Book a Demo Try Speak AI Free Login
Case study · Research · A research agency

How a Research Agency Launches a White-Label Qualitative Platform With Embedded Recorders

A qualitative research agency whose clients wanted larger samples in several languages embedded branded recorders and a client layer on Speak, instead of building its own capture, transcription and reporting platform.

Book a DemoTry Speak AI Free
As of May 2026Measured from Speak usage records
  • 5,600+
    Recordings processed
  • About 150
    Hours of audio and video
  • 14
    Branded recorders configured
  • 29
    Fields set up for study attributes

Summary

Clients wanted qualitative depth at a larger scale, in several languages, with the respondent's original wording preserved. A traditional interview programme could not reach that sample size on a normal budget.

The agency placed branded recorders inside its own platform so respondents answer on video or audio, added fields for the attributes it needed, and gave clients a white-label library to explore results themselves.

As of May 2026 the account held more than 5,600 recordings and 6,000 transcripts.

A confident woman researcher sits at a laptop in a modern studio office, looking toward the window.

Three lines to take away

Before

Clients asked for larger qualitative samples in several languages, with exact original wording, and a custom build looked costly.

With Speak

Embedded branded recorders capture respondent answers, and a white-label client layer lets clients explore the results.

After

The agency ran larger qualitative samples on a platform it did not have to build, with transcripts and structured fields behind each answer.

Larger samples, more languages, no platform to run them

The agency's clients wanted the richness of qualitative research at a scale it was not designed for.

A young woman sits on an office floor among printed pages and notes, comparing handwritten translations.

Samples out of reach

Moderated interviews do not stretch to the sample sizes clients now ask for. Each extra respondent added scheduling, moderation and transcription time.

Many languages

Respondents answered in different languages. The agency needed consistent permissions and exports for each, and the exact original wording alongside any translation.

Build or buy

Doing it in house meant building capture, transcription, translation and export tooling, and then supporting it. That was a large commitment before a single study ran.

The cost: The choice was between turning down larger qualitative projects and funding a custom platform, and neither was attractive for an agency of its size.

The turning point

Embed the capture and give clients a window

The turning point came from seeing that the agency did not need to build a platform. It needed the recorder inside its own surveys and a way for clients to see the results without a training session. An embedded recorder meant respondents never left the agency's environment, and a white-label library meant clients opened something that looked like the agency's own product.

The sceptical question was fidelity. A research agency cannot work from paraphrase, so the transcripts had to keep the respondent's own words in each language. Once the team saw transcripts alongside the recordings and the fields it had defined for its study attributes, it moved the first larger-sample studies onto the setup.

How they use Speak AI

Respondents answer inside the agency's platform, and clients see structured results in a branded library.

  1. Step 1

    Embedded recorders

    Branded audio and video recorders sit inside the agency's own platform, so respondents answer a prompt without installing anything or leaving the page.

  2. Step 2

    Structured fields

    Fields hold the attributes each study needs, and the agency loads respondent lists and past material in bulk by import and API.

  3. Step 3

    Multilingual transcripts

    Each answer is transcribed in the language spoken, keeping the original wording available beside any translation.

  4. Step 4

    Client library

    Clients explore recordings, transcripts and themes through a white-label library and ask questions across a whole folder.

Honest friction: Fields, recorders and imports needed careful setup for each study, and the agency worked through several rounds of guided configuration. This page describes the build as it stood in May 2026.

Two colleagues at a table record a short video message on a phone propped upright.

Book a Demo

Results

What the platform held as of May 2026

As of May 2026, the account held more than 5,600 recordings and about 150 hours of audio and video, most of them short respondent answers, brought in by a mix of recorders, API and import.

6,000+
Transcripts held
2,100+
Recordings brought in by bulk import
11
Automations configured
How we calculated this

An estimate, not a measurement

The usage numbers on this page are measured from Speak records. This block is a separate estimate of the manual effort the same workload would take. It uses published time benchmarks at the low end of their ranges and official wage statistics at the 25th percentile. Most teams already have transcripts, so Speak’s value here is the analysis layer.

Headline for this study

Starting from: already using a transcription service

Counts only the analysis, grading and notes effort. The conservative baseline.

At least 35 hours
about US$1,500 at the low end

Up to about 150 hours and US$8,200 at the high end of the published ranges.

Alternative baseline

Starting from: no transcription today

Baseline (a) plus manual transcription. The upper bound.

At least 600 hours
about US$14,000 at the low end

Up to about 1,100 hours and US$36,000 at the high end of the published ranges.

Which baseline we lead with: Not recorded in our calls or notes how responses were handled before, so the conservative baseline (a) leads.

Task Manual time (published range) With Speak (our assumption)
Coding and theming interview responses 0.75 to 1.52 hours per hour of audio
Assessing the time required for qualitative analysis: a comparative methodological study of coding interview data in health services research		 0.5 hours per hour of audio
Our assumption: 0.5 hours per hour of interview to review AI-suggested themes.
Manual transcription (baseline b only) 4 to 7 hours per hour of audio
Realities Toolkit #08: Transcribing your own qualitative data		 0.25 hours per hour of audio
Our assumption: 0.25 hours to check an AI transcript.

Workload (measured): 150 hours of audio. Speak media records for the account (not deleted, each recording capped at 12 hours, zero-length rows ignored), pulled 2026-10-11.

Hourly cost: US$28.05 (25th percentile) to US$37.87 (median) for Market research analysts and marketing specialists; transcription is costed at US$16.42 to US$19.43. Both are multiplied by 1.43 to include benefits and employer costs. US wages are used as a neutral reference.

Hours saved = workload x (manual hours – hours with Speak). Value = hours saved x hourly wage x 1.43.

Equivalent manual effort for the recorded hours. Many responses are short, and not every one would have been coded by hand.

As of May 2026. This estimate is not customer-reported, is not audited, and does not include software costs.

What the team found, in short

Paraphrased from our own conversations with the team. Not quotes.

Agencies often choose to embed capture instead of building it, because respondents answer better inside a familiar page than in a separate tool.

Paraphrased from our calls with the team

Larger qualitative samples only work if each answer is short and structured, which is why fields matter as much as the recorder itself.

Paraphrased from our calls with the team

A branded client layer changes what the agency delivers, from a report at the end to a view of the evidence clients can open themselves.

Paraphrased from our calls with the team

What is next

Two colleagues lean over one laptop beside a bright window and discuss results together.
  • Per-study templatesTeams like this often turn a first study's recorders and fields into a template, so the next study starts from a known set-up.
  • Theme summariesChat across a folder lets analysts ask for themes and representative answers without opening each recording.
  • Client dashboardsA shared dashboard can give each client a view of its own study next to the recordings and transcripts.

More customer stories

Customer name withheld at their request. Figures are from Speak usage records, frozen as of May 2026.

See how a white-label research platform could work for your agency

Talk with us about embedded recorders, multilingual transcripts and a client library that carries your brand.

Book a DemoTry Speak AI Free