Amit Agarwal helped scale Datadog from its early stages into a major public software company as its Chief Product Officer and President. He now leads Standard Template Labs, serves on the boards of Datadog and Vercel, and is focused on shifting service management from tracking tickets to resolving work with AI. This profile covers his approach to product-led growth, enterprise software, and context-aware infrastructure.

Visual summary of operating lessons from Amit Agarwal.

Part 1: Product-Led Growth and Scaling

  1. On scaling developer tools: Transforming a developer product into an enterprise standard requires maintaining the deep customer obsession that fueled its early adoption. — Reference: cityam.com
  2. On multi-product expansion: A crucial factor in building a successful, scaled SaaS company is determining the right timing and strategy for launching a second product line. — Reference: saastr.com
  3. On early marketing strategies: In the early days of a startup, you do not necessarily need a formal marketing strategy; success can stem purely from product-led growth and adoption. — Reference: saastr.com
  4. On planning without pretending to predict: Five- or ten-year plans rarely survive contact with a changing market, so teams should keep listening to customers, challenge their assumptions, and make small course corrections as new information arrives. — Reference: ICONIQ interview on uncommon leadership
  5. On growing out of every job: In a seven- or eight-person company, leaders do nearly everything themselves. As specialists take over those tasks, the leader's job becomes identifying the two or three priorities that can genuinely move the business. — Reference: ICONIQ interview on uncommon leadership
  6. On hands-on leadership after delegation: Delegating responsibility should not turn a leader into a spectator; he believes leaders still need to go deep, put themselves on the line, and drive the few priorities that matter most. — Reference: ICONIQ interview on uncommon leadership
  7. On making failure discussable: Leaders create room for experimentation by candidly naming their own mistakes, which gives teams permission to learn rather than disguise failed ideas as successes. — Reference: ICONIQ interview on uncommon leadership

Part 2: The Evolution of Enterprise Service Management

  1. On industry stagnation: "IT service management was revolutionised two decades ago, but then the industry stopped evolving," — Source: techfundingnews.com
  2. On legacy system design: Legacy IT service platforms were designed primarily as systems of record and routing rather than systems that actually resolve issues. — Reference: iconiq.com
  3. On the trap of deep customization: The deep customizations enterprises build into older workflow platforms often create high switching costs that initially look like defensibility, but ultimately become a fragile liability. — Reference: iconiq.com
  4. On the burden of navigation: "The best IT professionals today spend most of their time navigating systems, chasing approvals, and re-gathering context that should already be at their fingertips." — Source: siliconangle.com

Part 3: Designing Agentic AI Platforms

  1. On superficial AI additions: "Adding a chatbot to a decades-old platform doesn’t make it intelligent — it just gives people a faster way to file the same tickets." — Source: siliconangle.com
  2. On maintaining context: To move beyond routing requests to resolving them, software must be grounded in a continuously updated digital twin that maps the relationships between employees, systems, and policies. — Reference: techfundingnews.com
  3. On automated resolution: AI can interpret the intent behind IT requests and automatically trigger the workflows necessary to fix them without human intervention for the majority of common tickets. — Reference: techfundingnews.com
  4. On human oversight: While AI agents can assemble context and resolve requests safely, humans must remain in the loop for complex situations that require judgment and oversight. — Reference: iconiq.com
  5. On changing how software teams work: With coding agents, he expects developers to spend more time specifying, testing, and parallelizing work, while product managers use AI to sharpen requirements before implementation. — Reference: The Uncommon Path — Amit Agarwal on AI Service Management
  6. On making AI work inspectable: Enterprise AI products need trust and safety mechanisms that let people inspect, supervise, and intervene in agent work instead of treating automation as an opaque handoff. — Reference: The Uncommon Path — Amit Agarwal on AI Service Management

Part 4: Go-to-Market and Pricing Strategy

  1. On early enterprise sales: Selling to large enterprise customers is notoriously difficult in the early stages of building a SaaS company. — Reference: saastr.com
  2. On the pricing valley of death: Startups must navigate the dangerous pricing gap between $25,000 and $100,000, where deals are too small for heavy enterprise sales motions but require too much friction for self-serve models. — Reference: saastr.com

Part 5: Customer Obsession and Feedback

  1. On solving complex challenges: "We believe that this placement reflects Datadog’s continued commitment to solving our customers’ most sophisticated challenges and building products that provide unmatched visibility into the performance, security, and cost of their traditional, cloud-based, or hybrid tech stack—from code to production." — Source: datadoghq.com
  2. On the role of analyst reports: Analyst reports are fundamentally rooted in customer feedback, serving as recognition of how effectively a platform supports users in their daily operations and digital transformation initiatives. — Reference: datadoghq.com
  3. On uncovering friction points: Direct conversations with IT leaders often reveal that teams are forced to throw headcount at manual coordination problems that software should have solved. — Reference: iconiq.com
  4. On reducing context switching: He treats the time users lose moving between systems as a product problem; reducing that friction can make existing work faster and reveal natural paths into additional products. — Reference: The Uncommon Path — Amit Agarwal on AI Service Management
  5. On software's practical purpose: He judges product work by whether it makes customers more productive, not by the company's branding or the amount of attention surrounding a launch. — Reference: The Uncommon Path — Amit Agarwal on AI Service Management

Part 6: Breaking Down Organizational Silos

  1. On unified visibility: A successful observability platform breaks down internal silos, allowing development, security, operations, and business teams to collaborate based on a single source of truth. — Reference: datadoghq.com
  2. On complex access environments: Managing internal IT requires understanding the intricate web of users, permissions, and access policies to avoid sluggish response times across fragmented tools. — Reference: techfundingnews.com
  3. On mapping organizational relationships: A useful service-management system should understand how people, applications, devices, permissions, and timing relate so it can resolve requests and plan changes with less operational disruption. — Reference: The Uncommon Path — Amit Agarwal on AI Service Management

Part 7: The Limitations of Legacy IT Systems

  1. On manual configuration databases: When systems change in a traditional IT environment, requiring administrators to update asset records manually leads to stale data that restricts the usefulness of automation. — Reference: siliconangle.com
  2. On brittle integrations: Over time, IT operations often become dependent on unreliable systems held together by custom code that is difficult to maintain and scale. — Reference: iconiq.com
  3. On operational drag: Service management has quietly morphed into one of the largest drags on productivity, with employees frequently waiting weeks just to get the basic access required for their jobs. — Reference: iconiq.com

Part 8: Building Generational Companies

  1. On platform evolution: "Vercel has done something incumbents rarely do. Rather than being displaced by the AI wave, it evolved into the agentic infrastructure that many AI-native builders reach for first," — Source: cityam.com
  2. On long-term vision: "I look forward to supporting the team as they take a platform developers already love and build it into a generational company." — Source: cityam.com
  3. On adding an operator's perspective in the boardroom: He joined Vercel's board in August 2026 with experience scaling a developer-focused product through enterprise adoption, an IPO, and more than $2.5 billion in annual revenue at Datadog. — Reference: Vercel board announcement via City AM
  4. On rebuilding from first principles: Foundational software categories are periodically rebuilt during major technology shifts; the agentic AI era presents an opportunity to reconstruct mission-critical infrastructure from the ground up. — Reference: iconiq.com
  5. On getting close to a technology shift: His year working with founders at ICONIQ gave him a direct view of small teams building useful AI products in months, which convinced him the moment was large enough to start another company. — Reference: The Uncommon Path — Amit Agarwal on AI Service Management