Best Practices for Sample Utterances and Custom Slot Type Values


The usability of the skill directly depends on how well the sample utterances and custom slot values represent real-world language use.

Building a representative set of custom values and sample utterances is an important process and one that requires iteration. During development and testing, try using many different phrases to invoke each intent. If you can observe other users during testing, note the phrases that they speak to invoke each intent. Continually update the custom values and sample utterances file to ensure that it includes instances of your users' most common phrasings.

The following sections provide recommendations for developing your set of sample utterances.

Sample Utterance Phrasing

Given the flexibility and variation of spoken language in the real world, there will often be many different ways to express the same request. For example, to ask for a horoscope a user might say:

  • إيش هي توقعات الابراج اليوم؟
  • أعطيني توقعات برجي اليوم
  • قوليلي توقعات برجي اليوم
  • شحالها توقعات برجي اليوم؟

Or any other variations on the above forms:

  • "إيش" and "إيش هو" and "إيش هي"
  • "طلع", "قولي", and "أعطي"
  • "حقي" and "ال"

If you have requests that are full sentences, think about shortened ways that users might say them, especially when combined with the Ask/Tell starting phrases, with patterns like أبغى أسأل…عن…. For instance, if you have "كيف الطقس؟" consider also just "الطقس". This works well with patterns such as "أبغى أسأل <invocation name> *عن *الطقس."

Incorporate variants that include words such as "من فضلك" in various positions within the utterance, such as: "من فضلك أعطيني توقعات برجي" or "قوليلي توقعات برجي اليوم".

If your skill lets users "open" or "use" items in some way, utterances that use these words in combination with slots may sound very natural and are also good to include. For example, a skill that streams audio such as radio stations could support an interaction like this:

User: *اليكسا، افتحي *مشغل الراديو

اليكسا: …مرحبا بك في مشغل الراديو
User: افتحي المفضلات الكلاسيكية.

In the above example, you would need a sample utterance such as this, where stationName is a slot:

افتحي {stationName}

Providing these different phrases in your sample utterances will help improve voice recognition for the abilities you add to Alexa. It is important to include as wide a range of representative samples as you can – all the phrases that you can think of that are possible in use (though do not include samples that users will never speak). Alexa also attempts to generalize based on the samples you provide to interpret spoken phrases that differ in minor ways from the samples specified.

Given the flexibility and variation of spoken language in the real world, there will often be many different ways to express the same request. For example, to ask for a horoscope a user might say:

  • what is the horoscope
  • get me my horoscope
  • tell me the horoscope
  • how's my horoscope today

Or any other variations on the above forms:

  • "what's" and "what is"
  • "get", "tell", and "give"
  • "my" and "the"

If you have requests that are full sentences, think about shortened ways that users might say them, especially when combined with the Ask/Tell starting phrases, with patterns like Ask…to… and Ask…about…. For instance, if you have "what's the weather," consider also just "weather". This works well with patterns such as "Ask <invocation name> for weather." or "Ask <invocation name> about weather".

Incorporate variants that include words such as "please" in various positions within the utterance, such as "please give me my horoscope" or "tell me my horoscope please".

If your skill lets users "open" or "use" items in some way, utterances that use these words in combination with slots may sound very natural and are also good to include. For example, a skill that streams audio such as radio stations could support an interaction like this:

User: Alexa, Open My Radio Player

Alexa: Welcome to My Radio Player…
User: Open classical favorites.

In the above example, you would need a sample utterance such as this, where stationName is a slot:

Open {stationName}

Providing these different phrases in your sample utterances will help improve voice recognition for the abilities you add to Alexa. It is important to include as wide a range of representative samples as you can – all the phrases that you can think of that are possible in use (though do not include samples that users will never speak). Alexa also attempts to generalize based on the samples you provide to interpret spoken phrases that differ in minor ways from the samples specified.

Given the flexibility and variation of spoken language in the real world, there will often be many different ways to express the same request. For example, to ask for a horoscope a user might say:

  • quel est mon horoscope
  • donne-moi l'horoscope
  • dis-moi mon horoscope
  • c'est quoi mon horoscope aujourd'hui

Or any other variations on the above forms:

  • "c'est quoi" and "qu'est-ce que c'est"
  • "dis-moi", "je veux", "dites-moi", "trouve"
  • "mon/ma/mes" et "le/la/l'/les"

If you have requests that are full sentences, think about shortened ways that users might say them, especially when combined with the Ask/Tell starting phrases, with patterns like Demande à…quel est… and Demande à…. For instance, if you have "quelle est la météo," consider also just "la météo". This works well with patterns such as "Demande à<invocation name> la météo." or "Demande à<invocation name> quelle est la météo".

To make sure your sample utterances reach the correct intents, adhere to the following spelling rules:

  • Add a space after each apostrophe. For example: "quel est l' horoscope".
  • For verbs ending in -er, for polite or plural requests, use imperative verb forms (-ez) rather than the infinitive (-er). For example, use "continuez".

Also use the correct accentuation in the sample utterances, as shown in the following examples:

  • "a quelle heure commence shabbat a {ville}" should be accented as "à quelle heure commence shabbat à {ville}"
  • "prière du pape benedict" should be accented as "prière du pape bénédict"

Providing these different phrases in your sample utterances will help improve voice recognition for the abilities you add to Alexa. It is important to include as wide a range of representative samples as you can -– all the phrases that you can think of that are possible in use (though do not include samples that users will never speak). Alexa also attempts to generalize based on the samples you provide to interpret spoken phrases that differ in minor ways from the samples specified.

