# Listen Labs

Aiolum story: https://aiolum.com/story/listen-labs/ · added 2026-11-12 · updated 2026-11-12

**In short.** Listen Labs (AI interviewers for customer research, United States): US$30M a year run rate, September 2026. Evidence grade D: reported, not verified. Team: 2 founders in 2023; over 80 employees in July 2026, per Contrary Research; US$100M raised by January 2026, per the company. Launched April 2025; sold 17 months after launch.

AI interviewers for customer research · United States · AI-native · started small, grew after funding · sold · group: Marketing and social content

- Outcome: US$30M a year run rate, September 2026 (evidence grade of the headline number: D; headline in US dollars: 30000000)
- Launch: April 2025; months from launch to the outcome: 17
- Team: 2 founders in 2023; over 80 employees in July 2026, per Contrary Research; US$100M raised by January 2026, per the company
- Outside funding: US$100M in total by January 2026, per the company
- Main factor: Own problem
- Meters, 1 to 4: evidence strength 1, copyable pattern 1, capital needed 3
- Tags: AI-native, Small team, Funded, Exit, B2B, Market research

Links:

- Visit listenlabs.com: https://listenlabs.com/

**The product.** An AI interviewer that runs customer research. A company describes a study, the software finds participants, interviews them by video, voice or text, and writes up what they said. Customers pay on a subscription or usage basis through order forms.

**Why it worked.** The founders built it for their own problem. An app of theirs gained 20,000 users overnight. They did not know who those users were, so they built an AI that could interview them. The company later said it grew its annualized revenue 15 times in nine months.

Evidence grades: A = disclosed by a public company, a filing or a payment processor, or verified by one; B = two or more independent reports agree, or a journalist confirmed it; C = a founder claim that no one else verified; D = reported, not verified: one founder claim or a third-party estimate with no second source; E = no revenue figure found, so the headline is a non-money fact and the other facts carry their own grades; calc = computed on this page from graded facts.

## Summary

Listen Labs sells software that runs AI-moderated interviews. An AI asks the questions, follows up on the answers and records the replies. A company can talk to hundreds of customers at once. The company's legal name is Merlin AI, Inc., doing business as Listen Labs, a Delaware corporation. Its customers include Microsoft, Anthropic and Sweetgreen, named by the company and by the press.

A company blog post says two founders moved into a San Francisco hacker house in January 2023. SiliconANGLE attributes the post to co-founder and Chief Executive Alfred Wahlforss. An app they had built had picked up 20,000 users overnight, and the founders did not know who the users were. Over a hackathon they built an AI that could interview them. The company launched Listen in April 2025. In January 2026 the company wrote that it had grown annualized revenue 15 times since the launch and had raised US$100 million in total. Contrary Research counted over 80 employees in July 2026.

The headline figure is a report, not a verified number. SiliconANGLE wrote on 1 October 2026 that annualized revenue was running at about US$30 million. It took the figure from a TechCrunch report of September 2026. Neither company disclosed the terms of the Salesforce deal, and the price of about US$2 billion is a press report. Salesforce signed an agreement to buy the company, and no source read here says the deal has closed.

## Why did Listen Labs work?

### The founders built the interviewer for their own users

The company's post says an app of the founders had 20,000 users overnight and they did not know who the users were. They built an AI to interview them over a hackathon. The product is that tool, grown into a platform that also finds the participants and analyzes the replies.

### The buyers are large companies that pay by credits and order forms

The Master Services Agreement describes subscription or usage sales under an order form, with credits for fees. The company names Microsoft, Anthropic and Sweetgreen. Microsoft's research manager says in a case study that she can reach hundreds of users at one third of the cost.

### The company says it raised US$100 million as revenue grew 15 times

The company says it raised US$100 million in total by January 2026. In that post it also says its annualized revenue grew 15 times in nine months. No source here shows how the money was spent.

## How it was built

Few build details are published. These come from the company's own pages and are grade C.

- The product is software as a service. Customers design a study, reach participants from a network, and read the analysis.
- The interviews are AI-moderated and run in more than 120 languages, according to Salesforce's statement as SiliconANGLE reports it.
- The engineering blog of July 2026 says the company runs Claude Code in a loop in a cloud environment. The loop improves its own product against defined metrics.
- In 2026 the company added a simulation product, Listen Twins, which the home page describes as a new product. The product page says each twin is a model of one real participant, built from that person's interview.

