Seven services, one delivery pipeline
Collection, generative AI data, annotation, transcription, translation, evaluation and managed workforce. The same task engine, the same QC tiers and the same reporting behind all of them.
01
AI data collection
The data your model has never seen, recorded by real people in the conditions your product ships into.
What we take on
- Speech collection
- Image collection
- Video collection
- Text and prompts
- Sensor and device
- Demographic quotas
Industries that use it
- Automotive
- Healthcare
- Banking and finance
- Retail and e-commerce
- Telecom
- Real estate
- Safety and surveillance
- Consumer devices
How it is collected
- 1Recruited against your quota
- 2Captured on real devices
- 3Checked at the moment of capture
02
Generative AI services
Alignment data with a written reason behind every judgement, so you can audit why a response won.
What we take on
- Prompt and response
- Preference ranking
- Chain of thought
- Rubric rating
- Multi-turn dialogue
- Red teaming
Industries that use it
- Technology and software
- Banking and finance
- Healthcare
- Retail and e-commerce
- Telecom
- Education
How it is collected
- 1Written by qualified people
- 2Briefed with a taxonomy
- 3Rated against a versioned rubric
03
Data annotation
Labels scored by how correct they are, not by how fast they were drawn.
What we take on
- 2D bounding boxes
- Polygon and mask
- Keypoint and pose
- 3D cuboids and LiDAR
- OCR and document
- Text labelling
Industries that use it
- Automotive
- Retail and e-commerce
- Healthcare
- Banking and finance
- Safety and surveillance
- Real estate
- AR and VR
- Manufacturing
How it is collected
- 1You send the raw data, or we collect it
- 2Guidelines with worked examples
- 3Pre-labelling where it helps
04
Data transcription
Transcription in the transcriber's own language, checked against the script before a human sees it.
What we take on
- Verbatim transcription
- Clean read
- Speaker diarisation
- Timestamping
- Phonetic transcription
- Domain transcription
Industries that use it
- Telecom
- Banking and finance
- Healthcare
- Automotive
- BPO and call centres
- Media and entertainment
How it is collected
- 1Audio from you or from our collection
- 2Machine pre-fill, human correction
- 3Style guide per project
05
Model evaluation
An evaluation you can defend in a review, with the disagreement made visible.
What we take on
- Search relevance
- Assistant evaluation
- ASR benchmarking
- Bias and fairness
- Output ranking
- Error taxonomy
Industries that use it
- Technology and software
- Banking and finance
- Healthcare
- Telecom
- Automotive
- Retail and e-commerce
How it is collected
- 1Test sets built from your use case
- 2Raters qualified on your rubric
- 3Blind comparisons
06
Crowd-as-a-service
Our people, your platform, your guidelines, with the quality machinery still attached.
What we take on
- Dedicated teams
- Language pods
- Expert pools
- On-platform work
- Managed payments
- Fraud controls
Industries that use it
- Technology and software
- Banking and finance
- Healthcare
- Automotive
- Retail and e-commerce
- Telecom
How it is collected
- 1Sourced from the community and vendors
- 2Verified before paid work
- 3Managed day to day
07
Translation and localisation
Language data that reads like it was written there, in every language pair your product ships in.
What we take on
- Parallel corpora
- Document translation
- Localisation
- Multilingual prompt data
- Translation evaluation
- Rare and low-resource languages
Industries that use it
- Technology and software
- Banking and finance
- Healthcare
- Retail and e-commerce
- Telecom
- Media and entertainment
How it is collected
- 1Source text from you, or written for you
- 2Translated by native professionals
- 3Terminology controlled
How every project runs
The same six steps behind all seven services, so quality and reporting look the same whichever you choose.
- 01
Scoping call
Volume, languages, quotas, formats and what counts as a rejected item.
- 02
Guidelines and gold set
Written guidelines with worked examples and a test every contributor passes.
- 03
Pilot batch
A small paid batch you accept before anything scales.
- 04
Production
Capture checks, automated QC, peer review and expert review.
- 05
Delivery
Incremental batches with a data card, QA report and agreement scores.
- 06
Iteration
Rejections feed back into the guidelines.
Not sure which of these you need?
Describe the failure you are trying to fix. We will tell you which service applies, and say so if the answer is none of them.
Common questions
If the answer you need is not here, the contact form reaches the delivery team rather than a sequence.
How quickly can a project start?
A pilot on an existing language and task type usually starts within a week. New languages, specialist domains or KYC-gated workforces take two to four weeks to recruit and qualify.
What do you need from us to quote?
Modality, volume, languages, any demographic or device constraints, your deadline, and ideally an example of what the model gets wrong today. Guidelines help but are not required to start.
Do you run a pilot?
On every project. A small paid batch against your real guidelines, which you accept before anything scales. It has saved every client who took it.
What is the minimum engagement?
There isn't a hard floor, but below roughly a week of production work the setup cost dominates and you are better served by a catalogue dataset.