AI in Recruiting: Insights from ICONIQ Builder's Playbook 2025 State of AI Report

Artificial Intelligence (AI) continues to reshape industries, and recruitment is no exception. The evolving landscape of AI in recruiting is marked by rapid innovation, strategic adoption, and emerging challenges. A recent report by ICONIQ, titled the Builder's Playbook 2025 State of AI, offers a comprehensive look into how software companies—especially those focused on generative AI products—are navigating this transformative era.

Drawing on a survey of 300 executives across software companies of various sizes and growth stages, this report sheds light on the trends shaping the future of AI-powered recruitment and software development. This article delves deep into the findings, unpacking key areas such as building generative AI products, go-to-market strategies, compliance, organizational structures, AI costs, internal productivity, and the top AI tools dominating the market.

Table of Contents

• Understanding the Survey Landscape: Who Took Part?
• From Traditional SaaS to AI Native: The Spectrum of AI Adoption
• Model Usage: Proprietary vs Third-Party AI Models
• Dominant AI Models and Training Techniques
• Performance Monitoring and AI Agent Deployment
• Go-to-Market Strategy and Compliance in AI-Enabled Companies
• Conclusion: The Future of AI in Recruiting and Software Development

Understanding the Survey Landscape: Who Took Part?

The foundation of the ICONIQ report is a robust survey involving 300 executives from software companies. These companies varied widely in size and revenue, creating a rich dataset that reflects diverse perspectives:

• Revenue Range: Companies spanned from those generating under $10 million to giants pulling in over $1 billion annually, with 26% of respondents falling within this spectrum.
• High Growth Companies: About 13% of the respondents represented high-growth organizations, often characterized by rapid scaling and innovation.
• Established Companies: The remainder consisted of organizations with at least $10 million in annual revenue, offering insights from mature market players.

This diversity allows for a nuanced understanding of how different types of companies approach AI integration, particularly in the recruitment space, where accuracy and efficiency are paramount.

From Traditional SaaS to AI Native: The Spectrum of AI Adoption

The report categorizes companies into two broad groups to illustrate the spectrum of AI adoption:

1. Traditional SaaS Companies: These include familiar names like Atlassian and Miro, which historically focused on software-as-a-service products without a core AI focus.
2. AI Native Organizations: Newer companies such as Cursa and Eleven Labs, which are built from the ground up around AI technologies and generative AI products.

This distinction is crucial in understanding how companies innovate and scale AI solutions. AI native companies, by design, are more agile in leveraging AI capabilities and integrating them deeply into their products.

Building Generative AI Products: Market Fit and Scaling

One of the most critical insights from the report is the stage of product development and market adoption among AI-focused companies:

• Only 13% of all surveyed organizations have successfully proven product-market fit for their AI products and are now focused on scaling.
• Within the AI enabled category—companies that augment existing products with AI features—this 13% figure highlights early but promising traction.
• In contrast, 47% of AI native companies are ready to scale their AI offerings, reflecting their inherent focus and momentum in the AI domain.

This data reveals that while AI adoption is widespread, achieving a strong market fit and moving into scaling phases is still a milestone that many companies are striving to reach, especially outside the pure AI native ecosystem.

What AI Products Are Companies Building?

The report highlights two primary types of AI products being developed:

• Agentic Workflows: These are AI-driven processes that automate tasks and decision-making workflows. A striking 79% of companies are focused on building these, indicating a strong trend towards automating complex workflows.
• Core AI Technologies: Only 27% of companies are developing foundational AI technologies themselves, suggesting that most firms rely on existing AI frameworks rather than building from scratch.

Within AI native companies, the focus on agentic workflows remains high at 79%, signaling an emphasis on creating intelligent, autonomous systems that can drive significant operational efficiencies.

The report also notes that 62% of AI enabled companies are working on agentic workflows, showing that even traditional SaaS firms are embracing more sophisticated AI applications.

Model Usage: Proprietary vs Third-Party AI Models

When it comes to AI model usage, the report distinguishes between companies developing proprietary AI models and those relying on third-party APIs:

• High Growth Companies: A significant 71% rely on third-party AI models rather than building their own, which allows them to leverage cutting-edge technology without the heavy investment of in-house development.
• Other Respondents: Even more pronounced, 80% use third-party AI APIs.
• Proprietary Model Development: Only 32% of companies are developing proprietary AI models from scratch, highlighting the challenges and costs associated with creating custom AI technology.

This reliance on third-party models is especially relevant for recruitment technology providers, where integrating accurate and reliable AI is key to delivering value.

Accuracy: The Core Concern in AI Recruitment

Among all AI performance metrics, accuracy stands out as the most critical factor. This is especially true in recruitment, where AI-driven decisions directly impact candidate selection and hiring outcomes.