Given the flexibility and variation of spoken language in the real world, there will often be many different ways to express the same request. For example, to ask for a horoscope a user might say:

  • ich möchte mein Horoskop hören
  • gib mir mein Tageshoroskop
  • was ist mein Tageshoroskop
  • wie wird mein Tag heute sein

Or any other variations on the above forms:

  • "ich will" and "kann ich"
  • "sage mir", "zeige mir" and "erzähle mir"
  • "wie ist" and "wie lautet"

If you have requests that are full sentences, think about shortened or different ways that users might say them, especially when combined with the Frage/Sage starting phrases, with patterns like Frage…ob… and Frage…nach…. For instance, if you have "wie ist das Wetter," consider also just "dem Wetter" or "wie das Wetter ist". This works well with patterns such as "Frage <invocation name> nach dem Wetter." or "Frage <invocation name> wie das Wetter ist".

Incorporate variants that include words such as "Bitte" in various positions within the utterance, such as "bitte gib mir mein Horoskop" or "sag mir mein Horoskop bitte".

If your skill lets users "open" or "use" items in some way, utterances that use these words in combination with slots may sound very natural and are also good to include. For example, a skill that streams audio such as radio stations could support an interaction like this:

User: Alexa, öffne mein Radioplayer

Alexa: Wilkommen zu meinem Radioplayer…
User: Öffne klassische Stücke.

In the above example, you would need a sample utterance such as this, where stationName is a slot:

Öffne {stationName}

Providing these different phrases in your sample utterances will help improve voice recognition for the abilities you add to Alexa. It is important to include as wide a range of representative samples as you can -– all the phrases that you can think of that are possible in use (though do not include samples that users will never speak). Alexa also attempts to generalize based on the samples you provide to interpret spoken phrases that differ in minor ways from the samples specified.

Given the flexibility and variation of spoken language in the real world, there will often be many different ways to express the same request. Hindi speakers will use Hindi and English words interchangeably. For example, to ask for a horoscope a user might say:

  • राशिफल क्या है
  • horoscope क्या है
  • मुझे मेरा राशिफल बताओ
  • मुझे मेरा horoscope बताओ
  • मेरा आज का राशिफल कैसा है

Or any other variations on the above forms:

  • "क्या" और "क्या है"
  • "बताओ", "दिखाओ" और "दो"
  • "मेरा"

If you have requests that are full sentences, think about shortened ways that users might say them, especially when combined with the Ask/Tell starting phrases, with patterns like ….से पूछो…. and ….के बारे में पूछो. For instance, if you have "मौसम कैसा है," consider also just "मौसम". This works well with patterns such as "मौसम के लिए पूछो." or "मौसम के बारे में पूछो" or मौसम कैसा है ये पूछो.

Include variants of English and Hindi, as users might use them interchangeably. For instance, "weather कैसा है" or "how is the मौसम".

Incorporate variants that include words such as "कृपया" in various positions within the utterance, such as "कृपया मेरा राशिफल बताओ" or "मेरा राशिफल बताइये कृपया".

If your skill lets users "open" or "use" items in some way, utterances that use these words in combination with slots may sound very natural and are also good to include. For example, a skill that streams audio such as radio stations could support an interaction like this:

User: Alexa, मेरा Radio खोलो*

Alexa: मेरा Radio आपका स्वागत करता है…
User: पसंदीदा शास्त्रीय संगीत खोलो.

In the above example, you would need a sample utterance such as this, where stationName is a slot:

खोलो {stationName}

Providing these different phrases in your sample utterances will help improve voice recognition for the abilities you add to Alexa. It is important to include as wide a range of representative samples as you can – all the phrases that you can think of that are possible in use (though do not include samples that users will never speak). Alexa also attempts to generalize based on the samples you provide to interpret spoken phrases that differ in minor ways from the samples specified.

Given the flexibility and variation of spoken language in the real world, there will often be many different ways to express the same request. For example, to ask for a horoscope a user might say:

  • qual è l'oroscopo
  • qual è il mio oroscopo
  • dammi il mio oroscopo
  • dimmi l'oroscopo
  • com'è il mio oroscopo oggi
  • ….

Or any other variations on the above forms:

  • "oggi", "del giorno" and "giornaliero"
  • "dire", and "dare"
  • "il mio" and "l'"

If you have requests that are full sentences, think about shortened ways that users might say them. For instance, if you have "Com'è il tempo," consider also just "il tempo".

Providing these different phrases in your sample utterances will help improve voice recognition for the abilities you add to Alexa. It is important to include as wide a range of representative samples as you can – all the phrases that you can think of that are possible in use (though do not include samples that users will never speak). Alexa also attempts to generalize based on the samples you provide to interpret spoken phrases that differ in minor ways from the samples specified.

Given the flexibility and variation of spoken language in the real world, there will often be many different ways to express the same request. For example, to ask for a horoscope a user might say:

  • 星座の運勢はどう
  • 私の星座の運勢を知らせて
  • 星占いを聞かせて
  • 今日の私の星座の運勢はどう
  • 星座の運勢を教えて

Or any other variations on the above forms:

  • "何" and "何ですか"
  • "教えてください" and "教えてよ"
  • "私の" and "自分の"

If you have requests that are full sentences, think about shortened ways that users might say them, especially when combined with the Ask/Tell starting phrases. For instance, if you have "天気はどう", consider also just "天気". This works well with patterns such as "<invocation name>で天気を調べて".

