Source: https://internal-agents.com/agents/plaid-ai-annotator

# Plaid — AI Annotator

Plaid's internal labeling agent for its own model training. AI Annotator automates large-scale labeling of anonymized transaction data, with human oversight on the labeled output. Plaid reports greater than 95% human alignment at a lower cost and time than manual labeling.

- Approach type: Task agent
- Deployment stage: Scaled
- Autonomy: Drafts reviewed
- Evidence strength: Detailed primary
- Status: Internal
- First reported year: 2025
- Work: Data
- Invocation: Background
- Entry reviewed: 2026-09-09

## Where people stay involved

- **raw transactions → labeled training data** — Work product review · Level 3

## Overview

### Summary

Plaid's internal labeling agent for its own model training. AI Annotator automates large-scale labeling of anonymized transaction data, with human oversight on the labeled output. Plaid reports greater than 95% human alignment at a lower cost and time than manual labeling.

Fact · Reported · High confidence · `plaid-ai-annotator--summary`

Confidence reason: A linked first-party source states the claim.

Evidence:

- Supports · [1] [AI agents at Plaid (June 2025)](https://plaid.com/blog/ai-agents-june-2025/) · [Preserved copy](https://github.com/steel-experiments/internal-agents-map/blob/main/archive/sources/plaid-ai-annotator-source-1/content.md)

## Supervision evidence

### Operating model assessment

Level 3 for raw transactions → labeled training data; human attention boundary: work-product-review.

Inference · Catalog judgment · Medium confidence · `plaid-ai-annotator--operating-models-0`

Confidence reason: The reported greater than 95% human alignment implies that people review the labeled output.

Qualifications:

- Observation date: 2025-06

Evidence:

- Supports · [1] [AI agents at Plaid (June 2025)](https://plaid.com/blog/ai-agents-june-2025/) · [Preserved copy](https://github.com/steel-experiments/internal-agents-map/blob/main/archive/sources/plaid-ai-annotator-source-1/content.md)

## Implementation details

### Knowledge

Anonymized Plaid transaction data

Fact · Reported · High confidence · `plaid-ai-annotator--architecture-knowledge`

Confidence reason: A linked first-party source states the claim.

Evidence:

- Supports · [1] [AI agents at Plaid (June 2025)](https://plaid.com/blog/ai-agents-june-2025/) · [Preserved copy](https://github.com/steel-experiments/internal-agents-map/blob/main/archive/sources/plaid-ai-annotator-source-1/content.md)

## Reported metrics

### Headline claim

Greater than 95% human alignment at lower cost and time than manual labeling

Metric · Reported · Medium confidence · `plaid-ai-annotator--headline-metric`

Confidence reason: Plaid reported the figure in its own blog without independent verification.

Qualifications:

- Reported by: Plaid
- Scope: Transaction labels generated by AI Annotator in early use
- Denominator: Labels compared with human judgments; sample size unspecified
- Observation date: 2025-06

Evidence:

- Supports · [1] [AI agents at Plaid (June 2025)](https://plaid.com/blog/ai-agents-june-2025/) · [Preserved copy](https://github.com/steel-experiments/internal-agents-map/blob/main/archive/sources/plaid-ai-annotator-source-1/content.md) · Preserved content.md, lines 16–20

### Key observation

Greater than 95% human alignment with labeled data

Metric · Reported · Medium confidence · `plaid-ai-annotator--key-metrics-0`

Confidence reason: Plaid reported the figure in its own blog.

Qualifications:

- Reported by: Plaid
- Scope: Transaction labels generated by AI Annotator in early use
- Denominator: Labels compared with human judgments; sample size unspecified

Evidence:

- Supports · [1] [AI agents at Plaid (June 2025)](https://plaid.com/blog/ai-agents-june-2025/) · [Preserved copy](https://github.com/steel-experiments/internal-agents-map/blob/main/archive/sources/plaid-ai-annotator-source-1/content.md) · Preserved content.md, lines 16–20

### Key observation

Lower cost and time than manual labeling

Metric · Reported · Medium confidence · `plaid-ai-annotator--key-metrics-1`

Confidence reason: Qualitative company comparison; neither actual costs nor elapsed-time measurements are supplied.

Qualifications:

- The source does not report the denominator of this figure.
- Reported by: Plaid
- Scope: AI transaction annotation cost and time relative to manual labeling

Evidence:

- Supports · [1] [AI agents at Plaid (June 2025)](https://plaid.com/blog/ai-agents-june-2025/) · [Preserved copy](https://github.com/steel-experiments/internal-agents-map/blob/main/archive/sources/plaid-ai-annotator-source-1/content.md) · Preserved content.md, lines 20

## Sources

1. [AI agents at Plaid (June 2025)](https://plaid.com/blog/ai-agents-june-2025/)
   - Engineering blog · First party · Evidence
   - Original URL: <https://plaid.com/blog/ai-agents-june-2025/>
   - Published: 2025-06 · Accessed: 2026-08-13 · Last verified: 2026-08-31
   - Preserved copy in the repository: <https://github.com/steel-experiments/internal-agents-map/blob/main/archive/sources/plaid-ai-annotator-source-1/content.md>