Event Recap | How Can AI Startups Build Digital Operations?

🍟 Event Recap

Today at Linear is Linear Capital's offline speaker series. We host events in major cities from time to time, helping entrepreneurs better understand industry trends and explore opportunities at the frontier of technology.

On June 16, Linear Capital held a "Linear AI Meetup" in Beijing, featuring AI product manager hidecloud and Liu Yang, head of internet and overseas business at SensorsData, a Linear portfolio company. They shared practical guides and best practices, with discussions centered on digital operations for startups — exploring the opportunities and challenges of enterprise digitalization from different perspectives.

We've compiled course notes from both speakers, hoping to offer some useful insights.

📝 Course Notes

01

The Data-Driven Path for AI Startups: A Practical Guide and Best Practices

Speaker: hidecloud, AI Product Manager

Common data challenges for startup teams include: not knowing which metrics to track, how to collect data, how to analyze it, or what business value data can provide. In this session, hidecloud shared practical guidance and best practices for startup teams on becoming data-driven. Here are the key takeaways:

1. What metrics should startups focus on?

• Characteristics of good metrics: comparable, easy to understand, ratio-based, and actionable

• Two common metrics in the AI era: TTFT (Time to First Token) and engagement rate

• TTFT: The time between a user submitting a prompt and seeing the first response appear. It measures user experience and provides a clear direction for product optimization.

• Engagement rate: How much users interact with the product — how many prompts they input. This measures genuine user engagement depth and is a prerequisite indicator for whether product value is actually being delivered.

• Methods to improve engagement rate:

• Give your product ways to showcase AI capabilities beyond waiting for users to actively prompt the model

• For example, GPT's suggested follow-up questions feature — instead of requiring users to ask new questions, the AI automatically generates related questions below each response

2. How should startups collect data?

• Early product stage: having any number to look at is enough

• Metabase: https://www.metabase.com

• Retool: https://retool.com

• Growth stage: need more professional analytics

• Plausible: https://plausible.io

• Matomo: https://matomo.org

• Mature stage: use data platforms for deeper analysis

• Fullstory: https://www.fullstory.com

• SensorsData: https://www.sensorsdata.com

Regardless of stage, just start saving data somehow. User data is something that, once gone, never comes back. In the early days, you must find ways to preserve logs of user behavior and attributes in your product.

We recommend the open-source collection tool from SensorsData: https://github.com/sensorsdata

3. How should startups analyze data?

• Metrics without dimensions aren't worth looking at
• Metrics tell you the what; dimensions tell you the why
• Hypothesize first, then validate

4. What business value does data provide?

• For AI products, user data is the new prompt
• You don't necessarily need users to actively trigger prompts — their behaviors and attributes within the product can be trigger points
• User behavior and attributes can essentially serve as prompt enhancement

02

SensorsData Empowering Chinese Companies' Digital Operations for Global Expansion

Speaker: Liu Yang, General Manager of Internet & Overseas Business, SensorsData

SensorsData is a leading digital customer engagement software provider in China and a nationally recognized high-tech enterprise. With a highly open product architecture, flexible integration capabilities, practical digital customer engagement solutions, and a comprehensive data security and compliance framework, the company helps enterprises achieve digital customer engagement.

In this session, Liu Yang noted that digital operations are key for Chinese companies expanding overseas to seize market opportunities. He introduced how SensorsData empowers the full digital operations cycle for Chinese companies going global through its combination of philosophy, methodology, and engine. The framework covers four main modules:

1. Target锁定 and precision marketing: Before going overseas, conduct extensive market research on target audiences, value propositions, and monetization strategies.

2. Customer experience and conversion rates: Big data and marketing technology help companies better understand customer needs and expectations, enabling more personalized products and services that truly satisfy customers.

3. Data-driven decision making: In the early stages of overseas expansion, move fast and iterate — adjust business strategies and optimize decisions based on real-time market information and data feedback.

4. Security and compliance: As companies grow, they will inevitably encounter security and compliance challenges. SensorsData helps enterprises build frameworks that meet requirements across different countries and regions, enabling unified global operations with efficient business integration.

🤗 Upcoming Events

That wraps up this Linear AI Meetup. We'll continue hosting events on related themes — follow Linear Capital's WeChat official account for the latest updates.

About Linear Capital

Linear Capital is an early-stage investment firm focused on "frontier technology + industry" — specifically, frontier technologies such as data intelligence, digital new infrastructure, next-generation robotics, and new technological transformations in traditional sectors (biotech, materials, energy, etc.), applied across vertical industries to dramatically improve efficiency, solve pain points, and enable industrial upgrading. We capture outsized returns through substantial increases in industrial value. We currently manage ten funds with approximately $2 billion in total assets under management.

Our investment stage focuses primarily on leading angel through Series A rounds, with check sizes ranging from $1 million to $10 million (or RMB equivalent).

Our early investments include Horizon Robotics, Kujiale, SensorsData, Tezign, Rokid, Guandata, Agile Robots, and over 120 other startup teams. The combined valuation of Linear's portfolio companies is approximately $20 billion.

In the near term, Linear Capital is working to become the best "Data Intelligence Technology Fund." Over the long term, we aim to build the most influential "Frontier Technology Application Fund."