Sources: [5, 4, 11, 1, 15]

## Who tried this before Listen Labs, and why did they fail?

Three companies in our database of failed startups built tools for gathering feedback or opinions. The values below are from that database and are not independently verified.

The failed startups come from our database of dead startups, retold here in our own words; Its entries are AI-assisted summaries of public sources, so they carry no evidence grade.

### User Hero (2018 to 2021; funding US$5.0M total cash burned; cause: Competition)

- Promise: A platform to collect, analyze and show user feedback for product teams, with surveys and interviews.
- Why it died: The record says it struggled to differentiate itself in positioning and user acquisition. Survey platforms such as Qualtrics and SurveyMonkey dominate the field. The same record also says its description suggests the company was still active in 2024 or 2025.
- Record: 

### Team Voice (2017 to 2020; funding US$2.5M total cash burned; cause: No demand)

- Promise: A software service for feedback and communication inside teams.
- Why it died: The record says the platform did not match the human side of HR work, and demand and adoption stayed low.
- Record: 

### Pollground (2006 to 2009; funding US$150K total cash burned; cause: Competition)

- Promise: A place to create and answer public opinion polls.
- Why it died: The record says it could not compete with larger social networks that added the same function.
- Record: 

## What changed

Each of the three sold a way to collect opinions with forms, polls or feedback tools. The record pages say they were out-competed by larger platforms, or that buyers did not adopt them. The work that a person did, asking follow-up questions and reading the answers, stayed manual.

Listen Labs sells the interview itself, run by an AI that asks follow-up questions. According to the Microsoft case study, a human researcher can reach hundreds of users in days. That is the change from survey tools. Whether the figures hold up is not verified here.

## Can one person copy Listen Labs now?

**Verdict: Not for one person, today. The idea is open; the scale was not.**

Two founders reached large customers with an interviewing product, and raised US$100 million according to the company. A solo founder could copy the idea of an AI interviewer for a narrow audience, but not the reported scale.

1. Start from a research job you have done yourself, such as interviewing your own users.
2. Let an AI ask the follow-up questions, and keep the customer's own words in the report.
3. Sell to teams that already pay for research and want results in days, not weeks.

**What is still open.** A narrow audience, for example one industry's customer interviews, without the cost of a panel of 50 million participants.

## How much did Listen Labs make, and what was it sold for? The numbers, graded

A report of US$30 million in annualized revenue in September 2026, 17 months after launch. Annualized revenue is the latest month multiplied by 12, and no source here is a filing or a processor

| Fact | Value | Grade | Note | Sources |
|---|---|---|---|---|
| Annualized revenue, September 2026 | about US$30 million | D | SiliconANGLE, 1 October 2026, retelling a TechCrunch report of September 2026: "Annualized revenue was running at about $30 million at the time". The time meant is when the company left a planned funding round for the Salesforce talks. Neither source is a filing or a processor. | [1] |
| Growth of annualized revenue after launch | 15 times in nine months | C | The company's Series B announcement, 14 January 2026: "Since launching nine months ago, Listen has grown its annualized revenue by 15x". The base figure is not given. | [3] |
| Start of the work | January 2023, in a San Francisco hacker house | C | The company's blog post of 29 September 2026, which SiliconANGLE attributes to co-founder Alfred Wahlforss. The AI interviewer began as a hackathon project. | [2, 1] |
| Launch of Listen | April 2025 | C | The same blog post: "In April 2025, we launched Listen". An Internet Archive copy of 19 April 2024 already shows the AI interviewer with a "Request Demo" button. | [2, 14] |
| Founding team | 2 founders: Alfred Wahlforss, Chief Executive, and Florian Juengermann, Chief Technology Officer | C | SiliconANGLE names both as co-founders. The company's post says that Florian and the writer moved into the hacker house. | [1, 2] |
| Team today | over 80 employees in July 2026 | D | Contrary Research, profile updated 22 October 2026: "Listen Labs had grown to over 80 employees as of July 2026". Its summary line shows 99 employees. The company's own pages give no headcount. | [13, 8] |
| Company founded | September 2023 | D | Contrary Research: "Listen Labs was founded in September 2023". The company's own post dates the start of the work to January 2023, and its pages say "Estd. 2023". | [13, 2, 10] |
| Funding | US$27 million seed and Series A, April 2025; US$69 million Series B, January 2026; US$100 million in total | C | The Series B figure and the April 2025 round are from SiliconANGLE. The company's own Series B post states US$100 million in total and names Ribbit Capital as lead. Calculated here: the two rounds add up to US$96 million. | [1, 3] |
| Legal entity | Merlin AI, Inc., doing business as Listen Labs, a Delaware corporation | C | The company's Master Services Agreement, last updated 6 February 2026. No downloaded page dates the entity. An archive copy of April 2024 calls the product Merlin in two places. | [4, 14] |
| Named customers | Microsoft, Anthropic, Sweetgreen | C | The company's blog post names them. SiliconANGLE repeats that Microsoft and Anthropic have signed on as customers. The Microsoft case study quotes a Microsoft research manager. | [2, 1, 6] |
| Participant network | more than 50 million potential participants; interviews in more than 120 languages | C | SiliconANGLE reports these as Salesforce's statement. The home page also shows "50M+ participants" with no source. | [1, 5] |
| Deal with Salesforce | definitive agreement signed; price about US$2 billion reported; terms not disclosed | D | SiliconANGLE, 1 October 2026: the price is from a Business Insider report of 9 September 2026, and neither company disclosed terms. Close is expected in the fourth quarter of Salesforce's fiscal 2027. | [1] |
| Time from launch to the September 2026 figure | 17 months | calc | Calculated: April 2025 to September 2026. | [2, 1] |

