Analysis by a16z and Mercury across more than 200,000 early-stage firms found that about 60% of AI budgets go to horizontal tools, meaning general-purpose platforms such as assistants, creative software, and shared workspaces that can be used across roles and teams. The remaining 40% supports vertical tools designed for specific functions like finance, HR, or legal. Creative software accounts for the largest share of spend, underscoring how AI has shifted from a supportive layer to a core part of how teams execute their daily work.
For many founders, being AI-native doesn’t mean automating everything. It means designing work so people and systems reinforce each other. Alex Wu, Managing Partner at CFO Advisors, a firm that provides fractional CFO services to startups, told PYMNTS that structure is what determines AI success.
“We encourage startups to create a dedicated role that’s all about testing new tools and figuring out how to implement them, an AI enablement role that scouts tools, runs pilots, and hardwires wins into workflows,” he explained. He added that this structured approach helps avoid what he calls “tool fatigue,” where teams experiment without measurable outcomes. “The real value shows up after the novelty wears off. If a tool is still being used after two weeks and has become part of a workflow, that’s when it’s creating real productivity.”
Across industries, AI-driven efficiency is starting to register. Thomson Reuters found that professionals expect to save an average of five hours a week through AI in 2025, up from four hours in 2024. Those gains may appear modest individually, but they compound across teams, signaling that AI adoption is moving from experimentation to measurable output. Simon Wallace, U.K.-based startup founder, said to PYMNTS that his team applies a clear return test to every tool.
“We are somewhat ruthless when it comes to canceling subscriptions or trials,” Wallace said. “If it doesn’t enable the team to be more effective, or they spend more time managing the AI than doing it manually, we cut it.” His engineers use Cursor for pair programming and OpenAI’s ChatGPT for refining internal reports. “AI hasn’t replaced the need for any individual; it has made the existing team more productive,” he added.
Advertisement: Scroll to Continue
Creative functions are also being rebuilt around AI as teams find new ways to balance automation and human judgment. According to the a16z report, 10 of the top 50 startup AI applications fall under design, video or audio generation. Reflecting a broader trend among startups, Marlon Misra, co-founder and CEO of Assembly, shared with PYMNTS that his teams use AI as part of their everyday process.
“AI tools have been adopted fastest by our engineering and design teams,” he said. “We lean on AI for exploration and volume, and we lean on people for taste, judgment and setting direction.” Misra noted that even as AI accelerates output, human oversight remains essential to ensure creativity aligns with brand and purpose.
How Enterprises Are Embedding AI into Core Systems
PYMNTS research shows that 82% of enterprise CFOs are either using or actively exploring generative AI for forecasting, reconciliation, and risk management. Gen AI adoption has accelerated faster than any recent enterprise technology, with Karen Webster’s analysis noting that Gen AI has reached mainstream use nearly twice as quickly as cloud computing. She attributes the speed to accessibility. Employees experimented with consumer-grade tools like ChatGPT, pulling AI into organizations from the bottom up.
Larger enterprises are embedding AI into compliance, finance and operations systems originally built for manual oversight, while startups can architect around AI from day one. Across both settings, AI is evolving into a workflow layer that enhances precision and speed while keeping humans in control.
