The strategic power of AI

AI is not just a technological advancement. It’s a strategic imperative.

Last week, we sent an update to our portfolio of 100+ companies spread across Europe and across industries detailing our evolving thinking around artificial intelligence. We wanted to share with our founders that we remain convinced by the opportunities presented by AI.

More so, we also wanted to outline tangible considerations founders and their teams should note if they’re weighing whether or not to incorporate AI into their products or businesses.

We’re now sharing the takeaways more broadly for founders and operators outside of the Cherry Ventures portfolio.

Dear founders,

It’s been quite the week in artificial intelligence news. And, beyond the back-and-forth headlines, we at Cherry believe that fundamental AI technologies can continue to provide important breakthroughs, spanning workplace productivity to strides in healthcare.

It is also becoming clear that we have entered the next wave of AI. Over 300K+ models have been shared on HuggingFace, 50%+ of trending GitHub repositories have been about AI, and well over 8,000 AI apps have already been created. At the same time, applications are becoming more sophisticated and we are starting to see a real impact on the venture ecosystem, tech stacks, and the broader economy.

The AI sector has recently witnessed a substantial increase in funding led by an American large language model (LLMs) hype. Over the past few months, Cherry has been intensively diving into numerous AI startups, particularly those specializing in the application layer. We also engaged with our portfolio companies to discuss the integration of LLM vision technologies and traditional AI methods into their operations.

Since starting Cherry, we have teamed up with founders harnessing AI and will continue to do so — several new companies building with AI, excitingly, have yet to be announced. We also want to keep you, our broader portfolio, informed of our evolving thinking and share lessons from our network on how they’re using AI, should you consider doing so.

Takeaways for CEOs

It’s imperative to understand AI’s strategic value.

AI extends innovation. It can help your company and product gain a competitive advantage and enhance customer experiences among other qualities. The coolest aspect is its low-cost availability to do first experiments for everyone.

Here are some key, at-a-glance strategies we recommend:

• Product research and exploration: Assign at least one team member to stay on top of AI trends and explore adoption possibilities. If advancements in AI, like generative text, video, and image creation, could impact your core product, dedicate a team to research and product development.
• Assign a set budget for AI: Allocate a budget for AI research, keeping in mind its exploratory nature and the potential for uncertain outcomes.
• Internal AI application: Begin by exploring AI to automate repetitive tasks, focusing on augmentation before full automation. For most companies, AI can enhance everyday tasks such as email composition, internal memos, and content research. Appoint someone to lead task optimization experiments using AI, this can be the same person who looks into product research.

Takeaways for CTOs

We trust you to find good technical approaches to integrate AI and the exact way will be different for every company. So, we will simply provide you with some considerations when coming up with your strategy:

• Feedback loop: Once you incorporate AI into your product (and workflows) you will be dealing with non-deterministic outcomes. Be prepared to tighten your feedback loop (with automatic triggers, employees, and users) and to adjust swiftly in case the outcome is not as hoped.
• Model flexibility: Explore using different models, such as GPT-4 and Claude, to reduce dependency on a single model and increase flexibility.
• Data access and privacy: Understand the nuances of data usage for fine-tuning AI models while respecting data privacy and ensuring long-term access to a large enough, relevant data set.

Key questions to ask

Through these meetings, research, and more, we’ve also come up with six core questions you should ask yourself to determine whether or not your product needs AI.

1. Do you have access to the necessary structured data that represents the target customer segment? Low-quality data leads to low precision
2. Does the product quality increase with the number of possible decisions? If it can be solved by rule mechanisms like a decision tree, AI is not needed.
3. Do you benefit from being able to model a lot of complex decisions? In deterministic software development, you try to minimize the decisions because you have to build software for all of them. With AI you can even model recommendations for unexpected instances.
4. Does the problem change over time? If the problem is slow-moving or inherently static AI is just expensive. If a problem shifts in real time and needs to accommodate changing variables, parameters, and data responses AI is great.
5. Can the solution tolerate imperfect results? AI solutions are imperfect because they rely on probabilities. No model will be correct 100% of the time, even after years of optimization.
6. Will the solution require exponential scaling? AI capabilities are a good choice if one expects the solution to scale fast and generate exponential data.

All in all, AI is not just a technological advancement. It’s a strategic imperative. As part of the Cherry portfolio, you are uniquely positioned to harness this potential and lead in this new era.

Let’s embrace this opportunity together.

Founders first,
Team Cherry 🍒

—

Contributors: Nadja Reischel and Jasper Masemann