Summary: Process automation is a never ending game and companies want to level up all the time. VC backed software companies may not be the answer to deliver that process automation.
What does the current investment landscape in generative AI look like?
Most AI funding (70%) targets model development, aiming to build smarter algorithms. However, there’s a noticeable underinvestment in AI applications and operations. I think this is understandable because the winners are not clear, market sizes are uncertain, and business models may change.
How significant is the potential for AI to improve productivity?
AI promises to drastically boost productivity, potentially increasing revenue per employee to $200K. Heck, maybe $300K. Yet, funding heavily favors model development over operations. My take is that this productivity boost can be achieved in multiple cost-efficient ways and not just by selling VC-backed software. Enterprises want process automation 24.7 and better process automation solutions are being shipped too quickly. Now, the best in class solution could be software + services to do this end-to-end process automation.
A key assumption I am making is that stitching together a few AI applications can transform Marketing processes, Sales processes, HR processes, Finance processes, supply-chain processes, and Support processes but the accompanying business model to deliver these services can be non-VC-backed.
What are the current adoption rates of AI within enterprises, and what does this imply?
Despite 60% of enterprises showing interest in AI, less than 10% have implemented it. This gap highlights both integration challenges and a significant market opportunity for user-friendly AI solutions. And, these user-friendly solutions can be a combination of humans doing complex tasks and automation of automatable tasks. Ultimately you’d want to implement this technology to automate boring things in your organizations.
Can AI significantly impact professional services?
Yes, AI can transform professional services, with metrics showing a 12% increase in task completion rates, a 25% reduction in task completion time, and a 40% improvement in task quality.
These metrics were achieved by BCG professionals so there is no doubt that every services business can improve on these metrics. See, everything is getting automated but it is all use case driven and these use cases may not be fit for venture scale outcomes. Some could be 🙂
How does data scarcity impact the development of AI?
Data scarcity, especially in high-quality text, poses a significant challenge, potentially decelerating AI development by 2026. Not the only way but you do need high-quality synthetic data and learning approaches to better Large Language Models that generate amazing text, audio, video, and other cool stuff for us. Companies and investors will make money but this is all high technology development and that’s very different from solving problems of enterprises.
What should we look forward to in AI development?
The next few years according to Coatue lie in creating reasoning engines that enhance human capabilities, shifting from basic computational tasks to complex reasoning and decision-making. This is inevitable and the moment this happens, you can run complex processes very efficiently with low error rates. You will not need thousands of tools to solve these problems and even if you did, you’d want good experts aka humans to manage those for you.
My Takeaways:
- Balanced Investment Strategy – VC + Non-VC: We need to invest beyond model development to applications and operations for AI growth and real-world applications but that investment may not be venture capital, although it could be.
- Enhance Integration and Usability: Enterprises and businesses in general care about low-cost end-to-end problem-solving. They want to focus on what makes their beer taste better and not necessarily run recruiting, or payroll in-house, unless these cost-centers are actually tied to revenue-producing activities in a meaningful way. And, to bridge the gap between AI interest and adoption, the focus should be on developing integrable, user-friendly solutions that demonstrate clear business value – revenue increase, cost reduction, and risk mitigation.
- Data and how will VCs make money?: On the technology side, I think addressing data scarcity with synthetic data and advanced learning techniques will be critical to sustaining and advancing AI capabilities. I think some of these solutions might be home runs from a returns perspective but it is too early to tell.
If you want to chat about this topic or critique, or give detailed feedback, pls do reach out at rohan.manchanda28@gmail.com
Rohan Manchanda 