Quality assurance · Production

Speak cuts data labeling time by 50% and improves model accuracy by 35% with Labelbox

The problem

Speak's AI models required vast amounts of high-quality labeled data across audio, transcriptions, and video to support complex language dynamics, but managing and labeling this data at scale — previously handled through individual contractors and spreadsheets — became a significant challenge as the user base grew.

First attempt

The previous data labeling process relied on a small number of individual contractors and scattered spreadsheets, making it difficult to maintain consistency or scale data operations.

Workflow diagram · grounded in source
1
Diverse data sources collected
trigger
“working with diverse data sources, including audio clips, transcriptions, and video-based conversations”
2
Centralized annotation in Labelbox
integration
“Labelbox allows Speak to centralize everything into one place, especially with a primary focus on speech-related tasks”
3
Collaborative annotation by internal and external teams
human_review
“collaborative environment where Speak's internal team, along with external data annotators, could work together seamlessly”
4
Quality assurance and validation
validation
“Labelbox's real-time feedback loops and quality assurance tools helped maintain high labeling accuracy”
5
Automatic daily labeling loop
feedback_loop
“an automatic labeling loop that lets us evaluate our speech systems daily on production data. This loop allows us to continuously retrain our models”
6
Faster feature and language releases
output
“Speak's data labeling time was cut by nearly 50%, allowing them to release new features and languages at a faster pace”
Reported outcome

Speak's data labeling time was cut by nearly 50%, model accuracy improved by 35%, and model development accelerated 2x after adopting Labelbox, enabling faster release of new features and languages.

Reported metrics
Data labeling timenearly 50%
Model accuracy improvement35%
Model development speed2x increase in speed
Reported stack
LabelboxOpenAIGoogle Cloud Platform (GCP)
Source
https://labelbox.com/customers/speak-customer-story
Read source ↗

Frequently asked questions

What did this team achieve with this AI workflow?

Speak's data labeling time was cut by nearly 50%, model accuracy improved by 35%, and model development accelerated 2x after adopting Labelbox, enabling faster release of new features and languages.

What tools did this team use?

Labelbox, OpenAI, Google Cloud Platform (GCP).

What results were reported?

Data labeling time: nearly 50%; Model accuracy improvement: 35%; Model development speed: 2x increase in speed (source-reported, not independently verified).

What failed first in this deployment?

The previous data labeling process relied on a small number of individual contractors and scattered spreadsheets, making it difficult to maintain consistency or scale data operations.

How is this quality assurance AI workflow structured?

Diverse data sources collected → Centralized annotation in Labelbox → Collaborative annotation by internal and external teams → Quality assurance and validation → Automatic daily labeling loop → Faster feature and language releases.

Related quality assurance cases