In Part 1 of this series, we looked at how app entrepreneurs can best leverage AI to create more software revenue, improve customer experiences, and build products in an AI-native way.
Part 2 of this series examined how artificial intelligence applications also show promise in building the new infrastructure stack for restructuring data, agent-driven automation, and evaluation.
Now, concluding this series inspired by a report from Bain Capital Ventures on The Top Generative AI Opportunities in AI and apps and infrastructure, we look at the road ahead for the types of companies most likely to succeed in the years ahead.
QUICK TAKES
AI could add between $2.6 and $4.4 trillion of economic value annually across 63 examples of uses analyzed, ranging from banking to life sciences, according to a June 2023 McKinsey estimate.
The AI market is already worth $136.55 billion, reports Grand View Research, which forecasts a compound annual growth rate of 37.3% from 2023 to 2030.
94% of business leaders believe AI will be important for their success in the next five year, according to Deloitte.
As the world has seen during the past few months, predictions about AI tend to fall flat when the circumstances change nearly every day. One moment, ChatGPT has 100 million users; the next, it’s criticized for its shortcomings. LLMs are simultaneously praised for saving time and for causing extra work. Still, as the BCV report authors point out, it’s possible to pick the types of enterprises that show signs of prevailing through their use of generative AI.
First are foundation model providers. Experts see companies such as OpenAI and Anthropic directly partnering with enterprises and selling through “hyper-scalers” such as Azure/Open AI, Oracle/Cohere and Google/Bard and consulting firm partnerships. “They will provide foundation models (Claude 5 or GPT-7) with enterprise functionality (running in the customer’s VPC), and create an ecosystem of prompt engineers around themselves (much as Salesforce has done with app administrators and consultants),” write the BCV authors. “Open-source models will continue to proliferate throughout the enterprise, offering Databricks/MosaicML, Together and others an opportunity to capture the training and serving stack.” By competing on specific “ergonomic qualities” that help them outperform in specific situations, foundation model providers will direct their focus away from sheer size.
Next are domain specific models, or full stack applications that have trained their own models. Because large foundation models can’t easily absorb proprietary information and also because of their emergent capabilities, there’s a need for new code generation models to infer cloud architecture and code performance characteristics by taking advantage of emergent capabilities. “These are emergent because the model could be trained only on Git data, not architecture schematics or code profiling charts,” according to the BCV report. “The trade-off is that the model gives up creativity—you can’t get code profiling summaries the way Shakespeare would have delivered them, but the model would still produce preferable results from a developer’s perspective.”
Finally, platforms for building autonomous agents are also promising because of the need for such integrated components as data extractors, data storage and workflow engines. By acting like an infinite army of researchers, these agents can be trained on a company’s knowledge and can be kept current through retrieval systems to automate niche, company-specific workflows, concludes the BCV report.
From improving customer experiences and building AI-native products to building new infrastructure stacks and broadening model capabilities, generative AI offers myriad opportunities for building companies to entrepreneurs and investors ready to do their due diligence.