How to read this table. Grade D means reported, not verified: one press retelling of a report by another outlet. Grade C is the company's own statement. A "calc" value was computed on this page from the rows above it.

## Pricing

Listen Labs has no published price list; the company's pages offer a demo and describe its charging terms

| Plan | Per month, billed yearly (US$) | Per month, billed monthly (US$) | Limit |
|---|---|---|---|


The home page offers "Book a demo" and shows no prices. The Master Services Agreement says services are sold on a subscription or usage basis under an order form, with credits bought to pay fees. The company's September 2026 release note describes a dashboard that shows credit spend per study. Contrary Research reports a typical entry point of about US$20,000 a year plus US$300 to US$400 per completed interview, from buyer reviews.

## When did Listen Labs launch, and how did it grow? Timeline

From a hacker house in January 2023 to a signed agreement with Salesforce in October 2026. Each entry carries its grade.

- **January 2023** (grade C [2]): Two founders move into a San Francisco hacker house after an app they built picks up 20,000 users overnight. A hackathon project becomes the AI interviewer.
- **April 2025** (grade C [2, 1]): Listen launches. The company announces a US$27 million seed and Series A round led by Sequoia Capital, according to SiliconANGLE.
- **14 January 2026** (grade C [3, 1]): Series B announced. The company says total funding is US$100 million and annualized revenue has grown 15 times since launch. SiliconANGLE gives the Series B as US$69 million.
- **September 2026** (grade D [1, 9]): TechCrunch reports annualized revenue of about US$30 million, according to SiliconANGLE. The company releases a usage and billing dashboard on 13 September.
- **29 September 2026** (grade C [2]): The company's blog announces that Listen Labs is joining Salesforce. The chief executive and chief technology officer stay on, inside Salesforce AI Labs.
- **1 October 2026** (grade D [1]): SiliconANGLE reports a definitive agreement. The price of about US$2 billion is from Business Insider; neither company disclosed terms.

## What could go wrong for Listen Labs?

- **What is not verified.** The US$30 million annualized revenue rests on one press retelling. SiliconANGLE, 1 October 2026, took it from a TechCrunch report that this page could not download. A reader could look for the TechCrunch report or a filing by Salesforce.
- **Sale not closed, price reported only.** The sale is agreed, and no source read here says it has closed. SiliconANGLE reports a signed agreement and an expected close in Salesforce's fourth fiscal quarter of 2027. The company's home page already says "now part of Salesforce". The price of about US$2 billion is a Business Insider report from the talks; neither company disclosed terms. Salesforce's own announcement could not be opened.
- **Home page logos, counters and quotations are the company's own.** The home page shows customer logos under "The research partner for hundreds of leading brands", with no source. It shows counters such as "50M+" participants, "120+" languages and "3x" longer responses. Its three quotations carry a name and a job title. The tiles "150+" and "300+" are figures from the company's own case studies. A web archive copy of listenlabs.ai from March 2022 shows a different project with the same name.
- **Simulated customers.** Listen Twins simulates the answers of real research participants, and the pages read do not say how participants agree to that. The product page calls each twin "a 1:1 model of a real person in your target audience". It says each answer can be traced to the participant's interview. The Study Privacy Policy lists running the study and improving the service as purposes, and does not name twins. The company's AI code of conduct says it does not use collected data "for training our AI models". No page read offers a copy of a participant's voice or face.
- **Founding date and earlier name.** The start year 2023 rests on the company's own post and on Contrary Research. The post dates the work to January 2023, and Contrary Research gives September 2023 as the founding date. The legal name is Merlin AI, Inc., and an archive copy of April 2024 calls the product Merlin in two places. No downloaded page gives a date for the legal entity. The company dates its launch to April 2025, but the April 2024 copy already offered a demo.
- **The founders' earlier app.** The app that came before Listen was BeFake, according to Contrary Research. It describes BeFake as "an AI avatar generation and social network" and says the founders sold it. No Listen Labs page read here sells or links to that app.