Incorporate variants that include words such as "ください" in various formats within the utterance, such as "星座の運勢を教えてください" or "星占いを聞かせて".

If your skill lets users "開く" or "使用する" items in some way, utterances that use these words in combination with slots may sound very natural and are also good to include. For example, a skill that streams audio such as radio stations could support an interaction like this:

User: アレクサ、マイラジオプレーヤーを開いて

Alexa: マイラジオプレーヤーにようこそ…
User: クラシックのお気に入りを開いて

In the above example, you would need a sample utterance such as this, where stationName is a slot:

{stationName}を開いて

Providing these different phrases in your sample utterances will help improve voice recognition for the abilities you add to Alexa. It is important to include as wide a range of representative samples as you can – all the phrases that you can think of that are possible in use (though do not include samples that users will never speak). Alexa also attempts to generalize based on the samples you provide to interpret spoken phrases that differ in minor ways from the samples specified.

Given the flexibility and variation of spoken language in the real world, there will often be many different ways to express the same request. For example, to ask for a horoscope a user might say:

  • qual é o meu horóscopo
  • me dá o meu horóscopo
  • diga o horóscopo do dia
  • busque o meu horóscopo de hoje

Or any other variations on the above forms:

  • "dia", "hoje", "de hoje" and "diário"
  • "me dê", "me diz" and "buscar"
  • "meu/meus" and "o/os"

If you have requests that are full sentences, think about shortened ways that users might say them, especially when combined with the Perguntar/Pedir starting phrases, with patterns like Pergunte para…sobre… and Peça ao…pela…. For instance, if you have "qual é a previsão do tempo", consider also just "previsão do tempo". This works well with patterns such as "Pergunte para o <invocation name> sobre a previsão do tempo." or "Pergunte para a <invocation name> sobre a previsão do tempo."or "Peça ao <invocation name> pela previsão do tempo".

You can incorporate variants that include words such as "por favor" in various positions within the utterance, such as "por favor me dê o meu horóscopo" or "me dê o meu horóscopo por favor".

If your skill lets users "open" or "use" items in some way, utterances that use these words in combination with slots may sound very natural and are also good to include. For example, a skill that streams audio such as radio stations could support an interaction like this:

User: Alexa, Abra Minha Rádio

Alexa: Bem-vindo à Minha Rádio…
User: Abra minhas músicas clássicas favoritas.

In the above example, you would need a sample utterance such as this, where stationName is a slot:

Abra {stationName}

Providing these different phrases in your sample utterances will help improve voice recognition for the abilities you add to Alexa. It is important to include as wide a range of representative samples as you can – all the phrases that you can think of that are possible in use (though do not include samples that users will never speak). Alexa also attempts to generalize based on the samples you provide to interpret spoken phrases that differ in minor ways from the samples specified.

To make sure your sample utterances reach the correct intents, please use the correct accentuation in sample utterances, as shown in the following examples:

  • "muda a sequencia" should be accented as "muda a sequência"
  • "adiciona linguica" should be accented as "adiciona linguiça"
  • "mais informacoes" should be accented "mais informações"
  • "toca radio um" should be accented as "toca rádio um"

Given the flexibility and variation of spoken language in the real world, there will often be many different ways to express the same request. For example, to ask for a horoscope a user might say:

  • cuál es el horóscopo
  • dame mi horóscopo
  • dime el horóscopo

Or any other variations on the above forms:

  • "cuál es" y "qué"
  • "obtén", "di", y "da"
  • "obtenme", "dime", y "dame"
  • "mi/mis" y "el/la/los/las"

If you have requests that are full sentences, think about shortened ways that users might say them, especially when combined with the Ask/Tell starting phrases, with patterns like pídele a…que… and pregúntale a…sobre…. For instance, if you have "cuál es el tiempo," consider also just "el tiempo". This works well with patterns such as "Pídele a <nombre invocación> que me dé el tiempo." or "Pregúntale a <nombre de invocación> sobre el tiempo".

Providing these different phrases in your sample utterances will help improve voice recognition for the abilities you add to Alexa. It is important to include as wide a range of representative samples as you can -– all the phrases that you can think of that are possible in use (though do not include samples that users will never speak). Alexa also attempts to generalize based on the samples you provide to interpret spoken phrases that differ in minor ways from the samples specified.

To make sure that your sample utterances reach the correct intents, use the correct accentuation in sample utterances as shown in the following examples:

  • "cuanto gas tengo en {deviceName}" should be accented as "cuánto gas tengo en {deviceName}"
  • "averigue el monto de la factura" should be accented as "averigüe el monto de la factura"

Sample Utterances for Starting a Conversation

The preferred phrase for beginning an interaction with an Alexa ability is أبغى أسأل. This can be used to ask questions or state commands in a very natural way. Your sample utterances should flow naturally when combined with these phrases.

For example, users might say the following to request a horoscope:

  • "اليكسا، أبغى أسأل توقعات الابراج اليومية عن برج الجوزاء"

Note the different forms of the sample utterances that work effectively with different variations of "أبغى أسأل".

