private companies, 2 million+ security holders, more than 38,000 409A valuations issued, and a customer base split four ways across founders, employees, investors, and law firms, the company has every structural reason to treat customer research as a conversational layer rather than a form layer. The lesson generalizes: any platform sitting between multiple stakeholders in a transaction needs research that adapts to each side's vocabulary, grounds itself in the respondent's actual product context, and turns "rate this experience" into "tell me what happened in your own words."

If you're running customer research for a multi-sided product — equity platforms, fintech, marketplaces, or any B2B2C SaaS — Perspective AI is built for exactly this problem. Start a free conversational research project at the Perspective AI research workspace, browse interview templates including the buyer persona and jobs-to-be-done interview templates, or see how the Interviewer agent runs multi-persona research at scale. The form era of customer research is closing. Carta is already operating in the next one.