In enterprise software, the "optimization (Customization) strategy" was long seen as a term representing old-fashioned IT. That view has recently shifted sharply. AI agents are the reason.

Industry sources say that as AI agents spread, optimizing software development to fit a company’s internal environment is no longer something to discard as the past. It is rapidly emerging as a keyword symbolising the future.

Details vary, but there is a growing view that the AI agent market needs work similar to system integration (SI), which develops software to what customers want. Some also say AI agents without optimization will struggle to take root in the enterprise market, which is seen as the major leagues of corporate AI.

This is also reflected in major AI companies expanding hiring of forward deployed engineers (FDE) who provide hands-on support on the ground so companies can properly build AI agents. Palantir is cited as a leading example that benefited from an FDE strategy, and other companies have recently been moving in a similar direction.

OpenAI and Anthropic have put forward the option of forming joint ventures to support enterprise AI adoption.

Anthropic plans to set up a joint venture focused on providing AI services for companies, together with Blackstone, Hellman & Friedman and Goldman Sachs, among others. Venture capital, hedge fund and private equity firms including Apollo Global Management, General Atlantic, GIC, Leonard Green and Sequoia Capital will also take part. The joint venture will adopt Palantir’s FDE model and operate by directly deploying engineering resources to individual companies.

OpenAI has also set up a joint venture, OpenAI Deployment Company, to support companies’ AI adoption and integration in cooperation with investment firms and consulting companies as part of its push into the enterprise market.

The joint venture was established jointly with 19 investment firms, consulting firms and system integrators including TPG, Advent, Bain Capital and Brookfield.

OpenAI also acquired AI consulting and engineering firm Tomoro alongside the launch of the joint venture. The acquisition will bring about 150 FDEs into the joint venture. The establishment of the joint venture is a move to increase enterprise customers’ AI spending. Anthropic also set up a separate joint venture in early May with private equity firms including Blackstone to target the enterprise market.

A recent report by The Information said Google also plans to create an FDE team within Google Cloud and hire several hundred engineers. Google Cloud Chief Revenue Officer Matt Renner said, "We are not just sending more salespeople. We will reach customers with more technical resources."

Industry moves toward FDE are based on the reality that building AI agents to fit a company’s internal circumstances is harder than expected.

Aaron Levie, CEO of collaboration software firm Box, said, "The forward deployed engineer role will become one of the most in-demand jobs in the technology sector. Deploying agents is a much more technical task than most people know. It is often far more complex than deploying software." He added, "To deploy agents properly, vendors need to deeply understand business processes." He added, "Companies need help identifying the model that best fits their workflows, and they also want support for workflow change management. Agent systems need to be continuously tuned to fit processes."

The industry expects demand for optimization related to AI agents is unlikely to be a temporary phenomenon seen at an early stage of AI and is likely to continue.

Yum Dong-hoon (염동훈), CEO of MegazoneCloud, said, "Companies will need to keep optimizing AI agents in line with changing circumstances." He said, "Just as the AWS platform provides countless building blocks like Lego blocks, AI agents also need to be combined to fit the nature of work. The AI needed by a marketing organisation and the AI needed by a finance team are different. Choosing the right AI for the work is a core capability in the AI era." He added, "When moving from on-premises to the cloud, only the platform changed while business processes stayed the same," and said, "In the AI era, business processes themselves need to change to maximise returns on AI investment."