Noun utterances:

  • "توقعات برج…"
  • "الجوزاء"

Question utterances:

  • "إيش هي توقعات برج…"

Verb utterances:

  • "أعطيني توقعات برج"
  • "قوليلي توقعات برج…"

In the syntax of the sample utterances, these would be specified as follows:

توقعات برج {Sign}
{Sign}
إيش هي توقعات برج {Sign}
أعطيني توقعات برج {Sign}
قوليلي توقعات برج {Sign}
...

Utterances that begin with specific question words ("إيش", "كيف", "وين", and so on) are especially important, as these work with the " question <invocation name> أبغى أسأل" variation. The following words are considered question words that work with أبغى أسأل:

  • منو, إيش, ليش, امتى, وين, اي, كيف
  • تقدري، تقدرين ، يقدر، تقدر، تستطيع،تستطيعي، يستطيع،
  • يا, مين, ليه (some variants of the above)

When writing utterances as questions, also consider multiple forms of the question. For example:

  • "إيش هي توقعات برجي؟". ("اليكسا، أبغى أسأل توقعات الابراج اليومية إيش هي توقعات برجي")

Not all of these question words may work well for all Alexa capabilities, but you should include utterances for these words when they represent real-world language use. For the full list of phrases users can say to begin interacting with your Alexa skill, see Understand How Users Invoke Custom Skills. Include utterances that work naturally with as many of these phrases as possible.

English invocation phrase Arabic (SA) invocation phrase Arabic (SA) example English translation Sample utterances for interaction model English translation of sample utterances

use Invocation Name to Request

استخدمي Invocation Name Request

استخدمي نغمات عربية لتشغيل قائمة الاغاني المفضلة حقي

use arabic tunes toplay my favorite playlist

لتشغيل قائمة الاغاني المفضلة حقي تشغيل قائمة الاغاني المفضلة حقي

toplay my favorite playlist play my favorite playlist

use Invocation Name to Request

استخدمي Invocation Name Request

استعملي نغمات عربية لتشغيل قائمة الاغاني المفضلة حقي

use arabic tunes toplay my favorite playlist

لتشغيل قائمة الاغاني المفضلة حقي تشغيل قائمة الاغاني المفضلة حقي

toplay my favorite playlist play my favorite playlist

open Invocation Name and Request

افتحي Invocation Name Request

افتحي لعبة الارقام وابدئي لعبة جديدة

open numbers game andstart a new game

وابدئي لعبة جديدة ابدئي لعبة جديدة

andstart a new game start a new game

start Invocation Name and Request

ابدئي Invocation Name Request

ابدئي نغمات عربية وشغلي قائمة الاغاني المفضلة حقي

start arabic tunes andplay my favorite playlist

وشغلي قائمة الاغاني المفضلة حقي شغلي قائمة الاغاني المفضلة حقي

andplay my favorite playlist play my favorite playlist

The preferred phrases for beginning an interaction with an Alexa ability are ask and tell. These can be used to ask questions or state commands in a very natural way. Your sample utterances should flow naturally when combined with these phrases.

For example, users might say one of the following to request a horoscope:

  • "Alexa, ask Daily Horoscopes for the horoscope for Gemini"
  • "Alexa, ask Daily Horoscopes about Gemini"
  • "Alexa, ask Daily Horoscopes what is the horoscope for Gemini"
  • "Alexa, ask Daily Horoscopes what's the horoscope for Gemini"
  • "Alexa, ask Daily Horoscopes to give me the horoscope for Gemini"
  • "Alexa, ask Daily Horoscopes to tell me the horoscope for Gemini"

Note the different forms of the sample utterances that work effectively with different variations of "ask":

  • Noun utterances:
    • "the horoscope for…"
    • "Gemini"
  • Question utterances:
    • "what is the horoscope for…"
    • "what's the horoscope for…"
  • Verb utterances:
    • "give me the horoscope for"
    • "tell me the horoscope for…"

In the syntax of the sample utterances, these would be specified as follows:

the horoscope for {Sign}
{Sign}
what's the horoscope for {Sign}
what is the horoscope for {Sign}
give me the horoscope for {Sign}
tell me the horoscope for {Sign}
...

Utterances that begin with specific question words ("what", "how", "where", and so on) are especially important, as these work with the "ask <invocation name> question" variation. The following words are considered question words that work with ask:

  • who, what, why, when, where, which, how
  • do, did, does
  • can, could, should, would, may, must, shall
  • is, was, will, am, are, were
  • have, has, had
  • Some variants of the above, such as: what's, whose, didn't, ain't

When writing utterances as questions, also consider multiple forms of the question. For example:

  • "what is my horoscope". ("Alexa, ask Daily Horoscopes what is my horoscope")
  • "what my horoscope is" ("Alexa, ask Daily Horoscopes what my horoscope is")

Not all of these question words may work well for all Alexa capabilities, but you should include utterances for these words when they represent real-world language use. For the full list of phrases users can say to begin interacting with your Alexa skill, see Understand How Users Invoke Custom Skills. Include utterances that work naturally with as many of these phrases as possible.

The preferred phrases for beginning an interaction with an Alexa ability are demande à and dis à. These can be used to ask questions or state commands in a very natural way. Your sample utterances should flow naturally when combined with these phrases.

