Tech • AI • Robotics
How Turing is unlocking new creative approaches by "AI-ing their way out" of problems
5/10TL;DR
Turing used a bottom-up AI adoption strategy to rapidly scale hiring, automate HR support, and cut content production times from weeks to hours while keeping humans focused on higher-empathy work.
KEY POINTS
Rapid onboarding at scale
Turing’s HR and talent team was asked on a Tuesday to onboard 800 people by the following Monday, and met the deadline. The effort became a defining example inside the company of using AI to solve urgent operational problems at speed. That experience helped establish the internal motto of finding ways to “AI” through difficult workflows rather than expanding headcount or relying solely on manual processes.
Bottom-up adoption over mandates
The company favored a bottom-up approach to AI adoption instead of ordering employees to use a specific tool from the top down. Leaders focused on helping staff discover practical uses in their own daily work, arguing that adults learn AI best by experimenting directly rather than sitting through compliance-style training. The strategy centered on curiosity, creativity and identifying repetitive tasks where AI could provide immediate leverage.
Offsites built around practical experimentation
One of the most popular internal programs has been a series of “AI our way out of this” offsites, where small groups work through real business problems. Using Gemini, teams designed solutions tied closely to operational needs rather than abstract demos. An initial concept created by talent operations and people leaders was later scaled by engineering, passed governance checks and is now expected to save several thousand hours of manual work annually.
HR support automation expanded quickly
In 2025, the people operations team was handling about 60,000 tickets manually with a flex team of roughly 12 to 16 people. To automate part of that workload, the company built a conversational AI assistant called Allen and divided requests into two categories: issues suitable for fast, ATM-style responses and cases that still needed a more human, “bank teller” interaction. That framing helped determine where automation would be accepted and where human judgment remained essential.
A sudden jump from 0.5 to 5 FTE of output
Early on, Allen handled a volume roughly equal to half of a full-time employee. After improvements to instructions and the underlying knowledge base, the system hit an inflection point and over about 72 hours began processing work equivalent to around 5 FTEs. The result was not only operational efficiency but also enough quality to satisfy both employees and HR teams using the service.
Automation increased demand rather than reducing it
Faster responses did not shrink HR demand. Instead, ticket volume at Turing roughly doubled, echoing a broader pattern in AI-enabled service operations: when people get useful answers quickly, they ask more questions. The company also expanded its knowledge base beyond traditional HR topics, extending the reach of the support model across more parts of the organization.
Humans shifted to higher-empathy work
Rather than eliminating the need for HR specialists, automation allowed them to focus on the more complex 20% of cases where people wanted deeper conversation, empathy and judgment. The company framed AI use cases intentionally as tools that augment human decision-making and creativity rather than replace them. That design choice was seen as central to building trust internally.
Training content creation collapsed from weeks to hours
On the talent development side, new coding capabilities in Gemini allowed non-engineers to rebuild parts of Turing University without software developers. A prototype module was created during a cross-country flight, illustrating how quickly internal tools could now be assembled. The team said content creation time dropped from an industry norm of 2 to 5 weeks to roughly 2 to 4 hours.
Trust depended on governance and familiar environments
For organizations cautious about AI, the main concerns remained safety, trust and control. A key lesson was that adoption moved more smoothly when AI tools lived inside trusted enterprise environments such as Google Workspace, reducing the need to justify unfamiliar third-party applications to security leaders. The broader challenge was described less as a technology rollout than as a human change management process.
A blueprint for scaling AI adoption
The company’s framework resembled a two-stage mission: first identify practical task-level use cases, then let those gains compound into larger transformation. Employees were encouraged to examine all of their responsibilities and ask where AI could provide leverage. Leadership advice emphasized empathy, acknowledging that skepticism toward generative AI remains real and must be addressed directly if adoption is to spread.
CONCLUSION
Turing’s experience suggests that the strongest early returns from AI come from employee-led experimentation, targeted automation and clear boundaries around human judgment. The broader payoff is not just efficiency, but a redesign of work that moves people toward more creative and empathetic tasks.
Explain this