Genesys Cloud - Developer Community!

#Integrations

------------------------------
Marcelo Breda
SOLUTION ARCHITECT SR
------------------------------

Hi Marcelo,

One question-are these 20,000 records already stored in a CRM or another backend system that can be queried by ClientID? If so, a real-time lookup using a Data Action (or a Function if additional processing is required) could avoid the need to maintain multiple Data Tables.

------------------------------
Phaneendra
Technical Solutions Consultant
------------------------------

------------------------------
Marcelo Breda
SOLUTION ARCHITECT SR
------------------------------

Original Message:
Sent: 07-21-2026 09:44
From: Phaneendra Avatapalli
Subject: External Contact

Hi Marcelo,

One question-are these 20,000 records already stored in a CRM or another backend system that can be queried by ClientID? If so, a real-time lookup using a Data Action (or a Function if additional processing is required) could avoid the need to maintain multiple Data Tables.

------------------------------
Phaneendra
Technical Solutions Consultant
------------------------------

Hi Marcelo,

If the records are in external DB, then you may use DB2Rest method along with Data Action in Genesys Cloud CX, so you can query your records in customer data.

Thanks

------------------------------
Serkan Kaya
Communication Systems Engineer
------------------------------

Thanks Serkan. Unfortunately, since I need to perform this lookup on 100% of inbound calls, the consumption of Data Actions becomes a concern due to rate limiting. We're trying to avoid to use data actions.
I appreciate your help and suggestion.

------------------------------
Marcelo Breda
SOLUTION ARCHITECT SR
------------------------------

Original Message:
Sent: 07-21-2026 14:17
From: Serkan Kaya
Subject: External Contact

Hi Marcelo,

If the records are in external DB, then you may use DB2Rest method along with Data Action in Genesys Cloud CX, so you can query your records in customer data.

Thanks

------------------------------
Serkan Kaya
Communication Systems Engineer
------------------------------

Marcelo,

We have a similar scenario in that we want our Agents to be aware of attributes of the customers before the conversation starts.  WE do this within a iFrame and do a data pull using rest APIs to the backend systems.   

It sounds like you could avoid the Data Table limitations, for your situation, by doing real time REST APIS, either through a data action or a separate iFrame application that would notify the agent/manager/whomever of a potential for complaints based on historical trends.  If you're able to attach this to the customers phone number, then you could automate this within a flow I believe.  For us we have customers enter a unique identifier, thus our system triggers off of that.

Hope this helps.

Graham

------------------------------
Graham Pearce
Senior Technology Specialist, Client Care Technologies
------------------------------

Hi Graham, thank you for your response. Two points here:

------------------------------
Marcelo Breda
SOLUTION ARCHITECT SR
------------------------------

Original Message:
Sent: 07-22-2026 08:09
From: Graham Pearce
Subject: External Contact

Marcelo,

We have a similar scenario in that we want our Agents to be aware of attributes of the customers before the conversation starts.  WE do this within a iFrame and do a data pull using rest APIs to the backend systems.   

It sounds like you could avoid the Data Table limitations, for your situation, by doing real time REST APIS, either through a data action or a separate iFrame application that would notify the agent/manager/whomever of a potential for complaints based on historical trends.  If you're able to attach this to the customers phone number, then you could automate this within a flow I believe.  For us we have customers enter a unique identifier, thus our system triggers off of that.

Hope this helps.

Graham

------------------------------
Graham Pearce
Senior Technology Specialist, Client Care Technologies
------------------------------

Marcelo,

My apologies for not considering call volumes, at a max of 2500 requests a minute our company hasn't hit that limitation yet.  However further investigation taught me about the 50 concurrent and I can see that being an issue.

I'd be very interested in how you solve this going forward.

Graham

------------------------------
Graham Pearce
Technology Manager, Client Care Technologies
------------------------------

Original Message:
Sent: 07-22-2026 08:54
From: Marcelo Breda
Subject: External Contact

Hi Graham, thank you for your response. Two points here:

1. Since I need to perform this lookup on 100% of inbound calls, the consumption of Data Actions becomes a concern due to rate limiting.

2. Looking up customers based on ANI is not 100% reliable, because customers do not necessarily call the IVR from the phone number registered in our system. That is why we use the clientId as the primary identifier.

That said, I appreciate your help and suggestions.

