Enrola Pty Ltd

Enrola operates AI-powered SMS sales agents that autonomously contact and qualify leads on behalf of businesses. Our agents handle the full conversational lifecycle — from initial outreach and qualification through to handoff — using OpenAI as the underlying large language model. We also use campaign automation to re-engage cold leads. We are continuously expanding our agent capabilities across high-consideration consumer sales verticals including education, insurance, finance, and real estate.

Yes. We identify and mitigate risks through a combination of human oversight and automated evaluation. Team leads and the CTO regularly review samples of agent conversations to assess quality, accuracy, and appropriateness. We also run structured evals against test scenarios to detect regressions or unintended behaviours before deploying model or prompt changes to production.

Frequency & method

Yes. We conduct ongoing (weekly) human-in-the-loop reviews involving the CTO, a dedicated quality engineer, our Customer Success team, and direct customer feedback. Conversation samples are regularly reviewed to assess response quality and flag issues. We also run automated evals prior to each production deployment. This multi-layered feedback loop ensures quality issues are identified and addressed quickly.

Example

Yes, partially. While there is no formal appeals process, AI engagement can be ceased at any point - by the business customer via the Enrola dashboard or by the end consumer via opt-out. Once the AI is paused, a human sales representative can pick up the conversation directly. This ensures that no lead is ever locked into an automated interaction and that human judgement can take over whenever needed.

Responsible for oversight

Yes. Yvette Quinby (CTO) is responsible for AI development, testing, and oversight, supported by a dedicated quality engineer and our Customer Success team. We maintain logs of agent conversations, model versions, and prompt changes. Customer feedback is actively collected and reviewed as part of our quality process. Configuration and deployment documentation is stored in our shared internal systems. We are building out more formal documentation practices as the team and product mature.

Methods

Methods

Other methods

Yes. Our AI agents are designed to engage all leads consistently regardless of demographic characteristics. We do not use protected characteristics (such as race, gender, or disability) as inputs to our models or decision logic. We conduct informal reviews of conversation samples to identify any patterns of differential treatment and address them promptly. As we scale, we plan to implement more structured bias detection processes.

N/A — Low-scale use. Enrola’s AI infrastructure runs on AWS, which has strong commitments to renewable energy and carbon reduction. Given our current scale, direct environmental impact is minimal. We will revisit this as we grow.

Enrola’s AI agents help businesses connect with leads who have already expressed interest in their product or service, improving conversion rates while reducing the need for disruptive cold-calling. By making sales follow-up faster and more consistent, we help small and mid-sized businesses compete with larger enterprises. As we scale, we plan to establish a formal AI governance framework, introduce an ethics advisory process, and deepen our alignment with the Australian AI Ethics Principles.