Business
2024 saw enterprise AI shift from innovation projects to core business tools. At the end of last year, Menlo Ventures published their "2024 State of Generative AI" report. Drawing insights from 600 U.S. enterprise leaders, the report reveals growing confidence in AI adoption, even as companies navigate the complexities of implementation.
In this analysis of the report, we'll examine five key developments that shaped enterprise AI in 2024:
Enterprise AI spending has exploded to $13.8 billion in 2024 - a staggering 6x increase from the $2.3 billion spent in 2023. This isn't just another technology trend; it's a fundamental shift in how businesses view AI's role in their operations.
What makes this surge particularly intriguing is its source. While 60% of spending still flows from innovation budgets (reflecting AI's emerging status), 40% now comes from permanent operational budgets. More tellingly, 58% of this permanent budget represents funds redirected from existing allocations. This signals a profound change: enterprises aren't just experimenting with AI; they're restructuring their entire operational framework around it.
Yet, this transition isn't without its paradoxes. Despite the massive investment surge, more than a third of surveyed companies still lack a clear vision for implementing AI across their organisations. This shouldn't surprise us though. We're witnessing a classic pattern in technological adoption where investment and understanding don't always move in lockstep. At Kiseki Labs, we've observed this firsthand - organisations often recognise AI's transformative potential before fully grasping how to harness it.
The key here is understanding that this apparent contradiction - massive spending coupled with strategic uncertainty - actually reflects the transformative nature of AI technology.
Unlike previous technological shifts that could be contained within specific departments or functions, AI has the potential to reshape entire organisational structures and workflows.
While foundation models continue to dominate headlines, 2024's most significant developments occurred in the application layer. Enterprise buyers have poured $4.6 billion into AI applications this year, representing an almost 8x increase from 2023's $600 million. This shift in spending patterns indicates that enterprises are more focused on practical implementation than technological novelty.
This focus on applications makes sense given the increasing commoditisation of foundation models. When companies like DeepSeek can match the performance of OpenAI's models at a fraction of the cost (as we explored in our blog post about DeepSeek v3), it becomes clear that sustainable competitive advantage lies not in the models themselves, but in how organisations apply them to solve specific business problems.
DeepSeek's achievement of matching GPT-4o and Claude 3.5 Sonnet's performance while spending significantly less on training demonstrates how the barriers to entry for frontier AI are lowering. The real value creation is happening at the application layer, where companies are building solutions that address unique industry challenges and deliver measurable business outcomes.
The adoption figures indicate which AI applications are delivering the most value for enterprises:
Code copilots have emerged as the clear leaders, achieving 51% adoption. GitHub Copilot is leading the charge in this area with an impressive $300 million revenue run rate, but tools like Cursor are also rapidly growing. Developers have likely become the most prominent AI power users, embracing everything from general coding assistants to specialised tools for DevOps and test automation. What makes this particularly interesting is how these tools are transforming the very nature of software development. They're not just making developers more productive; they're changing how developers think about and approach coding tasks.
Support chatbots have captured 31% of enterprise adoption, while enterprise search and data transformation solutions follow closely at 28% and 27% respectively. These adoption rates show that organisations are focusing on use cases that deliver immediate operational efficiency. The success of these implementations comes from their ability to solve specific, well-defined business problems like customer support automation or improving how employees find internal information.
A surprising finding from the Menlo Ventures report: only 1% of enterprise leaders consider price a primary factor when selecting AI solutions.
Instead of focusing on costs, organisations are prioritising measurable value delivery (30%) and industry-specific customisation (26%). This suggests that enterprises view AI as a strategic investment rather than just another technology purchase to be evaluated primarily on cost.
This shift in priorities reflects a deeper understanding of AI's potential impact. Enterprises are increasingly viewing AI investments through the lens of long-term transformation rather than short-term cost savings. They're willing to pay a premium for solutions that can demonstrate clear value and align with their specific industry needs.
Perhaps the most intriguing development of 2024 has been the emergence of AI agents. While still in their early stages, agentic architectures already power 12% of enterprise AI implementations.
At Kiseki Labs, our work with AI agents has revealed both their immense potential and current limitations. These systems excel at specific high-cognition tasks but still require human oversight for end-to-end processes. As we explored in our blog post about AI Agents, this is largely due to challenges with reliability and latency in multi-agent systems.
Agentic systems are transforming how organisations approach automation. Unlike traditional automation tools that follow rigid, predefined processes, AI agents can adapt to changing conditions and handle complex, multi-step tasks that require reasoning and decision-making. This adaptability comes from their unique architecture where multiple specialised agents work together to break down and solve complex problems, much like teams do in successful organisations.
Equally interesting is enterprises' approach to model selection.
Rather than relying on a single provider, organisations typically deploy three or more foundation models, routing to different models based on specific needs.
This pragmatic approach has reshaped the competitive landscape: OpenAI's market share has declined from 50% to 34%, while Anthropic has doubled its presence from 12% to 24% following Claude 3.5 Sonnet's release.
A shift has occurred in how enterprises approach AI implementation. The split between in-house development and vendor solutions now stands at nearly 50-50, with 47% of solutions developed internally and 53% sourced from vendors. This represents a significant change from 2023, when 80% of enterprises relied on third-party software.
From our experience at Kiseki Labs, this statistic requires careful interpretation. While more enterprises are indeed developing internal AI capabilities, many struggle to match the sophistication of specialised providers. The reality often lies somewhere between pure in-house development and complete outsourcing.
As we look towards 2025, several transformative trends are emerging:
First, AI agents will drive the next wave of transformation. We expect to see new business models emerge where value delivery, rather than traditional SaaS subscriptions, becomes the primary pricing mechanism. This shift could fundamentally change how enterprises consume and pay for AI services, moving towards models that more closely align with actual value creation.
Second, traditional IT outsourcing firms face an existential threat. If AI can consistently produce better code at a fraction of the cost, their current business model becomes unsustainable. We're already seeing early signs of this disruption, with some outsourcing firms scrambling to integrate AI into their offerings.
Finally, we're heading towards a severe talent drought. The gap between demand for AI expertise and available talent will widen, pushing organisations to fundamentally rethink their approach to technical recruitment and training. This isn't just about hiring data scientists or machine learning engineers; it's about finding people who can bridge the gap between AI capabilities and business needs.
Despite the impressive growth numbers and rapid adoption, we're still in the early stages of this transformation. The challenges around implementation, reliability, and talent show how much work remains to be done. What's clear is that the enterprises that succeed won't be those that simply adopt AI technologies, but those that fundamentally reimagine their operations around AI's capabilities.
At Kiseki Labs, we're excited to be part of this journey. Whether you're just starting your AI transformation or looking to scale existing implementations, we're here to help you navigate this rapidly evolving landscape. Book a free AI consultation with us to explore how we can help you build your AI-powered future.
LLMs
February 3, 2025
This blog post explains what DeepSeek v3 is and why it matters for frontier LLM development.
Read more
AI Agents
December 1, 2024
In this post I provide some reasons for why AI Agents are a useful paradigm for developing LLM-based applications that can tackle increasingly complex problems. Read on to learn more!
Read more
AI Agents
December 1, 2024
In this post I explain why I believe that Andrej Karpathy's vision for "Software 2.0" is finally within reach thanks to AI Agents. Read on to learn more.
Read more
