Mar 03, 2026 · Dev Interrupted · 5 min read

Monday.com's AI Strategy: Infrastructure First

Monday.com VP of RD Sergey Lykoveski details how the company paused its roadmap for 30 days to enable AI across 700 engineers. This strategy prioritized foundational infrastructure over quick wins, resulting in a four-tier AI product suite and significant operational efficiency gains.

Technology Business Strategy Engineering Management

The takeaways

Pause Roadmap for AI Enablement. Dedicate 30 days to halt standard roadmap work and focus exclusively on AI integration across the engineering organization. This intensive period accelerates skill acquisition and cultural shift more effectively than gradual training.
Prioritize Infrastructure Before AI Features. Resolve core technical debt and scale limitations before deploying AI features. A robust foundation ensures AI-driven traffic does not compromise system reliability or customer trust.
Design AI Products for User Intent. Segment AI offerings by user technical proficiency and intent, ranging from simple prompt-based solutions to complex vertical agents. This tiered approach maximizes adoption across diverse user bases.
Implement Zero-Bureaucracy Tool Adoption. Allow engineers to self-select and champion AI tools with minimal approval friction. Ownership of tool selection increases engagement and ensures the right tools are used for specific tasks.
Align Experiments with Production Goals. Require all AI experiments to have a clear path to production or specific business value. Avoid throwaway demos by tying innovation directly to customer commitments and product roadmaps.
Monitor Agent-Driven Infrastructure Metrics. Shift from CPU-bound to GPU-bound resource management, implementing fairness indices and concurrency controls for AI agents. New SLOs must account for agent fan-out and cost efficiency.

Strategic Pivot: Infrastructure as the AI Enabler

Monday.com’s approach to AI adoption challenges the common narrative of rapid feature deployment. Instead of immediately shipping AI capabilities, the company paused its product roadmap for 30 days to focus on internal enablement and foundational infrastructure. This deliberate pause allowed the 700-person engineering organization to shift its mindset, resulting in every developer using AI daily and the creation of a robust, four-tier AI product suite. The core insight is that AI scalability is contingent on underlying system reliability; without resolving technical debt and scale limitations, AI-driven traffic would have eroded customer trust rather than enhancing it.

The 30-Day AI Enablement Framework

The 30-day initiative was designed not as a hackathon, but as a production-oriented transformation. Key principles included aligning experiments with existing customer commitments and ensuring all outputs had a path to production. By removing bureaucratic barriers to tool adoption and empowering engineers to champion their chosen AI tools, Monday.com fostered a culture of ownership. This bottom-up approach generated 5,000 solutions in three months and 40,000 apps in two, demonstrating that structured experimentation yields higher adoption rates than top-down mandates.

Product Architecture: Tiered AI Solutions

Monday.com’s AI portfolio is segmented by user intent and technical proficiency, creating a compound value effect. Monday Magic serves as an entry point for prompt-based solution building, while Monday Vibe enables application creation within the platform. Sidekick functions as a horizontal copilot for general platform tasks, and Agent Factory allows for the construction of vertical, domain-specific agents. This tiered structure ensures that users at all skill levels can derive value, while the shared context between these tools enhances overall platform stickiness and data utility.

Operational Impact and Future Metrics

The strategic focus on infrastructure has yielded measurable operational improvements, including a reduction in complex ticket resolution time from three days to one and a doubling of test coverage. As the platform transitions from CPU-bound to GPU-bound workloads, new metrics such as fairness indices and agent concurrency are becoming critical. This shift requires redefining SLOs to manage cost and resource allocation for AI agents, marking a new frontier in SaaS reliability and operational management.

Key insights

AI adoption requires a foundational infrastructure overhaul before feature deployment. Monday.com resolved technical debt and scale issues to ensure AI traffic did not compromise system reliability.

Infrastructure Strategy →

Impact: Prevents customer churn caused by AI-induced performance degradation and builds trust for enterprise-scale AI usage.
A 30-day dedicated pause on the roadmap accelerates organizational AI enablement more effectively than continuous training. This intensive period shifts engineering culture and skills rapidly.

Organizational Change →

Impact: Increases developer productivity and AI utilization rates, leading to faster product iteration and innovation cycles.
Tiered AI product design, segmented by user intent and technical proficiency, maximizes adoption across diverse user bases. This approach ranges from simple prompts to complex vertical agents.

Product Strategy →

Impact: Expands the total addressable market by catering to non-technical and technical users, increasing platform stickiness and value.
Zero-bureaucracy tool adoption policies, where engineers champion their own tools, drive higher engagement and better tool selection. This reduces friction and increases ownership.

Engineering Culture →

Impact: Accelerates the adoption of best-in-class AI tools and fosters a culture of experimentation and continuous improvement.
AI agents introduce new infrastructure challenges, shifting workloads from CPU-bound to GPU-bound. This requires new metrics for fairness, concurrency, and cost management.

Operational Metrics →

Impact: Ensures sustainable scaling of AI features and prevents resource exhaustion, maintaining service level objectives for all customers.

Action items

Conduct a 30-day AI enablement sprint, pausing standard roadmap work to focus on internal AI integration and skill development. Define clear production-oriented goals for all experiments.

Impact: Accelerates organizational AI readiness and ensures that AI initiatives align with business objectives rather than remaining isolated demos.
Audit and resolve core technical debt and scale limitations before deploying AI features. Prioritize infrastructure upgrades that support high-concurrency, GPU-bound workloads.

Impact: Prevents performance degradation and maintains customer trust as AI-driven traffic increases, ensuring long-term platform stability.
Design a tiered AI product suite that segments features by user intent and technical proficiency. Create entry-level prompt tools, mid-level application builders, and advanced vertical agents.

Impact: Maximizes user adoption across different skill levels and creates a compound value effect through shared context and data utility.
Implement a zero-bureaucracy policy for AI tool adoption, allowing engineers to self-select and champion tools with minimal approval friction. Require tool champions to demonstrate value within a set period.

Impact: Increases engineer engagement and ensures the adoption of the most effective tools, reducing friction and accelerating innovation.
Develop new operational metrics and SLOs for AI agent workloads, including fairness indices, concurrency controls, and GPU cost management. Monitor agent fan-out and resource allocation.

Impact: Ensures sustainable scaling of AI features and prevents resource exhaustion, maintaining service reliability and cost efficiency.

Quotes

“trust is the currency for our customers and trustworthy experience”

“if you will not be able to experiment, if you will not give teams to experiment, it will be a failure”

“we are moving from SaaS that is CPU bound to SaaS that is GPU bound”