DXC Technology today revealed it has formed a multi-year alliance with Anthropic to provide organizations with external expertise needed to build, deploy, manage and secure artificial intelligence (AI) applications based on Claude large language models (LLMs).
Under the terms of the agreement, the two companies are extending an existing relationship to create applications that will be tailored for specific vertical industries and domains. Specific areas of focus will include the insurance sector and, more broadly, cybersecurity and application services.
The goal is to help organizations navigate a host of AI issues and challenges that most every organization is now encountering, says Holly Grant, senior vice president for strategy and innovation at DXC.
DXC is establishing a dedicated team of forward-deployed engineers to work directly inside customer environments. These engineers will be selectively recruited from DXC’s existing engineering talent, trained and certified in 90 days through the Anthropic Partner Academy. DXC has also developed an additional certification curriculum to continuously train those teams as additional AI advances are made.
Much of that expertise was originally gained by using the Claude platform to build and deploy OASIS, a managed IT service that DXC rolled out last month. DXC relied on Claude to generate 95% of the code for that AI-native platform, which accelerated delivery by an estimated factor of ten, the company claims.
As part of that strategy, DXC will be focusing its AI expertise on a narrow range of AI models that, at this juncture, does not include OpenAI, which last month launched its own consulting firm.
In general, organizations are still working through myriad issues that arise as they try to operationalize AI. Along with defining a set of best practices for governing AI agents, organizations are also now trying to rein in the total cost of AI. Going forward, organizations will need to invest in the harnesses that will be required to route prompts to the AI model that is best suited to automate that task based on the number of tokens consumed, says Grant. “Tokenomics is a thing,” she says.
Organizations also now realize that while AI improves the personal productivity of employees, that has not yet translated into gains in productivity for teams or the organization itself, adds Grant.
Of course, every organization is on a slightly different AI journey. Many business leaders are still keen to show a return on investment (ROI) in AI that excites investors, but there are significant workflow challenges that need to be addressed to ensure the outputs generated are reliable. As such, many organizations are now taking a more systematic approach to integrating AI into their workflows.
Additionally, others are forming teams, also known as pods, that enable them to bring to bear the human expertise and creativity needed to fully operationalize AI, notes Grant.
Regardless of approach, there will continue to be a significant amount of AI experimentation, especially as the reasoning capabilities of AI models continue to become more advanced. The challenge and the opportunity now is making sure the AI model being used is the best fit for that specific purpose. Otherwise, the total cost of AI is likely to far exceed the benefits actually realized by the business.