## Sources

1. SiliconANGLE, 1 October 2026: Salesforce to acquire AI customer research startup Listen Labs in reported $2B deal — https://siliconangle.com/2026/10/01/salesforce-to-acquire-ai-customer-research-startup-listen-labs-in-reported-2b-deal/ (grade D). Read in full. Source for the US$30 million figure, the funding rounds and the deal terms. It retells TechCrunch and Business Insider, which this page did not read.
2. Listen Labs blog, 29 September 2026: Listen Labs is joining Salesforce — https://listenlabs.com/blog/listen-labs-is-joining-salesforce (grade C). Read in full. The company's announcement. It is unsigned on the page; SiliconANGLE attributes it to Alfred Wahlforss.
3. Listen Labs blog, 14 January 2026: Announcing Listen's Series B — https://listenlabs.com/blog/announcing-listen-s-series-b (grade C). Read in full. Source for the US$100 million total and the 15 times growth.
4. Listen Labs Master Service Agreement, last updated 6 February 2026 — https://listenlabs.com/msa (grade C). Read in full. Source for the legal entity and for how services are charged.
5. Listen Labs home page, read 12 November 2026 — https://listenlabs.com/ (grade C). The copy of the home page read for this page was taken from listenlabs.ai on 12 November 2026, after the sale announcement. It shows counters and a demo form but no prices.
6. Listen Labs case study: How Microsoft uses Listen Labs — https://listenlabs.com/case-studies/microsoft (grade C). Read in full. The company's own case study quoting a Microsoft research manager.
7. Database of failed startups, local export — /method/#dead (grade C). Record pages read on the website on 12 November 2026. AI-written summaries, not independently verified.
8. Listen Labs careers page, read 12 November 2026 — https://listenlabs.com/careers (grade C). Read in full. It gives no headcount.
9. Listen Labs release note, 13 September 2026: New Usage and Billing Dashboard — https://listenlabs.com/whatsnew/new-usage-and-billing-dashboard (grade C). Read in full. It describes credit spend per study.
10. Listen Labs terms of use, last updated 13 May 2025, and privacy policy of 23 February 2026 — https://listenlabs.com/terms (grade C). Read in full. The terms name Merlin AI, Inc. and the pages say "Estd. 2023".
11. Listen Labs engineering blog, 14 July 2026: Autoresearch At Listen — https://listenlabs.com/blog/autoresearch-at-listen (grade C). Read in full. Source for the use of Claude Code.
12. Internet Archive copy of listenlabs.ai, 10 March 2022 — https://web.archive.org/web/20220310222608/https://www.listenlabs.ai/ (grade C). Read in full. A site with menu items such as "The Science" and "Share Your Voice"; used only to note that the domain held another site in 2022.
13. Contrary Research: Listen Labs business breakdown and founding story, updated 22 October 2026 — https://research.contrary.com/company/listen-labs (grade D). Read the sections on the founding, the team and the business model. It does not name its source for the headcount.
14. Internet Archive copy of listenlabs.ai, 19 April 2024 — https://web.archive.org/web/20240419083852/https://listenlabs.ai/ (grade C). Read in full. The company's own home page of that date, with "Request Demo" and the name Merlin in two places.
15. Listen Labs: Listen Twins product page, read 12 November 2026 — https://listenlabs.com/twins (grade C). Read in full, with the Study Privacy Policy and the AI code of conduct on the same site.

Contact: aiolum@livyden.com. Run Listen Labs? Send corrections or a verified figure to aiolum@livyden.com (subject: Correction: Listen Labs).