For example, users might say one of the following to request a horoscope from "signe astrologique" skill:

  • "Alexa, demande à signe astrologique de me donner l' horoscope des gémeaux"
  • "Alexa, demande à signe astrologique quel est l' horoscope gémeaux"
  • "Alexa, demande à signe astrologique l' horoscope du gémeau"
  • "Alexa, dis à signe astrologique gémeaux"
  • "Alexa, dis à signe astrologique de me lire l' horoscope des gémeaux"

Note the different forms of the sample utterances that work effectively with different variations of "demande à":

  • Noun utterances:
    • "l' horoscope des…"
    • "gémeaux"
  • Question utterances:
    • "quel est l' horoscope des…"
  • Verb utterances:
    • "donne-moi l' horoscope des…"
    • "écouter l' horoscope des…"

In the syntax of the sample utterances, these would be specified as follows:

l' horoscope des {Sign}
{Sign}
quel est l' horoscope du {Sign}
dis-moi l' horoscope de {Sign}
écouter l' horoscope {Sign}
me donner l' horoscope des {Sign}
...

Utterances that begin with specific question words ("quel", "quelle", "où", "de" and so on) are especially important, as these work with the "demande/dis à <invocation name> question" variation. The following words are considered additional question words that work with demande/dis à:

  • qui, que, quoi, où, pourquoi, quand, quel, quelle, quels, quelles, comment
  • si, s'
  • pouvoir et vouloir
  • avoir
  • verbes au subjonctif

When writing utterances as questions, consider multiple forms of the question. For example:

  • "puis-je avoir l' horoscope de…" ("Alexa, demande à signe astrologique puis-je avoir l' horoscope de…")
  • "tu as l' horoscope de…" ("Alexa, demande à signe astrologique si tu as l' horoscope de…")
  • "je veux que tu me donnes l' horoscope de…" ("Alexa, dis à signe astrologique que je veux que tu me donnes l' horoscope de…")

Not all of these question words may work well for all Alexa capabilities, but you should include utterances for these words when they represent real-world language use. For the full list of phrases users can say to begin interacting with your Alexa skill, see Understand How Users Invoke Custom Skills. Include utterances that work naturally with as many of these phrases as possible.

The preferred phrases for beginning an interaction with an Alexa ability are sage and frage. These can be used to ask questions or state commands in a very natural way. Your sample utterances should flow naturally when combined with these phrases.

For example, users might say one of the following to request a horoscope:

  • "Alexa, frage Tageshoroskop nach dem Horoskop für Fische"
  • "Alexa, frage Tageshoroskop nach Fische"
  • "Alexa, frage Tageshoroskop, was das Horoskop für Fische ist"
  • "Alexa, sag mir das heutige Horoskop für Fische vom Tageshoroskop"
  • "Alexa, frage Tageshoroskop, was das heutige Horoskop für Fische ist"
  • "Alexa, sage Tageshoroskop, ich möchte mein Fischhoroskop für heute"

Note the different forms of the sample utterances that work effectively with different variations of "frage":

  • Noun utterances:
    • "dem Horoskop für…"
    • "Fische"
  • Question utterances:
    • "was das Horoskop für Fische ist"
    • "was das heutige Horoskop für Fische ist"

In the syntax of the sample utterances, these would be specified as follows:

das Horoskop für {Sign} ist
{Sign}
was das Horoskop für {Sign} ist
sag mir das heutige Horoskop für {Sign}
was das heutige Horoskop für {Sign} ist
ich möchte mein {Sign}  Horoskop für heute
...

Utterances that begin with specific question words ("wo", "wie", "was", and so on) are especially important, as these work with the "frage <invocation name> question" variation. The following words are considered question words that work with frage:

  • wer, was, wieso, wann, wo, welche, wie
  • ob
  • könnte, hat, würde, sollte, müsste
  • ist

Not all of these question words may work well for all Alexa capabilities, but you should include utterances for these words when they represent real-world language use. For the full list of phrases users can say to begin interacting with your Alexa skill, see Understand How Users Invoke Custom Skills. Include utterances that work naturally with as many of these phrases as possible.

The preferred phrases for beginning an interaction with an Alexa ability are पूछो and बताओ. These can be used to ask questions or state commands in a very natural way. Your sample utterances should flow naturally when combined with these phrases.

For example, users might say one of the following to request a horoscope:

  • "Alexa, दैनिक राशिफल से मिथून राशि के horoscope के बारे में पूछें"
  • "Alexa, Gemini राशि के बारे में दैनिक राशिफल से पूछें"
  • "Alexa, दैनिक राशिफल से पूछें कि मिथुन राशि के लिए राशिफल क्या है"
  • "Alexa, दैनिक राशिफल को मिथून राशि का राशिफल पूछें"
  • "Alexa, दैनिक राशिफल को पूछकर Gemini राशि का राशिफल बताइये"
  • "Alexa, दैनिक राशिफल को पूछकर Gemini राशि का राशिफल बताओ"

Note the different forms of the sample utterances that work effectively with different variations of "ask":

  • Noun utterances:
    • "….का राशिफल…..हैं"
    • "मिथून राशि"
  • Question utterances:
    • "…..राशि का राशिफल क्या है"
  • Verb utterances:
    • "मुझे….राशि का राशिफल बताइये"
    • "मुझे….राशि का राशिफल बताओ"

In the syntax of the sample utterances, these would be specified as follows:

{Sign} का राशिफल क्या है
{Sign}
{Sign} का horoscope बताओ
{Sign} का राशिफल बताइये
{Sign} का राशिफल कहिये
...

