AttenTech
Language & Intelligence

Q&A

Answer questions over your own governed records in one call.

Get API AccessPrivate beta only
// OVERVIEW

How Q&A works

Ask a plain-language question and get back a grounded answer pulled directly from your organization's own documents and records — not a generic model guess. Every answer is traceable to its source passage and scored for confidence, so downstream teams know exactly how much to trust it.

Grounded, source-cited answers

Every answer is generated only from your governed record set and returned with the exact source document and passage it came from — no hallucinated facts, no unattributed claims.

Confidence scoring

Each answer ships with a calibrated confidence score, so your workflow can auto-route low-confidence responses to human review instead of surfacing them silently.

Works over records you already have

Indexes existing document stores and databases as-is — no need to restructure your records or maintain a separate knowledge base before asking your first question.

qa
attentechth.com/api/qa/ask
01// ask a question over your own records
02import { QA } from '@attentech/sdk'
03
04const qa = new QA({
05 region: 'th-bangkok',
06 pdpa: true,
07})
08
09const result = await qa.ask({
10 corpus: 'hr_policies',
11 question: 'How many annual leave days do I get?',
12})
13
14if (result.confidence > 0.8) {
15 const { answer, source } = result
16 reply(answer)
17}
// PROCESS

How Q&A answers questions from your own governed records

Every question triggers a search over your own indexed records for the most relevant passages, and each candidate answer is checked against its source before it ships — so what comes back is grounded and traceable, not a generic model guess.

?
ANSWER
SOURCE
CONFIDENCE
Ask
Retrieve
Score
Structured Output

Ready to integrate Q&A?

Get API access in days, or bring us in to run the full workflow for you — sovereignty by design at every step.

Get API AccessPrivate beta only