Certeza joins NVIDIA Inception to advance AI-powered debt collection

A milestone with a practical purpose

Certeza is now a member of NVIDIA Inception, NVIDIA's program for startups building artificial intelligence solutions. For our team, this is more than a milestone to announce. It is an opportunity to strengthen the technology behind a complete, measurable debt collection operation designed to scale across Latin America.

We are joining the program at an important stage in Certeza's development. Our goal is to take artificial intelligence beyond an isolated conversation and apply it to a tangible business problem: turning outstanding receivables into cash while preserving the context of every obligation, the controls approved by each customer, and a reliable record of what happened.

Revenue can remain trapped after the sale

Many companies have already completed the hard part. They acquired the customer and delivered a product or service. Collection, however, may still depend on disconnected spreadsheets, calls without continuity, conversations spread across channels, and payment receipts that must be reviewed by hand.

When receivables data is separated from contact history, payment commitments, and bank movements, teams spend valuable time reconstructing a case before they can act. It also becomes harder to understand which approach worked, which promise is about to expire, or which payment belongs to a particular invoice, obligation, resident, or unit.

That is why Certeza was not designed as another chatbot. It was designed as a specialized operating layer that organizes data, coordinates actions, and preserves evidence throughout the recovery process.

An end-to-end collection operation

The platform connects stages that are often managed in separate tools:

1. Portfolio integration and assessment: receive and organize data from accounting systems, ERPs, CRMs, and operational files.
2. Prioritization and segmentation: define treatments based on age, obligation type, risk, history, and approved policies.
3. Coordinated outreach: execute follow-up through available channels with shared context and continuity between interactions.
4. Promises and arrangements: record commitments, dates, amounts, and outcomes so follow-up does not become fragmented.
5. Receipts and payments: help receive supporting documents, associate them with the relevant obligations, and keep case status current.
6. Reconciliation and disputes: support payment matching and route discrepancies that require validation or human intervention.
7. Traceability and reporting: turn each action into useful information for monitoring recovery, productivity, and receivables decisions.

This approach is reflected in our business debt collection solution. Artificial intelligence operates inside a real workflow rather than as a demonstration disconnected from portfolio data and actual recovery.

Specialized agents with context, controls, and voice

A collection agent needs more than the ability to produce a convincing answer. It must identify the right obligation, understand the state of the case, follow contact rules, record the outcome, and recognize when an exception should be escalated.

At Certeza, we are building specialized agents that share context with the rest of the operation and act within verifiable boundaries. Our evaluations consider conversational naturalness, factual accuracy, continuity across channels, and the correct separation between communicating, recommending an action, and changing sensitive operational information.

Voice is an important part of that evolution. A useful voice interaction is not only one that sounds natural. It must respond at the right time, retain context, avoid unauthorized claims, and leave a record that the team can review. We therefore treat quality as an operational result, not merely as linguistic fluency.

What NVIDIA Inception enables for Certeza

NVIDIA describes Inception as a free program that supports AI startups from prototype to production. Its member benefits include developer resources and tools, technical training, ecosystem offers, and opportunities to connect with a global innovation community, subject to the conditions that apply to each member.

For Certeza, these resources can support four product priorities:

• More rigorous evaluation: expand testing for accuracy, consistency, safety, and correct resolution before new capabilities are scaled.
• Greater efficiency: explore ways to reduce latency and cost per useful interaction without compromising agent quality.
• More natural, controlled voice: continue improving turn-taking, context retention, response times, and escalation to people.
• Regional scalability: prepare configurable infrastructure that can adapt to the channels, language, and regulatory requirements of different Latin American markets.

Membership does not change our responsibility for the product. Certeza remains responsible for its architecture decisions, safety controls, and the way artificial intelligence is applied to collection operations.

What comes next

Our next chapter remains focused on measurable outcomes: integrating receivables data, helping teams prioritize, coordinating outreach, recording promises, reconciling payments, and understanding which actions produce recovery.

Joining NVIDIA Inception expands the resources available to pursue that vision. The real measure of progress will be what customers experience: more consistent agents, more efficient operations, stronger traceability, and a platform capable of supporting companies with recurring receivables as they grow.

If your organization needs to organize and scale its collection process, contact Certeza for an assessment. We would like to understand how your receivables operate today and show how disconnected information can become a coordinated recovery workflow.