One example cited is EQ.app, a recruitment AI platform that prioritizes accuracy as a core tenant of its offering. For companies considering AI in recruiting, accuracy is not just a feature—it’s a fundamental requirement that builds trust and effectiveness.

If you are exploring AI recruitment tools, platforms like EQ.app are worth considering for their focus on delivering precise, reliable AI insights that enhance hiring decisions.

Dominant AI Models and Training Techniques

The report identifies OpenAI’s GPT models as the dominant player in the AI ecosystem:

• 95% of respondents use OpenAI GPT models, underscoring their widespread adoption and reliability.
• Elon Musk’s xAI models currently hold only a 2% usage share, indicating OpenAI’s clear lead in the market.

Between these, companies like EQ.app develop proprietary models for specific use cases while also leveraging open source and top-tier models to ensure versatility and performance.

Regarding training techniques, the landscape is diverse but with clear favorites:

• Fine-Tuning: The most popular method, where existing models are adapted to specific tasks or datasets.
• Retrieval Augmented Generation (RAG): An advanced technique where AI models pull information from multiple data sources to improve relevance and accuracy. This technique is gaining traction for its ability to enhance AI responses by grounding them in real-time or specialized data.

Infrastructure and Challenges: Cloud, GPUs, and Hallucinations

Operating AI models at scale requires robust infrastructure, and the report highlights several key points:

• Cloud Adoption: Full cloud deployment is essential for scalability and flexibility.
• Hallucinations: The top challenge reported by 39% of respondents. Hallucinations refer to AI models generating inaccurate or misleading information, a critical concern in recruitment where decisions must be based on factual data.
• GPU Access: Surprisingly, only 5% cite accessing GPUs as a major challenge, despite the crucial role of GPUs in AI training and inference. This may reflect the growing accessibility of cloud GPU resources.

Companies like NVIDIA are experiencing growth due to their GPU products, but for now, infrastructure bottlenecks are less pressing than managing AI output quality.

Performance Monitoring and AI Agent Deployment

Ensuring AI models perform well throughout their lifecycle requires advanced monitoring, especially as products move from beta to general availability and scaling phases:

• Advanced Monitoring: 44% of companies use advanced monitoring tools to track AI performance, detect anomalies, and ensure reliability.
• Agentic Workflows Deployment: 32% of companies outside the high-growth segment actively deploy AI agents—autonomous AI-driven processes that can execute tasks without constant human oversight.
• This leaves 68% of companies not yet deploying AI agents, perhaps due to complexity, risk aversion, or lack of readiness.
• Among high-growth companies, 47% have embraced AI agents, reflecting their willingness to adopt cutting-edge AI workflows to accelerate operations and innovation.

Go-to-Market Strategy and Compliance in AI-Enabled Companies

The integration of AI into product roadmaps is a significant focus for many organizations:

• AI-Enabled Companies: Allocate approximately 20-35% of their product roadmap to AI-driven features.
• High Growth Companies: Commit an even larger share—30-45%—to AI-related development, signaling a strategic prioritization of AI capabilities.

This indicates that AI is not merely a buzzword but a core component of future product strategies. Compliance, particularly with evolving AI regulations and ethical standards, is also an implicit concern, especially as companies scale AI-driven products to broader markets.

Conclusion: The Future of AI in Recruiting and Software Development

The ICONIQ Builder's Playbook 2025 State of AI report paints a detailed picture of the current and near-future state of AI adoption in the software industry, with direct implications for recruitment technology:

• AI Adoption Is Accelerating: With nearly half of AI native companies ready to scale and a growing share of AI-enabled firms integrating agentic workflows, the momentum is clear.
• Accuracy Remains Paramount: Especially in recruitment, where AI decisions impact human lives and business outcomes, accuracy is the foundation of trust and effectiveness.
• Third-Party Models Dominate: Most companies leverage existing AI models like OpenAI’s GPT, balancing innovation with pragmatic resource management.
• Challenges Persist: Hallucinations and output reliability are key issues to address as AI systems become more autonomous.
• Strategic Roadmapping: AI features are becoming central to product development, with compliance and monitoring playing critical roles.

For recruitment professionals and software companies alike, understanding these trends is essential. AI is not just a tool; it is becoming the backbone of smarter, faster, and more effective recruitment processes. Platforms like EQ.app demonstrate how focusing on accuracy and leveraging AI can revolutionize hiring, making it more efficient and equitable.

As we move deeper into 2025, those who embrace AI thoughtfully—balancing innovation with ethical considerations and technical rigor—will lead the charge in reshaping the recruiting landscape.

If you're involved in recruitment or software development, keeping a pulse on these developments will help you stay competitive and harness the full potential of AI in recruiting.