Add a classification step
Call a published model from your application and use its label in your own workflow.
STAYCHARTED AMT / FOR SOFTWARE COMPANIES
Add customer-specific classification to your software
Use StayCharted AMT to train and measure category models, then call them from your application through the API.
FROM YOUR EXAMPLES
Customer message: “I need to invite a new teammate.”Account administrationIllustrative category · Not a live prediction
Where this fits in your work
Support customer taxonomies
Train separate models from the categories and examples each customer approves.
Reduce sprawling keyword rules
Compare learned categories with your current rules before deciding what to replace.
PREPARE YOUR EXAMPLES
Start with a file your team has already labeled.
Send a record to a published model and receive its category and confidence. Use the answer to help route a ticket, classify a product, or suggest a label in your application. Your software controls the user experience and decides when a result needs review.
Use one row per example with an approved answer. Keep the input representative of what the model will receive later, and agree on how to label ambiguous cases before training.
| Input text | Customer-approved label |
|---|---|
| I need to invite a teammate. | Account administration |
| Why was I charged twice? | Billing |
| Can the export include custom fields? | Feature request |
Illustrative rows and categories. Use your own approved labels.
From examples to reviewed results
Prepare
Resolve inconsistent labels and remove unnecessary sensitive information.
Train
Choose a model on your plan and train on representative examples.
Validate
Inspect held-out accuracy by category and the mistakes that remain.
Use and improve
Publish, fill a file or call the API, then review and feed corrections into the next version.
KEEP REVIEW IN THE WORKFLOW
Know where the model stops.
Keep deterministic controls where exact behavior is required. Review uncertain results, validate each customer’s model, and protect API keys on a trusted server. Your application owns retries, permissions, and how predictions are applied.
Evidence and implementation
Customer-specific AI
Different customers may define the same category differently. Build models from their approved examples in the appropriate workspaces, keeping permissions and API keys scoped to that workspace. Measure each model against its own validation examples before relying on it.
Replace category rules where examples work better
A model can learn recurring category decisions that have become difficult to express as keyword rules. Keep deterministic rules where the outcome must be exact. Compare a model’s mistakes with the rules it would replace and provide a review path for uncertain results.
Integrate for the volume you need
Use classify for a single flat answer or predict for batches within the documented limits. The API shares the workspace’s prediction allowance and rate limits. For bulk work, use file fills in the app. Keep API keys on a trusted server rather than exposing them in browser code.
Maintain the model alongside your application
Track label changes, review errors, and retrain with representative examples. Publish a new version after checking its validation report. An API connection is a building block for your integration, not a ready-made connector to every application.
File fills return category suggestions for your existing import process. API calls return results to your own software; your integration decides how to apply them. Plan allowances and workspace rate limits still apply.
Common questions
Is this an embedded UI or a white-label application?
This workflow uses AMT to train models and the API to call them. Your team builds and operates the application experience and integration.
Can each customer have different labels?
Yes. Use the appropriate workspace and representative examples for each customer’s taxonomy. Validate each model separately.
How should we handle bulk jobs?
Design around the API’s workspace rate limits. For large files, use the app’s file-fill workflow; see the developer guide for request limits and error handling.
Explore the details
Try it on your own examples.
Start with Free for text classification. Compare plans for AI models, Vision, and API access.