------------------------------
Marcelo Breda
SOLUTION ARCHITECT SR
------------------------------


Original Message:
Sent: 07-22-2026 08:09
From: Graham Pearce
Subject: External Contact

Marcelo,

We have a similar scenario in that we want our Agents to be aware of attributes of the customers before the conversation starts.  WE do this within a iFrame and do a data pull using rest APIs to the backend systems.   

It sounds like you could avoid the Data Table limitations, for your situation, by doing real time REST APIS, either through a data action or a separate iFrame application that would notify the agent/manager/whomever of a potential for complaints based on historical trends.  If you're able to attach this to the customers phone number, then you could automate this within a flow I believe.  For us we have customers enter a unique identifier, thus our system triggers off of that.

Hope this helps.

Graham

------------------------------
Graham Pearce
Senior Technology Specialist, Client Care Technologies
------------------------------

------------------------------
Marcelo Breda
SOLUTION ARCHITECT SR
------------------------------

Original Message:
Sent: 07-22-2026 09:28
From: Graham Pearce
Subject: External Contact

Marcelo,

My apologies for not considering call volumes, at a max of 2500 requests a minute our company hasn't hit that limitation yet.  However further investigation taught me about the 50 concurrent and I can see that being an issue.

I'd be very interested in how you solve this going forward.

Graham

------------------------------
Graham Pearce
Technology Manager, Client Care Technologies
------------------------------


Original Message:
Sent: 07-22-2026 08:54
From: Marcelo Breda
Subject: External Contact

Hi Graham, thank you for your response. Two points here:

1. Since I need to perform this lookup on 100% of inbound calls, the consumption of Data Actions becomes a concern due to rate limiting.

2. Looking up customers based on ANI is not 100% reliable, because customers do not necessarily call the IVR from the phone number registered in our system. That is why we use the clientId as the primary identifier.

That said, I appreciate your help and suggestions.

------------------------------
Marcelo Breda
SOLUTION ARCHITECT SR
------------------------------



Original Message:
Sent: 07-22-2026 08:09
From: Graham Pearce
Subject: External Contact

Marcelo,

We have a similar scenario in that we want our Agents to be aware of attributes of the customers before the conversation starts.  WE do this within a iFrame and do a data pull using rest APIs to the backend systems.   

It sounds like you could avoid the Data Table limitations, for your situation, by doing real time REST APIS, either through a data action or a separate iFrame application that would notify the agent/manager/whomever of a potential for complaints based on historical trends.  If you're able to attach this to the customers phone number, then you could automate this within a flow I believe.  For us we have customers enter a unique identifier, thus our system triggers off of that.

Hope this helps.

Graham

------------------------------
Graham Pearce
Senior Technology Specialist, Client Care Technologies
------------------------------

Hi Marcelo,

As you mentioned, the simplest option is to contact Genesys Cloud Customer Care and request an increase to the Data Table row limit.

The current limit is 10,000 rows per table, and Genesys allows customers to request a higher limit.

If a limit of at least 20,000 rows is approved, you can keep a single Data Table using "ClientID" as the key and perform the lookup directly from Architect.

Official limits:

https://developer.genesys.cloud/organization/organization/limits#architect-datatables

Another option is to use External Contacts (Create a repository of customer data ).

First, you would need to create the contacts in an organized way using:

POST /api/v2/externalcontacts/contacts

During the interaction, search for the contact using (You do not necessarily need to use all these APIs to search for a contact. You can also use the native features Get External Contact action):

GET /api/v2/externalcontacts/contacts?q={searchTerm}

If the contact is found, retain the "contactId". If no contact is found, create a new one using:

POST /api/v2/externalcontacts/contacts

If necessary, retrieve the complete contact details using:

GET /api/v2/externalcontacts/contacts/{contactId}

Then retrieve the conversation participants and communication IDs using:

GET /api/v2/conversations/{conversationId}

Finally, associate the external contact with the conversation using:

PUT /api/v2/externalcontacts/conversations/{conversationId}

For this specific scenario, increasing the Data Table limit would probably be the simplest solution. External Contacts would require more API calls and additional logic to create, search, and associate the contact with the conversation.

------------------------------
Luiz Rosa
Full stack developer
------------------------------

Olá, Marcelo.

