Investment experience of AI projects: 6 suggestions for early entrepreneurs in the AI track

When the original intention of product design can meet the rigid needs of the market, after being empowered by AI, the value of the product will be more easily recognized by the market and obtain considerable cash flow in the early stage. Take AI+credit management as an example. In the US market, credit scores will affect the loan interest rate for a person to buy a house or a car. Imagine that you only need to pay a few dozen dollars in annual fees for an AI+credit improvement APP product, and it is possible that the annual interest rate for buying a house or a car will be reduced from 9% to 7% or 6% due to the improvement of credit scores. For consumers, this is a huge attraction. When the product can meet the rigid needs of users, even if the product is not market-leading in AI technology, the functions it implements and the problems it solves are very likely to make its cash flow far exceed some AI tools with unclear goals. Of course, this type of market also faces risks: competitors may enter the market later with the same in-depth understanding of the business as early entrants, and with strong market capabilities and stronger technical capabilities. Therefore, it is extremely important for rigid demand products or entrepreneurial teams that do not have strong technical backgrounds to build a moat for themselves in the field through early user cultivation.

Compared with this type of rigid demand products, I have seen another type of entrepreneurial team that comes in with technology, and has a very strong determination and motivation to change the world with technology. However, during further communication, it seems that the team has not yet begun to explore which market segment they hope to solve. This is what I am more worried about from an investor's perspective. We do not rule out that technical experts can attract PMs or business partners who have a deep understanding of users and the market at a certain point. For technology-driven entrepreneurial teams, of course, having a strong technical background is a strong advantage. However, if the team cannot prove that its products can truly resonate with users, then this technical advantage may be quickly diluted by the market.

In the field of money, everyone competes with each other in a game similar to a Min-Max algorithm game, and the strategies of the players are very strong. There are many big and small players. It is very important to have some data models in the business side to verify the market acceptance of its products while having strong technology.

To sum up:

(1) Original rigid demand + AI tools = The original ability to obtain cash flow is amplified

(2) AI tools are difficult for target users to understand and solve, which means difficult cash flow

(3) Strong technical strength = strong potential for cash flow (but whether there is enough spark after the product and market collide needs to be verified)

My suggestions:

If the team itself has a strong technical background, do not ignore the market feedback.One feasible approach is to form an early data estimation model for user conversion rate.Through early customer acquisition data, we collect data such as natural traffic, conversion rate, payment ratio, retention rate after payment, etc. of our products to form an initial model for predicting financial changes. Entrepreneurship itself is a competition of comprehensive strength. Sometimes I hear the founder of a startup team say to me, "We have strong technical strength and the market does not need to be verified." At this time, I think the founding team may need to rethink how to view the market, users, and their own market strategy (in fact, I am a little afraid of such founders, thinking that they may need some time to connect with the real market, because the final product needs to be tested in the market, and users will vote with actions and payments).

Why is it important to “design this competitive barrier” + “prove through some actions that this barrier has been initially established”?

1. Barriers provide sustainable competitive advantage:Barriers are the key to a company's continued existence in the market. If other competing companies can simply copy the business model of early entrants, it will be difficult for this early entrant startup to achieve long-term profitability. When the entrepreneurial team clearly defines the barriers of its products, the entrepreneurial team is actually building its own competitive advantage.
2. ROI guarantee:From an investor's perspective, investing in a company with clear barriers means that our investment has a greater chance of getting a return. On the contrary, investing in a company that is easily imitated by competitors means a higher risk.
3. Technical barriers, business model innovation barriers, community barriers, etc. In short, the entrepreneurial team must have at least one or two barriers to raise the threshold:Some products built on GPT are based on the same core technology (provided by GPT), so some teams sometimes confuse the watershed between GPT's capabilities and their own team's capabilities during the pitch process. Startups must answer clearly what barriers they can build on large language models, and this is a question worth thinking about for a long time.

    

Two examples:

• Comet.ml: This is a platform that helps data scientists and engineers more easily track, compare, explain, and optimize machine learning experiments. While its underlying technology can be copied by other teams, its strong community, educational content, and integrated tools create a unique market position for it.
• Hugging Face: Despite being based on open source models, Hugging Face creates value for its users by providing high-quality models, a strong community, and an easy-to-use tool library.

For both companies, their technological foundation is not their only moat, but the startups’ communities, brands, and specific product features have kept them ahead of the curve in the market.

My suggestion is to take two steps:

“Design this competitive barrier” + “Prove through some actions that this barrier has been initially established”

AI and machine learning models are directly related to the quality of the data they use. It is also very important to ensure that startups have a proper data strategy and have made corresponding plans for data collection, storage and processing.

1. Quality and quantity of data:High-quality data is key to a machine learning model performing well. If the data is biased, erroneous, or noisy, the model's output may also be inaccurate or biased. In addition, although some models can work on less data, many deep learning models require a large amount of labeled data to achieve good results. Startup teams need to think about the source of high-quality data.
2. Diversity of data:In order for the model to perform well under a variety of conditions, it needs to be trained on data from a variety of scenarios and conditions. This ensures the model's generalization ability.
3. Data Privacy and Compliance：With the implementation of data privacy regulations, such as GDPR in Europe and CCPA in California, there are strict requirements for the collection, storage and processing of data. Startups need to ensure that their data strategies comply with the corresponding applicable regulations. (I personally think that it may not be the most important in the early stage, but in the process of product scaling, data privacy is an issue that needs to be addressed.)

