> ## Documentation Index
> Fetch the complete documentation index at: https://dialnexa.com/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# Stannp

> Connect DialNexa calls to Stannp for model run, extraction result, generated answer, transcript, classification, media analysis, or tool call workflows.

Stannp provides a direct mail API enabling users to send postcards and letters programmatically.

<Note>
  Use Stannp with DialNexa when the call needs classification, extraction, generation, retrieval, summarization, visual understanding, or another AI-assisted step.
</Note>

## Where Stannp fits in a DialNexa workflow

Stannp should receive DialNexa output when the conversation affects a model run, extraction result, generated answer, transcript, classification, media analysis, or tool call. The handoff should explain what the caller asked for, what DialNexa learned, which record or object is affected, and who owns the next step.

<CardGroup cols={2}>
  <Card title="Classify call intent" icon="check-circle">
    Turn messy speech into intent, category, urgency, language, sentiment, and confidence fields.
  </Card>

  <Card title="Generate useful drafts" icon="check-circle">
    Create replies, notes, briefs, summaries, or tasks from the exact call outcome and approved context.
  </Card>

  <Card title="Extract structured details" icon="check-circle">
    Pull names, dates, products, addresses, invoice numbers, image labels, or requested actions into fields.
  </Card>

  <Card title="Keep humans in control" icon="check-circle">
    Send low-confidence, sensitive, or account-changing outputs to review instead of acting silently.
  </Card>
</CardGroup>

## What DialNexa should capture for Stannp

* Transcript, summary, language, speaker role, media or file link, intent, confidence, and sensitive-data flag
* Knowledge source, prompt version, model ID, tool call, allowed action, and fallback path
* Generated draft, extracted fields, recommended next step, review reason, and owner
* Transcript link, recording link, DialNexa call ID, CRM link, ticket link, and output record URL
* Risk flags for hallucination, missing source, private data, low confidence, or restricted action

## High-value Stannp workflows

<AccordionGroup>
  <Accordion title="Classify support reason after calls">
    For this workflow, DialNexa should send Stannp a concise, action-ready handoff: matched caller, affected record, reason for the update, urgency, owner, next step, and links to call evidence.
  </Accordion>

  <Accordion title="Draft a CRM note or support reply">
    For this workflow, DialNexa should send Stannp a concise, action-ready handoff: matched caller, affected record, reason for the update, urgency, owner, next step, and links to call evidence.
  </Accordion>

  <Accordion title="Extract appointment or address details">
    DialNexa should capture the preferred time, timezone, owner, promise made, and contact channel before updating Stannp. The receiving team should see exactly why the follow-up exists and what the caller expects next.
  </Accordion>

  <Accordion title="Summarize a long escalation">
    For this scenario, DialNexa should treat Stannp as an escalation destination. Send the impact, urgency, affected customer or object, owner, and transcript link so the right team can act before the issue gets colder.
  </Accordion>

  <Accordion title="Answer from approved knowledge">
    For this workflow, DialNexa should send Stannp a concise, action-ready handoff: matched caller, affected record, reason for the update, urgency, owner, next step, and links to call evidence.
  </Accordion>

  <Accordion title="Use create campaign">
    Use create campaign only when DialNexa has a matched caller, a clear destination object, and enough call context to justify opening a new AI workflow result. If the caller is unclear, route to review instead of creating noise.
  </Accordion>

  <Accordion title="Use upload file">
    Use upload file when the call depends on a document, image, transcript, receipt, contract, recording, or attachment. Include why the file matters and where it should live.
  </Accordion>
</AccordionGroup>

## Workflows that pair Stannp with other integrations

* [Stannp](/integrations/stannp) + [HubSpot](/integrations/hubspot): HubSpot for CRM notes and tasks.
* [Stannp](/integrations/stannp) + [Zendesk](/integrations/zendesk): Zendesk for support replies and ticket summaries.
* [Stannp](/integrations/stannp) + [Slack](/integrations/slack): Slack for review of risky outputs.
* [Stannp](/integrations/stannp) + [Google Docs](/integrations/googledocs): Google Docs for long-form call briefs.
* [Stannp](/integrations/stannp) + [Notion](/integrations/notion): Notion for knowledge updates.
* [Stannp](/integrations/stannp) + [Google Sheets](/integrations/googlesheets): Google Sheets for extraction QA.

## Implementation notes

* Use the DialNexa call ID as the idempotency key before running Stannp actions.
* Write a short operational summary into Stannp and link to the full transcript or recording for audit.
* Map required fields before launch: destination object, owner, status, urgency, next step, and record URL.
* Create review paths for low-confidence matches, sensitive requests, high-value customers, and actions that change money, access, legal terms, or customer commitments.

## FAQs

<AccordionGroup>
  <Accordion title="How do we reduce hallucinations?">
    Use approved knowledge sources, cite source records, set fallback behavior, and route low-confidence answers to humans.
  </Accordion>

  <Accordion title="Can DialNexa write follow-up messages with AI?">
    Yes, but drafts should use the actual call outcome, approved tone, and customer context. Sensitive messages should be reviewed.
  </Accordion>

  <Accordion title="What data should not be sent to AI tools?">
    Secrets, payment data, private HR or health details, and anything your policy forbids unless the tool and workflow are approved.
  </Accordion>

  <Accordion title="How should extraction errors be handled?">
    Store confidence and route missing or conflicting fields to review rather than silently updating downstream systems.
  </Accordion>

  <Accordion title="When should an AI answer fall back to a human?">
    When source context is missing, confidence is low, the caller disputes the answer, or the next step changes money, access, or legal commitments.
  </Accordion>

  <Accordion title="What should be measured over time?">
    Intent accuracy, extraction accuracy, fallback rate, review overrides, bad answers, and customer outcomes after AI-generated follow-up.
  </Accordion>

  <Accordion title="Should AI outputs act without review?">
    Only for low-risk workflows with clear confidence thresholds. Account changes, money, access, legal, and sensitive support cases need review.
  </Accordion>

  <Accordion title="What should be logged for AI decisions?">
    Prompt version, source context, model or workflow ID, confidence, output, call ID, and reviewer when applicable.
  </Accordion>
</AccordionGroup>