Utterances that begin with specific question words ("क्या", "कैसे", "कहां", and so on) are especially important, as these work with the "ask <invocation name> question" variation. The following words are considered question words that work with ask:

  • कौन, क्या, क्यू, कब, कहां, कौन सा, कैसे
  • दो, दे, दिजिये, देना
  • कर सकते है, सकता है, चाहिए, होगा, हो सकता है, जरूर, करेगा, करेगी
  • है, हैं, था, करूंगा, करूंगी, थे
  • Some variants of the above, such as: नहीं था, किस का,नहीं है

When writing utterances as questions, also consider multiple forms of the question. For example:

  • "मेरा राशिफल क्या है". ("Alexa, दैनिक राशिफल से पूछो मेरा राशिफल क्या है")
  • "मेरा horoscope क्या है" ("Alexa, दैनिक राशिफल से मेरा horoscope क्या है पूछें")

Not all of these question words may work well for all Alexa capabilities, but you should include utterances for these words when they represent real-world language use. For the full list of phrases users can say to begin interacting with your Alexa skill, see Understand How Users Invoke Custom Skills. Include utterances that work naturally with as many of these phrases as possible.

The preferred phrases for beginning an interaction with an Alexa ability are chiedi a and domanda a. These can be used to ask questions or state commands in a very natural way. Your sample utterances should flow naturally when combined with these phrases.

For example, users might say one of the following to request a horoscope:

  • "Alexa, chiedi a Oroscopo del Giorno se può darmi l'oroscopo dei Gemelli"
  • "Alexa, domanda a Oroscopo del Giorno se può dirmi l'oroscopo del Toro"
  • "Alexa, chiedi a Oroscopo del Giorno l'oroscopo del Sagittario"
  • "Alexa, domanda a Oroscopo del Giorno l'oroscopo dei Pesci"
  • "…"

Note the sample utterances that work effectively with "chiedi a" and "domanda a":

  • Verb utterances:
    • "darmi l'oroscopo di…"
    • "dirmi l'oroscopo di…"

Noun phrases: phrases that can follow

"chiedi a <invocation name> …" or
"domanda a <invocation name> …"

  • "chiedi a <invocation name> il mio colore preferito"
  • "domanda a <invocation name> l'ora del mio appuntamento"

Questions: phrases that can follow

"chiedi a <invocation name> …" or "domanda a <invocation name> …"

  • "chiedi a <invocation name> dov'è la mia macchina" "domanda a <invocation name> dov'è la mia macchina"

In the syntax of the sample utterances, these would be specified as follows:

darmi l'oroscopo del {Segno}
{Segno}
darmi l'oroscopo dei {Segno}
dirmi l'oroscopo del {Segno}
dirmi l'oroscopo dei {Segno}
...

For the full list of phrases users can say to begin interacting with your Alexa skill, see Understand How Users Invoke Custom Skills. Include utterances that work naturally with as many of these phrases as possible.

The preferred phrases for beginning an interaction with an Alexa ability are "聞いて" and "教えて". These can be used to ask questions or state commands in a very natural way. Your sample utterances should flow naturally when combined with these phrases.

For example, users might say one of the following to request a horoscope:

  • "アレクサ、十二星座占いに双子座の運勢を聞いて。"
  • "アレクサ、十二星座占いに双子座の運勢について聞いて。"
  • "アレクサ、十二星座占いに双子座の運勢がどうなっているか聞いて。"
  • "アレクサ、十二星座占いに双子座の運勢がどうか聞いて。"
  • "アレクサ、十二星座占いで双子座の運勢を聞いて。"
  • "アレクサ、十二星座占いで双子座の運勢について聞いて。"

Note the different forms of the sample utterances that work effectively with different variations of "聞いて":

  • Noun utterances:
    • "…の星占い"
    • "双子座"
  • Question utterances:
    • "…の運勢はどう"
    • "…の運勢は何"
  • Verb utterances:
    • "…の運勢を教えて"
    • "…の星占いを教えて"

In the syntax of the sample utterances, these would be specified as follows:

{Sign}の星占い
{Sign}
{Sign}の運勢は何
{Sign}の運勢はどう
{Sign}の運勢を教えて
{Sign}の星占いを教えて
...

Utterances that begin with specific question words ("何", "どう", "どこ", and so on) are especially important, as these work with the "<invocation name>に質問聞いて" variation.The following words are considered question words that work with 聞いて:

  • 誰, 何, なぜ, いつ, どこ, どれ, どう
  • か, ですか
  • できる, してくれる, できますか
  • ですか, でしたか
  • して, やって, 実行して
  • Some variants of the above, such as: どうして, 誰の, しないの

When writing utterances as questions, also consider multiple forms of the question.For example:

  • "私の星座の運勢はどう". ("アレクサ、星座占いに私の星座の運勢はどうか聞いて")
  • "星座の運勢はどう" ("アレクサ、星座占いに私の星座の運勢はどうかチェックして")

Not all of these question words may work well for all Alexa capabilities, but you should include utterances for these words when they represent real-world language use. For the full list of phrases users can say to begin interacting with your Alexa skill, see Understand How Users Invoke Custom Skills. Include utterances that work naturally with as many of these phrases as possible.

