LiveLike has released its new multi-agent AI assistant, a fully embedded intelligence layer within the LiveLike CMS. Designed for editorial, marketing, product and analytics teams, the assistant enables users to create, schedule and analyse interactive experiences – including quizzes, polls and predictions – via natural language prompts.

This system is an intelligent agent network, where each agent mirrors a distinct user persona and collaborates in real time to execute complex engagement strategies.

“This assistant goes far beyond basic prompt-to-output functionality,” said Miheer Walavalkar, CEO of LiveLike. “It’s engineered to eliminate friction between content, marketing, product and analytics teams. That means faster time to market, greater campaign precision, and the ability to launch initiatives that were previously too complex to execute.”

Historically, producing the average 30 to 40 widgets per weekend – across quizzes, polls, and campaigns – required up to 400 minutes of manual work. With the Genie AI assistant, teams can now build and deploy the same output in under an hour, while also increasing personalisation and segmentation, according to LiveLike.

Thanks to agent coordination, a single prompt from a content editor can automatically incorporate marketing timing, product publishing rules and analytics triggers, without requiring cross-team planning.

“We’ve engineered coordination as a feature, not a by product,” Walavalkar continued. “When the editor acts, the analyst, product manager and marketer are already in sync through the agents that represent them.”

The assistant is designed to serve four core user personas:

• Editors: Create and deploy polls, quizzes and other content with a single prompt
• Marketers: Launch personalised, time-sensitive campaigns aligned with key events
• Product managers: Automate publishing logic and remove manual QA cycles
• Analysts: Access live performance data and generate optimisation recommendations

This inter-agent coordination produces second- and third-order benefits: analysts identify high-performing segments, marketers plan corresponding activations, and editors produce tailored content without duplication or delay.

Currently, the AI assistant can generate and publish polls, quizzes and prediction widgets with natural language, and image-based widgets for polls and quizzes; access real-time internet data to ensure freshness and contextual relevance; analyse historical widget performance and recommend next-step actions; produce code snippets for developers embedding LiveLike functionality; and support brainstorming and campaign ideation using internal LiveLike knowledge, including compliance documents and case studies.

One of the most advanced and exciting features now available is personalised loyalty experiences powered by dynamic user groups. Genie can now identify user segments based on behaviour and engagement patterns; assign personalised quests and incentives; and deliver campaign experiences unique to each user.

Future developments include emoji sliders that let audiences express sentiment and intensity through an interactive, visual scale; badges, leaderboards and rewards that keep fans coming back for more; the ability to instantly see which content hits hardest and how different audiences engage with deeper insights; and campaign segmentation using predefined or AI-generated user groups.

Genie is also gearing up for personalised loyalty journeys, where every fan gets their own quests, challenges,and rewards based on how they interact.

The assistant is built using AWS Bedrock, ensuring enterprise-grade security, data isolation and zero model training on client data. LiveLike is actively exploring co-marketing initiatives with AWS and anticipates strategic interest in extending this tool to AWS partners through its go-to-market ecosystem.

In parallel, LiveLike has developed and launched an MCP server that will allow external large language models to trigger CMS functionality via API. This opens access to the assistant’s capabilities for partners who do not currently use the LiveLike CMS, such as broadcasters, sports apps and third-party platforms.

“We believe the future of engagement is not just automated, but coordinated. That’s what this assistant enables – whether you’re a LiveLike client or part of the broader content ecosystem,” said Walavalkar.