Product Manager - Voice AI

About ClearGrid

ClearGrid is a leading AI-driven debt resolution platform operating throughout the UAE and Saudi Arabia. Its AI voice agents conduct borrower communications in Arabic and English for major banks and NBFIs, leveraging proprietary infrastructure. This is among the largest conversational voice AI implementations regionally, handling real client volume and regulated environments beyond demonstration phases.

Role Overview

We seek a Product Manager who will individually oversee the entire Voice AI product lifecycle. This high-autonomy role involves managing voice model budgeting and trade-offs with engineers as well as collaborating with commercial teams and banking clients. The role covers key areas currently without dedicated ownership: voice stack performance and economics, prompt engineering tools, and the platform's evolution for client configuration and monitoring.

Key Responsibilities

• Oversee complete voice AI product performance, including call quality, latency, transcription accuracy (focusing on Arabic and Gulf dialects), interruption management, and conversational reliability, partnering with engineering for monitoring and improvements.
• Manage voice infrastructure components such as speech-to-text, text-to-speech, model routing, telephony gateways, and dialing capacity, planning proactively to avoid production constraints.
• Control unit economics of voice calls by tracking and driving down costs per minute and per connected call while maintaining quality and making data-backed platform or vendor decisions.
• Develop and own prompt engineering tooling to streamline AI agent authoring, versioning, testing, and deployment, enhancing speed and reliability in launching and fixing agents.
• Establish production QA processes to identify and prevent live agent failures, ensuring compliance with financial regulations and rapidly addressing issues.
• Lead the progression of client-facing platform features that enable lenders to configure agents, monitor campaigns, view analytics, and access call recordings and transcripts, reducing manual client service latency.
• Keep abreast of voice AI technology trends, assess new models and features, and make decisions on supported languages and capabilities.
• Manage stakeholder engagement and product roadmap prioritization across prompt engineering, engineering, analytics, QA, compliance, and collections teams, while judiciously controlling scope.

Candidate Profile

• 4–6 years of product management experience delivering software products in collaboration with engineering teams.
• Strong expertise or experience in voice AI including voice agents, speech models, or voice-centric products preferred; equivalent depth in conversational AI, large language models, contact center, or speech technologies considered.
• Proven experience owning internal tools or workflow products designed for expert users, distinct from client-facing applications.
• Ability to build and defend financial models for product unit economics and actively manage cost efficiencies without sacrificing quality.
• Data-driven approach with capability to extract and analyze data independently.
• Technical background beneficial but hands-on experience is paramount.
• Experience developing Arabic or other non-English language products highly advantageous; fluency in English required, Arabic preferred.
• Experience working with regulated enterprise or financial services clients is an asset.
• Systems thinker comfortable with ambiguity, proactive, and hands-on.
• Effective communicator equally comfortable addressing engineering teams and senior banking stakeholders.
• Passionate about AI and emerging technologies, continually updated through personal interest.

Success Milestones

• Within 3 months: Complete mapping of voice stack and prompt engineering workflow; establish baseline metrics for deployment speed, fix times, and cost per connected call; agree on Phase 1 roadmap.
• Within 6 months: Significantly reduce manual prompt engineering effort; implement documented cost reduction plan for connected calls; launch client-facing setup and analytics to at least one lender.