Case Study AI Platform

Newo.ai

Our team participated in the rapid development and scaling of the Newo.ai platform for AI-powered "Digital Employees" that automate tasks and streamline workflows across diverse industries.

Overview

Newo.ai is a groundbreaking platform that empowers businesses to create highly sophisticated AI agents, or "Digital Employees." These AI-powered assistants are designed to mimic human behavior, including emotional intelligence and consciousness, making them capable of performing various tasks such as sales, customer support, coaching, training, and consulting. Digital employees are able to communicate with humans through various channels, including voice, email, and messaging.

Businesses can either build their Digital Employees from scratch or leverage Newo.ai's expertise for a rapid, tailored solution. Essentially, Newo.ai is transforming how businesses operate by providing intelligent, adaptable, and efficient virtual workforce members.

Newo.ai addressed our team with a clear mandate: to rapidly design and implement a robust, scalable infrastructure to support the dynamic needs of Newo.ai's intelligent agents. Lately, Mad Devs has provided not only DevOps, but Backend, UI/UX Design, and QA services and helped create a new platform design.

Challenges and Solutions

DevOps scope

Project Infrastructure

Phase 1

The initial brief was clear yet demanding: we needed to architect and deploy a scalable, resilient, and observable system within an incredibly tight timeframe.
This initial stage focused on building a solid foundation with the following requirements:

Key Results

Phase 2

In essence, the goal of this phase was to transform Newo.ai from a promising prototype into a robust, scalable, and secure platform capable of handling complex AI workloads.

Infrastructure and CI/CD overhaul:

The existing CI/CD pipelines were redesigned to accommodate new applications and improve efficiency.

Data infrastructure enhancements:

Pgbouncer: Implemented to pool database connections and optimize database usage.

Scalability and Performance:

Leveraged Keda for custom metrics-based scaling, allowing applications to scale based on specific requirements.
Improved monitoring and alerting to ensure system health and performance.
Implemented strategies to reduce cloud costs while maintaining performance.

Release management and Security:

Improved release processes, including notifications and automatic note-taking.
Implemented backup solutions for critical systems like CloudSQL, ClickHouse, and qdrant.
Strengthened security practices by conducting code and cluster vulnerability scans and regularly updating system components.

Data analytics and Optimization:

Optimized Kafka and ClickHouse clusters for better performance.
Refined Kubernetes application deployment and autoscaling strategies.

Key Results

Quality Assurance

The primary request for QA engineer was to implement automated tests for the backend using Python. Current Focus - developing automated tests for customer services to strengthen the testing suite further.

Key Achievements

Backend

We began work on the Newo.ai platform, focusing on designing and implementing new features. Key objectives included developing a new API gateway, enhancing system integration, and improving platform performance. Details are safeguarded under NDA restrictions.

Newo.ai design

Initially, the Newo.ai system may identify communications from a single person as coming from different actors. However, once the system recognizes that these actors belong to the same user, it merges them into a unified profile. This ability to map various communication channels to a single user's historical memory and persona distinguishes the Newo.ai platform.

During the second phase, when the decision was made to update the version of Newo.ai, it became necessary to redesign the platform completely. Our lead designer and the client-side frontend developer worked.

Workflow of Newo.ai

We've designed a user-friendly interface for creating Digital Agents that enables users to map out the entire customer journey, from first interaction to final response.

Conversation screens

Specifically paid attention to conversation screens. As Digital Employees can:

Provides answers according to the corporate knowledge base.
Uses up to date information from your CRM, ERP, Help Desk.
Has long-term memory.
Multiple communication channels (Digital Employees can call customers or reach out via email, online chat, Whatsapp, and Telegram).
Allows a real person to join a bot conversation and continue to speak for the bot (Human in The Loop).

Business app

We developed a business app design. It focuses on providing insights into agent performance and business outcomes. Here are examples of specific cases:

Offers insights into agent activity, including greeting phrases, transfers, and new contacts secured.
Shows specific customer interactions, such as table bookings, food orders, and large group reservations.
Displays conversion rates and associated revenue generated.
Presents overall business performance metrics, including total revenue and time saved.
Provides a summary of dialogue with a client (text, audio, and business metrics).

All of these options are customizable according to the customer’s business needs.

Overall result

Through this collaborative effort, your team successfully played a crucial role in developing a robust, scalable, and secure platform for Newo.ai's intelligent Digital Employees. Our team continues to work closely with Newo.ai to enhance the platform's capabilities further and expand its reach across various industries.