My suggestions:

Planning a startup’s data strategy

My personal advice to early stage teams is to build a team with multiple skills, including data scientists, engineers, business domain “experts”, and market strategy “experts”. From my observation, the earlier this step is, the better.

From an execution perspective, experts in a multidisciplinary team do not necessarily have to work full-time within the team, but the AI team definitely needs the corresponding perspectives to improve its products, such as some internal consultants or external consultants for regular communication.

In my opinion, the formula for a relatively low-risk team model after investment is as follows:

Ideal team = technical "god" + domain knowledge expert (providing domain knowledge) + product manager "god" (conversion of technology to product matrix) + marketing expert (2C or 2B)

(Of course, this is also a relatively one-sided model. If we invest in the seed round, then strong technical strength is also a very good factor, but if it is a higher amount of investment, such as the A round, the team with comprehensive strength is definitely more competitive.)

Certain disciplines need to focus on the explainability of algorithms: AI’s decision-making process should be transparent and explainable, especially when they involve critical decisions. In some security-related fields, such as finance and healthcare, if the founding team cannot explain the algorithm clearly, it may not be a good signal for potential investors.

1. Safety and trust: Safety is particularly important for technologies related to humans. In Tesla's self-driving cars or certain medical prediction decisions, it is crucial to understand why the system makes specific decisions. For AI, in certain specific areas, if the founding team cannot understand and explain the model's decisions, it is difficult for us to trust the startup team's grasp of the product, especially in some key decisions of the product.
2. Public and regulatory acceptance: Transparent and explainable AI systems are more likely to gain the trust of the public and regulators. As AI applications in healthcare, finance, and other critical areas increase, explainability will become part of regulatory requirements.
3. Fault diagnosis and optimization: When a model makes incorrect or unexpected decisions, explainability can help engineers and researchers diagnose the problem and optimize or correct the model.
4. Competitive advantage: In a competitive business environment, providing customers with an AI solution that they can understand and trust can be a significant competitive advantage, especially in the enterprise market.

For most users and customers, AI may be a relatively unfamiliar and complex field. How to effectively operate the product in practical applications may not be a very intuitive thing. Making the target customers/users think that the product is intuitive to use is one of the missions of the entrepreneurial team.Providing users with necessary guidance and training has become one of the actions that AI startups need to take. Another way is to use the design strategy of Design Thinking to lower the threshold of product use and improve product usability from the original intention of product design.

1. Improve user confidence and comfort: Understanding and becoming familiar with a new product or technology can greatly improve user confidence. Through the onboarding features designed into the corresponding product, users will feel that they have more control over the product and are more willing to use and recommend it.
2. Reduce potential errors and misunderstandings: When users don’t understand how to use a product correctly, they may make poor decisions or misunderstand it. Guiding users ensures they know how to use the product most effectively and avoid common errors and misunderstandings.
3. If it is an enterprise-level scenario, it is necessary to strengthen communication with customers and reach a corresponding consensus. We will see that some startup teams will provide products that are not very symmetrical with the role requirements of B-side customers: for example, the problems solved by the developed products are biased towards the technical level, and customers cannot understand how this technical product can be implemented in the business scenario.
4. Expand the application scope of the product: Through guidance, users may discover new uses or different application methods of the product, thereby increasing user dependence and satisfaction with the product.
5. Meeting compliance requirements: In some industries, such as healthcare and finance, the use of AI tools may require following specific training and education guidelines to ensure that end users use AI tools correctly and safely.

Some practical suggestions for action:

• Design product interfaces and processes through Design Thinking: Provide users with corresponding invisible support for using AI products, and make it easier for users to accept new AI products emotionally and in terms of usage.
• Online tutorials and courses: Provide users with clear and easy-to-understand online tutorials, from basic knowledge to advanced techniques, to help them better use AI products.
• Workshops and seminars: Organize actual events (such as AMAs) that allow users to personally experience, ask questions, and interact with the team and experts.
• Documentation and FAQ: Provide detailed product documentation and FAQs so that users can find and learn at any time.
• Community and Forum: Create a user community where users can ask questions, share experiences and best practices, or directly communicate and provide feedback to the core team. The user community can be established on Discord, WhatsApp or various platforms.

A successful entrepreneur does need to have a clear vision and strong execution, but this is still not enough to guarantee the long-term success of a company. An open mind is one of the decisive factors that can shape company culture, promote teamwork, and drive innovation.

1. Ability to adapt to change: The rapid pace of change in markets and technology means that a strategy that works today may be obsolete tomorrow. Open founders are able to adapt to these changes rather than clinging to outdated strategies or approaches.
2. Diversity and Inclusion: An open mindset fosters respect for and inclusion of diversity, which not only helps create a more positive, innovative work environment, but also attracts and retains top talent from all backgrounds.
3. Customer (user)-centric thinking: Open founders are more likely to listen to customer and user feedback and adjust products and services accordingly, which is crucial to ensuring the fit between product and market.

    

A founder with an open mind can not only better adapt to changes in the external environment, but also bring more opportunities and inspiration, thereby driving the company's continued success. On the road to entrepreneurship, openness and flexibility are more likely to bring long-term success than closedness and stubbornness.

I personally think that founders can listen to more other external voices, especially those from different domains or different perspectives, so that they can quickly enrich their awareness of the overall market.