FrenchTech Connect AI Agents Event Recap: Is Manus the Real Deal?

FrenchTech Connect AI Agents Event Recap: Is Manus the Real Deal?

On April 3rd, the MUST Grill in Shanghai buzzed with excitement as tech enthusiasts, entrepreneurs, and AI professionals gathered for FrenchTech Connect AI Agents – Is Manus the Real Deal? Hosted by System in Motion’s Founder & CEO, Stéphane Monsallier, the event delivered an insightful deep dive into the third wave of AI evolution: AI Agents.

Exploring the Eight “Brains” of AI

Stéphane kicked off the session with a compelling analogy, comparing today’s AI landscape to eight distinct “brains”—each specializing in different capabilities:

• Text models(e.g., GPT)
• Image models(e.g., DALL·E)
• Voice models(e.g., ElevenLabs)
• Video models(e.g., Kling)
• Music models(e.g., SUNO)
• AI-Based Search Models(e.g., Gemini)
• Reasonning Models(e.g., Deepseek R2)
• AI Agents(e.g., Manus)

This framework set the stage for understanding where AI Agents fit in the broader ecosystem. Unlike single-task models, AI Agents combine reasoning, tool usage, and autonomy—making them a potential third generation of AI.

Manus: Hype or Game-Changer?

The spotlight then turned to Manus, an AI Agent developed by the Monica team. Stéphane revealed its standout features:

• Autonomous computer operation (browsing, clicking, typing)
• Web navigation and task execution (e.g., data extraction, form filling)
• Handling multi-step workflows without constant human oversight.

But with great hype comes great skepticism. Stéphane addressed the elephant in the room: Is Manus truly revolutionary, or just a well-marketed tool? Benchmark scores showed promise, but he emphasized real-world limitations:

• CAPTCHA barriers (a familiar foe for automation)
• API dependency uncertainties (some integrations may fail)
• Task complexity thresholds (not all workflows are seamless yet).

Audience-Driven Demos: Manus in Action

The most engaging segment came when Stéphane handed the reins to the audience. Via Mentimeter, attendees voted for their top 3 use-case demos (with a surprise fourth added due to popular demand). The selected scenarios showcased Manus’s versatility:

• Strategy - Branding

Objective: Generate a list of twenty new brand names, Check which of these brands are available for domain registration.Ensure the brand names are ligned with brand elements, easy to pronounce.

• Sales - Predictive Analytics

Objective: Eliminate the last 3 month of sales. Build a predictive model, train the model with the truncated data, then to predict the sales the next 3 months. Analyze the accuracy of the model against the real data. Show the difference between real data and predicted data in one graph in the final report.

• Sales - Customer Journey Mapping

Objective:

With defined variables and prompts: identify customer segments, define personas, map customer journeys with success criteria and objections, draft emails addressing objections for each stage, and summarize results in a document.

Key Takeaways: ROI and the Future of AI Adoption

Stéphane’s closing remarks resonated deeply:

• “AI isn’t magic—it’s a tool.” Companies must evaluate ROI—does the efficiency gain justify the investment?
• “Today’s ‘limitation’ could be tomorrow’s update.” Staying informed is critical as AI evolves rapidly.
• “Start small, scale smart.” Pilot AI Agents for discrete tasks before enterprise-wide deployment.

For more use cases, follow our:

• YouTube Channel
• Bilibili Account

We are Here to Empower

At System in Motion, we are on a mission to empower as many knowledge workers as possible. To start or continue your GenAI journey.