Para esse cenário, eu avaliaria utilizar o External Contacts como uma camada de armazenamento operacional das informações dos clientes, em vez de distribuir os aproximadamente 20 mil registros entre várias Data Tables.

Em fluxos de chamadas inbound, o Genesys Cloud disponibiliza a variável nativa:

Call.ExternalContactId

Quando a interação já está associada a um External Contact, essa variável contém o identificador interno do contato. Nesse caso, não vejo necessidade de criar uma Data Action customizada apenas para pesquisar o registro.

O fluxo poderia utilizar diretamente a ação nativa Get External Contact, passando:

Call.ExternalContactId

A lógica seria semelhante a:

Chamada inbound
      ↓
Call.ExternalContactId está preenchido?
      ├── Sim
      │    ↓
      │ Get External Contact
      │    ↓
      │ Informações necessárias estão preenchidas?
      │    ├── Sim → continuar o atendimento
      │    └── Não → consultar a API do cliente
      │
      └── Não
           ↓
       Consultar a API do cliente pelo ClientID
           ↓
       Utilizar as informações retornadas
           ↓
       Criar ou atualizar o External Contact

Na implementação que utilizo atualmente, o comportamento é parecido:

1. O fluxo verifica se existe um External Contact associado à interação.
2. Quando existe, recupera as informações já armazenadas.
3. Se os campos necessários estiverem preenchidos, o fluxo continua sem consultar o backend do cliente.
4. Se o contato não existir ou estiver sem as informações necessárias, o fluxo consulta a API do cliente.
5. A resposta da API pode ser utilizada para criar ou enriquecer o External Contact, permitindo que interações futuras utilizem as informações já armazenadas.

Esse modelo reduz a dependência de consultas em tempo real ao sistema do cliente e utiliza o External Contacts como uma espécie de cache operacional.

Um ponto importante é que Call.ExternalContactId depende de a interação já ter sido associada a um contato.

Como você comentou que o ANI nem sempre é confiável, podem existir chamadas em que a variável esteja vazia ou NOT_SET. Nesse cenário, o fallback por ClientID ainda seria necessário.

Mesmo assim, isso não significa que todas as chamadas precisariam consultar o backend. A consulta seria realizada apenas quando:

• O External Contact não estivesse identificado.
• O contato ainda não tivesse sido criado.
• As informações necessárias não estivessem preenchidas.
• Os dados estivessem vencidos ou precisassem ser recalculados.

Para os dados adicionais, como propensão de reclamação ou classificação do cliente, poderiam ser utilizados custom fields do External Contact.

Atualização das informações

No meu caso atual, não tenho necessidade de atualizar constantemente os dados já armazenados.

Entretanto, se as informações precisassem permanecer sincronizadas com o sistema de origem, vejo duas alternativas principais.

A primeira, e que considero mais eficiente, seria o sistema do cliente enviar um webhook para o backend de integração quando alguma informação fosse modificada:

Alteração no sistema do cliente
          ↓
Webhook para o backend
          ↓
Atualização do External Contact

Dessa forma, somente os registros alterados seriam atualizados.

A segunda alternativa seria utilizar uma trigger programada para iniciar um Workflow ou chamar um backend responsável por consultar os registros modificados desde a última sincronização:

Scheduled Trigger
       ↓
Workflow ou backend
       ↓
Consulta por registros modificados
       ↓
Atualização dos External Contacts

Eu evitaria varrer os 20 mil registros a cada execução. O ideal seria utilizar algum campo como:

updatedAt
lastModifiedDate
lastSyncDate

para executar uma sincronização incremental.

Volume e limites

Se o volume de chamadas simultâneas ou de atualizações fosse muito alto, eu também recomendaria abrir um ticket com o suporte da Genesys.

O objetivo do ticket seria solicitar uma avaliação de capacidade e dos limites aplicáveis ao caso de uso, apresentando informações como:

• Chamadas por minuto no horário de pico.
• Quantidade de consultas por interação.
• Número esperado de atualizações.
• Concorrência máxima.
• Crescimento projetado.
• Endpoints utilizados.

A partir desses dados, a Genesys poderia avaliar se os limites atuais atendem ao cenário ou se existe alguma possibilidade de ajuste.