The preferred phrases for beginning an interaction with an Alexa ability are different variations of perguntar para a, perguntar para o and pedir à, pedir ao. These can be used to ask questions or state commands in a very natural way. Your sample utterances should flow naturally when combined with these phrases.

For example, users might say one of the following to request a horoscope:

  • "Alexa, perguntar para o Horóscopo Diário sobre gêmeos"
  • "Alexa, perguntar para o Horóscopo Diário qual é o horóscopo de gêmeos"
  • "Alexa, pergunte para o Horóscopo Diário qual é o horóscopo de gêmeos"
  • "Alexa, pergunte pro Horóscopo Diário pelo horóscopo de gêmeos"

Note the different forms of the sample utterances that work effectively with different variations of "perguntar para a", "perguntar para o" and "pedir à", "pedir ao":

  • Noun utterances:
    • "o horóscopo de…"
    • "gêmeos"
  • Question utterances:
    • "qual é o horóscopo de…"
  • Verb utterances:
    • "me dar o horóscopo de"
    • "me dizer o horóscopo de…"

In the syntax of the sample utterances, these would be specified as follows:

o horóscopo de {Signo}
{Signo}
qual é o horóscopo de {Signo}
me dar o horóscopo de {Signo}
me dizer o horóscopo de {Signo}
...

Utterances that begin with specific question words ("o que", "como", "onde", and so on) are especially important, as these work with the "perguntar para a, perguntar para o and pedir à, pedir ao <invocation name> question" variation. The following words are considered question words that work with pedir/perguntar:

  • quem, o que, por que, onde, qual, como
  • ser, sou, é, são, será, foi, era, etc.
  • poder, posso, possa, pode, podemos, poderíamos, podia etc.
  • ter, tenho, tem, temos, teria, teríamos, tinha, etc.

When writing utterances as questions, also consider multiple forms of the question. For example:

  • "qual é meu horóscopo". ("Alexa, pergunte para o Horóscopo Diário qual é meu horóscopo")
  • "qual é o meu horóscopo" ("Alexa, pergunte para o Horóscopo Diário qual é o meu horóscopo")

Not all of these question words may work well for all Alexa capabilities, but you should include utterances for these words when they represent real-world language use. For the full list of phrases users can say to begin interacting with your Alexa skill, see Understand How Users Invoke Custom Skills. Include utterances that work naturally with as many of these phrases as possible.

The preferred phrases for beginning an interaction with an Alexa ability are pregúntale a and dile a. These can be used to ask questions or state commands in a very natural way. Your sample utterances should flow naturally when combined with these phrases.

For example, users might say one of the following to request a horoscope:

  • "Alexa, pídele a Horóscopo Diario el horóscopo de Géminis"
  • "Alexa, pídele a Horóscopo Diario que me diga el horóscopo de Géminis"
  • "Alexa, pregúntele a Horóscopo Diario sobre Géminis"
  • "Alexa, pregúntale a Horóscopo Diario si me puede decir el horóscopo de Géminis"
  • "Alexa, pídele a Horóscopo Diario que me dé el horóscopo de Géminis"
  • "Alexa, pregúntale a Horóscopo Diario cuál es el horóscopo de Géminis"

Note the different forms of the sample utterances that work effectively with different variations of "ask":

  • Noun utterances:
    • "el horóscopo de…"
    • "Géminis"
  • Question utterances:
    • "cuál es el horóscopo de…"
    • "cuál es el horóscopo para…"
  • Verb utterances:
    • "dame el horóscopo de"
    • "dime el horóscopo de…"

In the syntax of the sample utterances, these would be specified as follows:

el horóscopo de {Signo}
{Signo}
cuál es el horóscopo de {Signo}
cuál es el horóscopo para {Signo}
dame el horóscopo de {Signo}
dime el horóscopo para {Signo}
...

Utterances that begin with specific question words ("qué", "cómo", "dónde", etc.) are especially important, as these work with the "pregúntale a <invocation name> question" variation. The following words are considered question words that work with pregúntale a:

  • quién, qué, por qué, cuándo, dónde, cuál, cómo, etc.
  • puede, pudo, podría, etc.
  • es, era, fue, sera, sería, etc.
  • tiene, tuvo, tendría, tendrá, etc.

When writing utterances as questions, also consider multiple forms of the question. For example:

  • "cuál es mi horóscopo". ("Alexa, pregúntele a Horóscopo Diario cuál es mi horóscopo")
  • "qué horóscopo tengo" ("Alexa, pregúntale a Horóscopo Diario qué horóscopo tengo")

Not all of these question words may work well for all Alexa capabilities, but you should include utterances for these words when they represent real-world language use. For the full list of phrases users can say to begin interacting with your Alexa skill, see Understand How Users Invoke Custom Skills. Include utterances that work naturally with as many of these phrases as possible.

Number of Sample Utterances

For each intent, include as many variations of the phrases as you expect users to speak. For example, for the utterance "إيش هو برجي اليوم", include variations such as:

  • "ايش هو…"
  • "ايش…"
  • "قولي لي…"
  • "أعطي…"
  • "أعطيني…"
  • "طلعي…"
  • "طلعيلي…"
  • "القي…"
  • "القيلي…"

It is better to provide too many samples than to provide too few, so test different phrases and add additional phrases as needed.

