In a significant development for the business technology sector, a new McKinsey & Company survey has revealed that 32% of organizations have chosen to skip purchasing at least one software product or feature, opting instead to build the needed functionality in-house using AI agents in business workflows. The survey, released August 30, 2026, indicates that adoption of AI agents is not only accelerating but is also reshaping how enterprises make product and procurement decisions. Within the first 100 words, these findings illustrate how agentic automation is fundamentally shifting enterprise technology strategies.
The rise of AI agents in business strategy
AI agents—automated software systems that act on behalf of users or organizations for specific intentions—are increasingly integral to business operations. According to McKinsey’s State of AI in 2026 survey, not only are these agents becoming more capable, but their impact is also being felt across both product development and technology purchasing pipelines. Large enterprises, in particular, are leading the shift, with 40% reporting active deployment of agents in one or more organizational functions, up from 27% last year. In contrast, adoption among smaller firms remained flat at around 22%.
What is driving the adoption of AI agents?
- Cost savings: By developing internal features with AI agents, companies reduce reliance on external software vendors.
- Customization: In-house agentic solutions provide flexibility to tailor functionality to business-specific needs.
- Rapid iteration: Internal tech teams can update, improve, or repurpose AI agent–built features quickly—sometimes in days rather than weeks or months.
- Competitive advantage: Building proprietary agent-powered tools can deliver process efficiencies not available to rivals using generic off-the-shelf software.
Detailed findings from the McKinsey survey
McKinsey’s 2026 survey, which included more than 550 engineers and technology leaders across the US and UK, found the following key changes over the past year:
- Agent use grew sharply at large enterprises, rising from 27% to 40% of companies deploying agents in at least one major business function.
- Smaller firms’ reported use of agents plateaued at around 22%—suggesting a scale gap in who can fully leverage agentic automation.
- 80.8% of respondents now use AI agents on a daily basis, up from 47.3% a year earlier—a 70.8% relative increase in habitual use.
These statistics demonstrate that AI agents in business have moved beyond experimental pilots and now underpin core process workflows for many large organizations.
Real-world examples and applications
Agentic automation is being adopted across sectors. For instance:
- Financial services: Major banks are rebuilding chatbots and customer consultation tools as AI agents that can securely handle transactions, automate onboarding, and even escalate complex cases to human staff when necessary.
- Retail: Companies are deploying agents to automate inventory management, personalize online shopping experiences, and streamline supply chain communications.
- Manufacturing: AI agents are increasingly tasked with predictive maintenance, quality assurance, and dynamic resource allocation in factories.
A notable example comes from South Korea, where telecom company KT is partnering with Woori Bank to rebuild its AI chat and consultation system with agent-connect features, allowing task handoff between automated assistants and human managers, ensuring cohesive customer service and data integrity across channels.
Challenges: Security, governance, and talent
As agentic coding becomes standard practice, organizations face technical and regulatory hurdles:
- Security and provable authorization: Surveyed technology leaders highlighted the challenge of maintaining robust permission records as more processes become autonomous. Google’s recent “Secure AI Framework” whitepaper, and ongoing NIST policy work on agent identity and provenance, are notable responses.
- Governance: NIST’s concept of the “AI AGENT Act” (S.5051) and Google’s Agent Payments Protocol (AP2) are examples of regulatory and commercial mechanisms now emerging to ensure responsible deployment.
- Skills gap: To fully realize benefits, organizations need employees skilled in task-bounded automation, orchestration, and agent development.
Implications for software vendors and the broader technology market
The finding that nearly one-third of organizations are now building over buying for at least some features signals a major shift in enterprise procurement. This trend threatens the traditional software-as-a-service (SaaS) business model for generic features, while placing a premium on vendor products offering domain expertise, regulatory compliance, or highly specialized integrations.
Industry analysts note that mass SaaS vendors may need to focus more on composability and model extensibility, partnering with IT departments in a co-development role rather than acting only as outsourcers. Meanwhile, open-source AI agents and communities such as the Hugging Face ecosystem continue to lower technical barriers for in-house agent adoption—a trend likely to accelerate further.
Context and further reading
For technology leaders, these developments require ongoing monitoring of both internal agent orchestration efforts and the evolving regulatory landscape. Practical resources include the latest CyberProfi business technology coverage and CyberProfi artificial intelligence explainers, which regularly analyze the impact of emerging agentic automation.
Frequently asked questions
- What are AI agents and how are they used in business?
- AI agents are automated digital systems that perform tasks on behalf of users with little or no human intervention, increasingly automating tasks across functions such as operations, customer service, finance, and supply chain management.
- Why are more organizations building internal tools instead of buying software?
- Companies are finding they can better tailor solutions to their specific needs, iterate faster, and reduce costs by building with AI agents, especially for non-core or highly customized functionalities.
- Are there security risks with deploying AI agents at scale?
- Yes, securing agentic workflows is a foremost challenge, with new frameworks and policies (such as the Secure AI Framework and AI AGENT Act) emerging to address authorization, auditability, and data integrity.
- How can small firms catch up to large enterprises in AI agent adoption?
- Access to open-source models and platforms, as well as improved developer tooling, will be essential. Smaller organizations must also invest in workforce upskilling and process redesign.
- How should technology providers respond to the trend toward in-house agent development?
- Vendors should prioritize partnerships, extensibility, and domain expertise, helping buyers build on their platforms rather than simply offering packaged solutions.