Minha proposta arquitetural seria:

External Contacts
    = cache operacional no Genesys Cloud

API do cliente
    = fonte oficial das informações

Call.ExternalContactId + Get External Contact
    = consulta nativa dentro do fluxo

Webhook ou sincronização incremental
    = atualização dos dados

API em tempo real
    = fallback

Esse modelo não elimina completamente a API do cliente, mas evita chamá-la em todas as interações.

A consulta ao backend acontece apenas quando o Genesys Cloud ainda não possui um contato identificado ou quando os dados armazenados não são suficientes.

Você já avaliou utilizar Call.ExternalContactId com a ação nativa Get External Contact dessa forma?

Hi Marcelo,

For this scenario, I would consider using External Contacts as an operational storage layer for customer information instead of distributing approximately 20,000 records across multiple Data Tables.

In inbound call flows, Genesys Cloud provides the following built-in variable:

Call.ExternalContactId

When the interaction is already associated with an External Contact, this variable contains the contact's internal identifier. In that situation, I do not see a need to create a custom Data Action only to search for the record.

The flow could directly use the native Get External Contact action and pass:

Call.ExternalContactId

The logic could be structured as follows:

Inbound call
      ↓
Is Call.ExternalContactId populated?
      ├── Yes
      │    ↓
      │ Get External Contact
      │    ↓
      │ Are the required fields populated?
      │    ├── Yes → continue the customer journey
      │    └── No → query the customer API
      │
      └── No
           ↓
       Query the customer API using ClientID
           ↓
       Use the returned information
           ↓
       Create or update the External Contact

The implementation I currently use follows a similar pattern:

1. The flow verifies whether an External Contact is associated with the interaction.
2. When a contact exists, the flow retrieves the information already stored in Genesys Cloud.
3. If the required fields are populated, the flow continues without querying the customer's backend.
4. If the contact does not exist or does not contain the required information, the flow calls the customer API.
5. The API response can be used to create or enrich the External Contact so that future interactions can use the stored information.

This approach reduces the dependency on real-time calls to the customer's system and uses External Contacts as an operational cache.

One important consideration is that Call.ExternalContactId depends on the interaction already being associated with a contact.

Since you mentioned that ANI is not always reliable, there may be calls where this variable is empty or NOT_SET. In that situation, the fallback lookup using ClientID would still be required.

However, this does not mean that every interaction must query the backend. The API call would only be necessary when:

• The External Contact was not identified.
• The contact has not yet been created.
• The required information is missing.
• The stored data has expired or must be recalculated.

Additional information, such as complaint propensity or customer classification, could be stored in External Contact custom fields.

Keeping the information updated

In my current implementation, I do not need to continuously refresh information that has already been stored.

However, if the information had to remain synchronized with the source system, I would consider two main approaches.

The first, and in my opinion the most efficient, would be for the customer's system to send a webhook to the integration backend whenever a record changes:

Change in the customer system
          ↓
Webhook to the backend
          ↓
Update the External Contact

This way, only changed records would be updated.

The second option would be to use a scheduled trigger to start a Workflow or call a backend service responsible for retrieving records modified since the last synchronization:

Scheduled Trigger
       ↓
Workflow or backend
       ↓
Retrieve modified records
       ↓
Update External Contacts

I would avoid scanning all 20,000 records during every execution. An incremental synchronization should preferably use a field such as:

updatedAt
lastModifiedDate
lastSyncDate

Volume and platform limits

If the expected volume of concurrent calls or contact updates were high, I would also recommend opening a support ticket with Genesys.

The objective would be to request a capacity and rate-limit assessment for the use case, including information such as:

• Peak calls per minute.
• Number of lookups per interaction.
• Expected number of updates.
• Maximum concurrency.
• Projected growth.
• APIs and endpoints involved.

Based on these details, Genesys could evaluate whether the existing limits support the scenario or whether any adjustment is possible.

My proposed architecture would be:

External Contacts
    = operational cache in Genesys Cloud

Customer API
    = source of truth

Call.ExternalContactId + Get External Contact
    = native lookup inside Architect

Webhook or incremental synchronization
    = data-update mechanism

Real-time API lookup
    = fallback

This model does not completely eliminate the customer API, but it avoids calling it during every interaction.

The backend is queried only when Genesys Cloud does not yet have an identified contact or when the stored information is insufficient.

Have you considered using Call.ExternalContactId together with the native Get External Contact action in this way?

------------------------------
Matheus Mendonca
------------------------------