Oct 21, 2026 · Latent Space: The AI Engineer Podcast · 5 min read

10X AI: Output-Based Compensation and Enterprise Adoption

10X AI shares its strategy for transforming enterprises from AI-absent to AI-native. The discussion covers output-based compensation models, the critical gap in non-engineering agent deployment, and the strategic shift from vertical SaaS to horizontal agent platforms.

Technology Business Entrepreneurship

The takeaways

Pay for Output, Not Time. Shift compensation models from hourly billing to output-based metrics like story points. This aligns engineer incentives with AI leverage, rewarding high productivity and driving a 4x variance in earning capacity.
Standardize Processes Before Automating. You cannot automate a process that does not exist. Enterprises must first standardize workflows and documentation to create durable context for AI agents, enabling reliable background execution and evals.
Non-Engineering Agents Are Rare. Only 5% of large enterprises have non-engineering agents in production. The biggest diffusion gap lies in deploying agents for back-office tasks, hindered by permissioning issues and lack of standardized processes.
Token Costs Drive Model Choice. Monitor token spend relative to employee spend. As frontier models become cheaper and faster, companies should calibrate model usage to maximize ROI, avoiding unnecessary costs from premium models for simple tasks.
Vertical SaaS Faces Horizontal Pressure. Vertical-specific AI solutions struggle against horizontal frontier models unless they possess unique data moats. The future favors platforms that provide APIs and MCPs for integration rather than standalone siloed tools.
Reallocate, Don't Just Replace. Frame AI adoption as job reallocation rather than displacement. Elevate employees from repetitive tasks to strategic roles, leveraging AI to handle drudge work while humans focus on high-value problem solving.

The Shift to Output-Based Engineering

10X AI, an applied AI partner, is redefining enterprise software development by transitioning from time-based billing to output-based compensation. By paying engineers based on story points rather than hours, 10X has created a high-variance earning model where top performers can earn millions. This approach leverages AI to amplify individual intellect, resulting in a 2x increase in engineer output over the past year. The company argues that as AI solves technical problems, clients will increasingly pay for outcomes rather than the time required to build them.

The Enterprise Adoption Gap

Despite widespread AI interest, enterprise adoption remains fragmented. While 90% of large companies have rolled out general-purpose LLMs, only 50% have deployed coding agents, and a mere 5% have non-engineering agents in production. The primary barriers are not technological but operational: lack of standardized processes and complex permissioning structures. 10X identifies that enterprises often fail to automate because they have not first standardized their workflows. The "final mile" of AI diffusion requires building durable, documented processes that agents can reliably execute in the background.

Strategic Implications for Software Vendors

The conversation highlights a bearish outlook on vertical-specific SaaS solutions that lack unique data moats. As frontier models become cheaper and more capable, horizontal platforms are gobbiling up vertical use cases. Companies like Salesforce are pivoting to provide APIs and MCPs for integration rather than building standalone agents. For startups, the key to survival is either owning proprietary data or building tools that change the user's process, rather than just replicating existing functions with AI.

Conclusion

The future of enterprise AI is not just about deploying models but about restructuring work. Companies must standardize processes, manage token costs carefully, and view AI as a tool for job reallocation. By focusing on output and durable context, organizations can bridge the gap between raw AI intelligence and tangible business value.

Key insights

Output-based compensation models align engineer incentives with AI leverage, creating significant variance in earnings. This model rewards productivity and quality over time spent, driving higher overall output.

Compensation Strategy →

Impact: Companies adopting output-based pay can significantly increase engineering productivity and reduce costs, as engineers are incentivized to maximize AI usage and efficiency.
The majority of enterprises have not deployed non-engineering agents in production due to lack of standardized processes and permissioning challenges. This represents a massive untapped opportunity for AI transformation.

Market Opportunity →

Impact: Firms that can solve the process standardization and permissioning hurdles will capture significant market share in enterprise AI deployment, particularly in back-office functions.
Token costs are a critical factor in model selection. As frontier models become cheaper, companies should calibrate their usage to avoid paying premium prices for simple tasks, optimizing for ROI.

Cost Management →

Impact: Effective token cost management can reduce AI operational expenses by up to 80%, allowing companies to scale AI usage without proportional cost increases.
Vertical-specific SaaS solutions are at risk from horizontal frontier models unless they possess unique data moats or change the user's process. The future favors integration platforms over standalone silos.

Competitive Landscape →

Impact: SaaS companies must pivot to provide APIs and MCPs for integration, or focus on proprietary data, to remain competitive against general-purpose AI models.
AI adoption should be framed as job reallocation rather than displacement. Elevating employees from repetitive tasks to strategic roles increases overall business value and employee satisfaction.

Workforce Strategy →

Impact: Companies that successfully reallocate workforce to high-value tasks will see improved productivity and retention, while those that focus solely on cost-cutting may face resistance and talent loss.

Action items

Implement output-based compensation models for engineering teams. Define clear story points and align incentives with productivity and quality metrics.

Impact: This will drive higher engineer output and reduce costs, as employees are motivated to maximize AI leverage and efficiency.
Standardize internal workflows and documentation before deploying AI agents. Create durable, documented processes that agents can reliably execute in the background.

Impact: This will enable reliable agent deployment and reduce errors, allowing companies to automate back-office tasks and improve operational efficiency.
Monitor token spend relative to employee spend and calibrate model usage. Use cheaper models for simple tasks and reserve premium models for complex problems.

Impact: This will optimize AI operational expenses and improve ROI, allowing companies to scale AI usage without proportional cost increases.
Evaluate vertical SaaS dependencies and consider building custom solutions or integrating with horizontal AI platforms. Focus on proprietary data or process changes to maintain competitive advantage.

Impact: This will reduce reliance on vulnerable vertical SaaS solutions and leverage the capabilities of frontier models, improving flexibility and reducing costs.
Frame AI adoption as job reallocation and invest in employee reskilling. Elevate employees from repetitive tasks to strategic roles, leveraging AI to handle drudge work.

Impact: This will improve employee satisfaction and retention, while increasing overall business value by focusing human capital on high-value problem solving.

Quotes

“You can't automate a process that does not exist.”

“AI is really just a leveraged bet on a human's intellect and work ethic.”

“The biggest diffusion gap is from raw intelligence to enterprise usage.”