10,000+ daily calls with structured POS support
A leading Nigerian digital payment company used localised Dyna.Ai Voice Agent calls and API callbacks to collect merchant POS feedback, confirm reported issues, and return structured issue tags to support teams. The workflow was designed to make merchant feedback usable for technical investigation rather than leaving support teams with disconnected recordings or inconsistent notes.
The operating problem emerged as the merchant network grew. A POS issue is valuable only when the support operation can capture the problem accurately, confirm what happened, and route the information quickly enough for a technical or service team to act. Manual follow-up calls became slow and inconsistent. Different conversations produced different levels of detail, and support teams had to interpret notes before they could begin investigation.
The company needed greater outreach capacity, but scale alone was not sufficient. More calls would have created more unstructured information unless the workflow could also classify the issue and return it to the customer’s existing support process.
The challenge
Merchant POS issues were lost in manual follow-up
Dyna.Ai provided a localised Voice Agent workflow for outbound merchant conversations. The agent could ask for the merchant’s description of the POS issue, gather the information required by the support process, and confirm the issue before submission.
The voice interaction was connected to internal systems through API callbacks. Language-model processing converted the conversation into categorized issue tags, allowing support teams to receive a structured starting point for investigation. The solution therefore combined three capabilities that matter to this use case: localised voice interaction, conversational issue confirmation, and structured system handoff.
The value was not only that the AI could speak with merchants. It was that the conversation produced an operational object that support teams could use. The workflow moved the organisation from “a call took place” to “this merchant reported this issue, it was confirmed, and it belongs to this category.”
The deployment
Voice Agent turned calls into structured issue data
The workflow was designed around the customer’s existing merchant and support systems. Merchant lists could be passed into the outbound process, calls could be placed in batches, and categorized results could be returned through API callbacks. The exact API systems, data fields, authentication model, and production ownership still need to be confirmed before publication.
The customer’s support teams remained responsible for diagnosis and resolution. Dyna.Ai collected and organized the issue signal; technical teams decided what the issue meant, what response was appropriate, and when a merchant needed further assistance.
That boundary is important for payment operations. POS issues may involve hardware, connectivity, merchant behaviour, transaction state, or account context. A structured tag can reduce the work required to begin investigation, but it does not replace technical diagnosis or the support decision.
How the workflow changed
API callbacks returned support-ready issue tags
Before the workflow, support teams depended on manually completed follow-up calls and inconsistent notes. After the workflow was introduced, merchant feedback could be collected at scale, confirmed during the conversation, classified by issue type, and returned to the support operation in a consistent format.
The workflow supported more than 10,000 daily outbound calls and 100% tagging accuracy in the reported case evidence. For support teams, the business result was not simply more conversations. It was a more reliable path from merchant feedback to a categorized issue, technical investigation, and the next support action.
The results
10,000+ daily calls with 100% tagging accuracy
For a digital payment company, the business value is the reduction of ambiguity between merchant outreach and technical action. Support teams receive a consistent signal about what happened, helping them move from a conversation to an investigation without reconstructing every issue from free-text notes or recordings.