For each intent, include as many variations of the phrases as you expect users to speak. For example, for the utterance "what is my horoscope", include variations such as:

  • "what is…"
  • "what's…"
  • "tell me…"
  • "give…"
  • "give me…"
  • "get…"
  • "get me…"
  • "find…"
  • "find me…"

It is better to provide too many samples than to provide too few, so test different phrases and add additional phrases as needed.

For each intent, include as many variations of the phrases as you expect users to speak. For example, for the utterance "quel est mon horoscope", include variations such as:

  • "quel est…"
  • "donne…"
  • "donne-moi…"
  • "dis…"
  • "c'est quoi…"
  • "je veux…"
  • "dis-moi quel est…"
  • "trouve…"
  • "trouve-moi…"

It is better to provide too many samples than to provide too few, so test different phrases and add additional phrases as needed.

For each intent, include as many variations of the phrases as you expect users to speak. For example, for the utterance "was ist mein Tageshoroskop", include variations such as:

  • "was ist…"
  • "gib mir.."
  • "sage…"
  • "wie lautet…"
  • "erzähle…"
  • "finde…"

It is better to provide too many samples than to provide too few, so test different phrases and add additional phrases as needed.

For each intent, include as many variations of the phrases as you expect users to speak. For example, for the utterance "what is my horoscope", include variations such as:

  • "…क्या है"
  • "मुझे बताओ…"
  • "मुझे बताइये…"
  • "दो…"
  • "मुझे दो…"
  • "मुझे दिजिये…"
  • "मुझे ढूढ़कर दो.."
  • "मुझे ढूढ़कर बताओ…"
  • "मुझे search कर के बताओ"
  • "search करो और बताओ"

It is better to provide too many samples than to provide too few, so test different phrases and add additional phrases as needed.

For each intent, include as many variations of the phrases as you expect users to speak. For example, for the utterance "qual è il mio oroscopo", include variations such as:

  • "qual è…"
  • "dimmi…"
  • "dai…"
  • "dammi…"
  • "cerca…"
  • "trovare…"

It is better to provide too many samples than to provide too few, so test different phrases and add additional phrases as needed.

For each intent, include as many variations of the phrases as you expect users to speak.For example, for the utterance "星占いを聞かせて", include variations such as:

  • "…は何"
  • "…は何ですか"
  • "…は何だろう"
  • "…を教えて"
  • "…を教えてよ"
  • "…を調べて"
  • "…を調べてください"
  • "…が知りたい"
  • "…を言って"

It is better to provide too many samples than to provide too few, so test different phrases and add additional phrases as needed.

For each intent, include as many variations of the phrases as you expect users to speak. For example, for the utterance "qual é o meu horóscopo", include variations such as:

  • "qual é…"
  • "dizer…"
  • "me fale…"
  • "me diga…"
  • "me dê…"
  • "me dá…"
  • "busca…"
  • "leia…"
  • "encontra…"

It is better to provide too many samples than to provide too few, so test different phrases and add additional phrases as needed.

For each intent, include as many variations of the phrases as you expect users to speak. For example, for the utterance "cuál es mi horóscopo", include variations such as:

  • "cuál es…"
  • "da…"
  • "dame…"
  • "dime…"
  • "dime, cuál es…"
  • "obtén…"
  • "obtenme…"

It is better to provide too many samples than to provide too few, so test different phrases and add additional phrases as needed.

Recommendations for Custom Slot Type Values

When using custom slot types, make sure the list of values provided for the slot comprehensively covers the expected input from your users. While this is relatively straightforward in the horoscopes example – in this case the set of twelve signs is small, finite and has few or no variations in terms of how people are likely to speak each sign – it can be more difficult as the complexity of input to a given slot increases. Tips to build representative coverage for more complex slot types include:

  • Start with data sources that reflect real usage. For instance, if you have a web version of your application that can use items you've logged from users via forms or search queries, use those. Otherwise look for other data sources representing likelihood of input.

    If you're building a list of ingredients for a recipe skill for instance, generate a list of all ingredients used in all recipes, and validate that against other lists you may be able to find online of common ingredients.

  • The total number of custom slot type values depends on the overall size of your interaction model, including the total number of slot types and values across all custom slots. If you need to scale back the number of values, be sure to keep the values most likely to be used.
  • If your list of custom values does not contain all the values your skill expects, provide a representative set of custom values with representative word counts. For instance, if values of one to four words are possible, use values of one to four words in your value list. But also be sure to distribute them proportionally. If a four-word value occurs in an estimated 10% of inputs, then include four-word values only in 10% of the values in your list.

Slot values are sent to your skill in written format. For example, both "fire h.d. 7" and "fire h.d. seven" would be sent to your skill as "Fire HD7". For better recognition, acronyms and other phrases involving spoken letters should either be all caps ("HD") or separated by periods and spaces ("h. d. "). Using lowercase for initialisms may lead to unreliable recognition since the spoken form may not correctly be detected. See examples.

Test your utterances, check for conflicts, and revise

As you iterate on your utterances and slot values, you may unintentionally create utterances that map to more than one intent in your model. You can use the developer console to find these utterance conflicts and then decide how you want to handle them.

Also consider using the utterance profiler and NLU evaluation tool to test your utterances and measure the accuracy of your model.


Was this page helpful?

Last updated: Nov 28, 2023