← Cases
United StatesFinance and insurance2026 – ongoing

Pacific Community Ventures

AIKKA voice survey tool that tags worker feedback with Claude

Pacific Community Ventures reached about 300 workers in one voice survey, up from 12 in a focus group, using Claude to code answers

WorkingVendor case studyVendor ↗

Problem

Pacific Community Ventures is a community lender for small business owners. Its research on worker experience relied on slow focus groups. Two years before the story, one focus group took several weeks and reached 12 people.

Approach

The team built AIKKA, a smartphone tool where workers answer open questions aloud in their own language. A transcription engine converts the speech, and Claude tags each response for sentiment, emotion and theme. The team checked accuracy in English, Spanish, Hindi and Punjabi. Data scientists moved from reviewing nearly every response to spot-checking.

Results

about 300workers reached in one nationwide voice survey, against 12 in an earlier focus group, per AnthropicVendor case study[1]
10x or moregrowth in qualitative research capacity, per AnthropicVendor case study[1]

Takeaway

Automated coding of spoken answers let a small research team reach far more people than a focus group could. The lender was also testing a Claude co-pilot for loan applications, and the page reported no result for it.

Categories

Innovation type
Process innovation
AI technology
Generative AI
Evidence
Vendor case study

Sources

Source summaries

[1]Anthropic14 Jul 2026

Anthropic's customer story describes how the community lender Pacific Community Ventures used Claude to tag spoken survey answers in its AIKKA tool. It gives about 300 workers reached against 12 in a focus group, and research capacity of 10x or more. The page carries no date; the Internet Archive first saved it on July 14, 2026.

Full details are in the original sources.

Added 4 Nov 2026 · Updated 4 Nov 2026
Open case page ↗