Prepared by

Yev Beg | Co-Founder & CEO

AI MVP Deployment Program for
Mirai Capital Global

Date

February 3rd, 2026

Project Objective

This project outlines the design and deployment of two production-grade AI MVPs intended to be implemented directly within operating portfolio companies.

The objective is not conceptual validation, but real-world deployment, usage, and performance evaluation inside live business environments.

The MVPs are designed to operate as standalone systems that can be expanded, refined, or replicated across additional companies following successful deployment. This approach allows for practical validation of both the technology and Vexia’s execution capability under real operating conditions.

MVP 1: Outbound Sales Conversational Voice AI

1. Purpose

Deploy an outbound voice AI capable of conducting natural, two-way sales or qualification conversations with external prospects.

The objective is to validate depth of conversation and domain understanding, not just call automation.

2. Core Functionality

  • Initiates outbound calls

  • Conducts conversational sales or qualification flows

  • Handles follow-up questions and objections

  • Adapts responses dynamically based on prospect input

  • Routes or escalates when appropriate

The benchmark for this trial is comparison to a trained SDR on first-touch outbound conversations.

3. Configuration Inputs

  • Product or service knowledge

  • Target audience and call objective

  • Approved messaging and boundaries

  • Region-specific calling logic

4. Compliance Considerations

  • Regional outbound calling and AI disclosure requirements must be respected

  • The leads must have opted in to prior and cannot be completely cold as per outbound AI laws in majority of regions.

  • Deployment can be limited by geography where needed

Compliance is treated as a deployment constraint, not a blocker.

5. Deployment Model

Lead list provided by the company

  • Controlled pilot volume

  • Defined success criteria (e.g., engagement, qualification, bookings)

  • Performance and call data reviewed for validation

  • Lead list context and background

6. Validation Focus

  • Ability to sustain natural conversations

  • Quality of responses under non-scripted conditions

  • Objection handling depth

  • Consistency across calls

  • Practical usefulness versus human SDRs

MVP 2: Internal Conversational Operations AI

  1. Purpose

Deploy a conversational AI assistant trained on company-specific operational knowledge to reduce internal dependency on managers for routine questions, onboarding, and task execution.

This system acts as a first-line internal reference, accessible via chat or voice, for employees across roles.

  1. Core Functionality

  • Conversational Q&A based on internal documentation

  • Context-aware responses based on role, department, or task

  • Step-by-step guidance for operational workflows

  • Support for onboarding, daily execution, and edge cases

  • Available 24/7 without manager involvement

This is not a static FAQ.

The system is designed to reason across company-specific knowledge and guide employees in real time.

3. Data Inputs

  • SOPs and internal process documentation

  • Training and onboarding materials

  • Policies, compliance rules, and internal guidelines

  • Product or service documentation

  • Historical internal Q&A where available

The system is trained on company-specific operational knowledge, not generic information.

4. Deployment Model

  • One company or department for pilot

  • Scoped knowledge base provided by the company

  • Usage and interaction data tracked for validation

5. Operational Impact (Validation Focus)

  • Reduction in repetitive manager questions

  • Faster onboarding and ramp-up time

  • More consistent task execution

  • Centralized institutional knowledge

  • Lower operational friction across teams

Questions / Confirmations

To proceed efficiently, we’ll need clarity on the following:

  1. Which portfolio company will each MVP be deployed in first?

  2. Who will be the primary point of contact on their side for each MVP

  3. Any internal constraints we should account for early (IT, legal, data access)

Once these are aligned, we can move directly into build and deployment.

Timeline for both MVPs

Desing Phase

Development Phase

Testing and Deployment

Week 1

Weeks 2-3

Week 4

Week 1 - 2

Week 1

Week 1 - 2

Design Phase

Design Phase

Design Phase

Week 2 - 3

Weeks 2-3

Week 2-3

Development Phase

Development Phase

Development Phase

Week 4

Week 4

Week 4

QA & Testing

Testing & Deployment

Testing and Deployment

Next steps

1

Confirm Scope & Milestones

2

Schedule kickoff date

3

Technical Planning

4

Development and Technical Execution

This overview serves as a framework for our collaboration. Please review and confirm your agreement to proceed.

We are looking forward to being your next partner.

Best regards,

Yev Beg
Yev Beg
Yev Beg

Co-Founder & CEO