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

# The agent mishears names, numbers or Indian language words

> Improve transcription accuracy on Bolna Voice AI calls: keyword boosting, per-language transcriber choice, precise transcripts, and prompt-level confirmation patterns.

Mishearing is a transcription problem, not an LLM problem. Fix it on the [Languages tab](/docs/agent-setup/audio-tab) before touching the prompt.

## Work through these in order

<Steps>
  <Step title="Confirm what was actually heard">
    Open the call in [Call history](/docs/agent-setup/call-history) and read the transcript. If the transcript is already wrong, the transcriber is the problem. If the transcript is right and the agent still got it wrong, it is a prompt problem — go to [Write better prompts](/docs/prompting/introduction).
  </Step>

  <Step title="Boost the words that matter">
    Add brand names, product names and other unusual terms to the **Keywords** field in the format `word:boost_value` — for example `Bruce:100`. Keyword boosting only takes effect with **Deepgram** transcription; the field is ignored by other providers.
  </Step>

  <Step title="Match the transcriber to the language">
    Providers differ sharply by language. Sarvam is tuned for Indian languages such as Hindi, Tamil and Telugu; Deepgram is strong on low-latency English. On a multilingual agent you can set **a different transcription provider per language** — see [Per-language configuration](/docs/customizations/multilingual-languages-support#per-language-configuration).
  </Step>

  <Step title="Turn on precise transcripts">
    Enable **Generate precise transcript** on the [Engine tab](/docs/agent-setup/engine-tab) for higher-accuracy transcription. Use it when accuracy matters more than the small processing cost — see [Capture precise transcripts](/docs/customizations/capturing-precise-transcripts).
  </Step>

  <Step title="Cut the noise reaching the transcriber">
    Enable **noise cancellation** on the [Call tab](/docs/agent-setup/call-tab). Callers on speakerphone in a shop or a car are the most common source of garbled transcripts.
  </Step>
</Steps>

***

## Digits, codes and spellings

Numbers are where mishearing costs the most. Two defenses:

* **Ask the agent to confirm.** Instruct it to read digits back before acting on them — "I have 9 8 7 6, is that right?". Prompt-level confirmation catches what no transcriber setting will.
* **Give the caller a keypad option.** For account numbers, OTPs and menu choices, [collect keypad input (DTMF)](/docs/inbound/dtmf) instead of relying on speech.

***

## Code-switching

Indian callers routinely mix English words into Hindi sentences. That is normal input, not an error, and it does not by itself trigger a language switch — see [How language switching works](/docs/customizations/language-switching-behavior) for what does. If the agent switches language when it should not, the fix is on that page, not in transcription settings.

***

## Next steps

<CardGroup cols={3}>
  <Card title="Choose voice and language" icon="language" href="/docs/agent-setup/audio-tab">
    Providers, models, keywords
  </Card>

  <Card title="Set up a multilingual agent" icon="globe" href="/docs/customizations/multilingual-languages-support">
    Per-language providers
  </Card>

  <Card title="Collect keypad input" icon="keyboard" href="/docs/inbound/dtmf">
    Skip speech for digits
  </Card>
</CardGroup>


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.