By Daniel Marcous
Artificial intelligence is evolving rapidly, and 2025 is poised to be a transformative year. For investors, the opportunity lies in looking beyond buzzwords and focusing on companies that deliver practical, scalable solutions to real-world problems.
As the industry shifts toward specialization, modularity and trust-building, here are five key factors to consider when evaluating AI investments.
RAG is reshaping scalability and cost efficiency
RAG, or retrieval-augmented generation, is emerging as a game-changer in AI. By allowing systems to access external, real-time databases for domain-specific knowledge, RAG eliminates the need for costly, ongoing fine-tuning of models. Instead of static AI that can become outdated, RAG enables applications to dynamically adapt to changing conditions — a major advantage for industries like finance, healthcare and legal services, where up-to-date information is critical.
For investors, the appeal of RAG lies in its scalability and cost efficiency. When evaluating companies, look for those with access to proprietary, high-quality datasets or strong partnerships with live data providers.
Also, examine whether the company has mechanisms for ensuring the accuracy of retrieved information, as industries like healthcare and financial compliance require high levels of reliability. Businesses leveraging RAG to create adaptive, industry-specific solutions will likely lead the next wave of AI innovation, offering a robust return on investment.
Composable AI: Adaptability through modularity
AI systems built with modular, interchangeable components — known as composable AI — are driving a new era of adaptability and efficiency. These architectures allow companies to iterate quickly, customize their solutions and reduce overhead.
Similarly, pre-assembled “plug-and-play” AI kits are democratizing access to AI by enabling businesses to deploy domain-specific tools without needing extensive technical expertise.
Investors should prioritize companies that focus on modularity as a way to serve underserved markets and adapt to industry-specific needs. Are they offering scalable architectures that let users easily integrate new capabilities?
Do their solutions address clear pain points in industries like finance or healthcare? Modular systems and plug-and-play tools are especially attractive because they lower barriers to entry, unlocking new revenue streams for companies while reducing customer friction.
Domain-specific AI as the new standard
The failure of general-purpose AI in 2024 underscored a clear lesson: specialization beats generalization. Businesses and consumers alike are seeking AI tools designed to excel at specific tasks rather than attempting to be all things to all users. Domain-specific AI, whether tailored for healthcare diagnostics, financial modeling or personalized education, offers greater precision and trust by focusing on solving well-defined problems.
For investors, the first question to ask is whether a company has a deep understanding of its target industry. Are its solutions developed with input from subject matter experts? Does the AI product align with clear customer pain points?
Domain-specific AI often benefits from higher adoption rates because it integrates seamlessly into workflows and delivers measurable value. Startups creating focused, purpose-built AI copilots are well-positioned to dominate their niches and drive significant returns.
Collaborative intelligence builds trust and drives adoption
AI is no longer about replacing humans — it’s about augmenting human capabilities. Collaborative intelligence, which combines AI’s scalability and efficiency with human judgment, is proving to be the most effective model across industries such as healthcare, finance and scientific research. These systems foster trust by positioning AI as a tool that enhances human decision-making rather than replacing it.
Investors should evaluate whether companies are building systems designed to work in tandem with human users. Do their platforms include robust feedback loops and intuitive interfaces? Are they positioning their solutions to solve problems that require human oversight, such as creative processes or ethical decision-making? Collaborative intelligence isn’t just a buzzword — it’s a framework for building practical, scalable AI solutions that end-users trust and adopt.
Risk management and regulatory adaptation
As AI adoption accelerates, its risks are becoming clearer. Liability for AI-generated errors, misinformation and evolving regulatory frameworks are now key concerns for businesses and investors alike. Companies that address these challenges head-on, building resilience into their systems, are more likely to succeed long-term.
For investors, it’s critical to assess how a company is mitigating these risks. Are they incorporating validation layers to ensure accuracy and reliability? Do they proactively engage with regulators to stay ahead of compliance requirements? Businesses that treat risk management and regulatory adaptation as core competencies, rather than afterthoughts, will have a competitive edge. Resilient systems not only inspire trust but also reduce costs associated with errors, lawsuits and compliance failures.
The future belongs to practical, purpose-driven AI
In 2025, the AI sector is shifting away from hype-driven automation and toward real-world impact. Investors should focus on companies that prioritize adaptability, collaboration and domain-specific expertise. The most successful businesses won’t be chasing buzzwords — they’ll be solving meaningful problems, building trust and creating scalable solutions that align with user needs.
The future of AI is about alignment, not disruption. Companies that integrate seamlessly into workflows, address specific industry challenges, and proactively manage risks will deliver the most enduring value. For investors, the opportunity lies in backing these practical, purpose-driven innovators.
Daniel Marcous, the co-founder and CTO of april, brings his rich experience in tech and data science to solve tax and save people time and money. After spending seven years at Google as a data wizard, managing tech projects and teams as well as forming AI strategy for Google Cloud, he became data science lead and CTO at Waze. Prior to that, Marcous spent several years in the Israeli Defense Forces, founding and leading its first big data and AI team (now an entire division), leading cyber intelligence programs using ML, and founding a center of excellence for tech innovation at the IDF.
Illustration: Dom Guzman
Stay up to date with recent funding rounds, acquisitions, and more with the
Crunchbase Daily.
