Automate Workflows w/ Autonomous Tools Real-Time AI Agents | Aytekin Tank | Founder & CEO | Jotform
Aytekin Tank, Founder and CEO of Jotform, discusses how AI is revolutionizing software development and customer support, highlighting "vibe coding" and Jotform's 80% AI-handled support with zero layoffs. He also shares his "Theory of Constraints" playbook for scaling businesses and explores the potential of ChatGPT apps, all within Jotform's 20-year bootstrapped journey.
Aytekin Tank, Founder and CEO of Jotform, delves into how AI is fundamentally reshaping software development, customer support, and strategic business growth, drawing insights from Jotform’s 20-year bootstrapped journey.
AI’s Transformative Impact on Coding
Tank highlights that AI writing production-level code is the most significant and rapidly evolving trend. Jotform’s developers actively engage in “vibe coding,” where AI generates the majority of the code. Tank personally demonstrated this by rebuilding broken websites using AI (specifically Claude Code) without writing a single line, emphasizing that the new skill is defining the end goal, not prescribing the steps. He notes that this capability, particularly with advanced models like Claude Code and Codex, arrived sooner than expected and can effectively manage large, established codebases, likening its impact to the initial launch of ChatGPT. While AI’s prowess in UI/UX and design is still developing, its advanced code generation capabilities are a groundbreaking achievement.
AI in Customer Support and Workforce Evolution
Jotform has remarkably transitioned 80% of its customer support to AI agents, handling over 50,000 daily conversations without any layoffs among its 200-person support team. This was achieved by evolving the support staff into “AI trainers” who review, correct, and improve AI agents. Their responsibilities now include identifying AI failures, creating training data, and reporting product issues to developers. This strategy retains invaluable institutional knowledge and experience, making the human workforce significantly more productive and allowing them to focus on complex, nuanced problems. This integration also dramatically accelerates product development and support efficiency, with daily AI-generated reports informing product teams of recurring user issues.
The Rise of ChatGPT Apps and AI Agents
Tank predicts that “ChatGPT apps” (and similar connector frameworks like Claude’s) will become as transformative as mobile applications. He observes that while OpenAI initially released an app framework, its full potential wasn’t immediately realized. Jotform’s own ChatGPT app quickly gained traction with 5,000 daily uses, showcasing the power of integrating diverse products within a conversational AI interface. Tank emphasizes the convenience of stating needs to AI for tasks like car rentals, bypassing traditional UIs. Internally developed Jotform AI Agents, initially for support, are now a product for businesses to deploy AI-powered conversational assistance across various platforms. The ultimate vision for Jotform is for users to interact with the platform entirely through natural conversation, eliminating the need for traditional UI.
Building AI Architecture and LLM Strategy
Jotform prioritizes developing its core tools and frameworks for critical functionalities, such as its AI agent product. While they utilize external Large Language Models (LLMs) for raw intelligence (primarily OpenAI and Claude), they build proprietary layers on top for specialized tools, skills, and integrations (referred to as MCP, likely standing for Multi-agent Coordination Platform or similar). They rigorously test new LLM versions to identify optimal performers based on results and cost-efficiency. Tank notes that self-hosting local models generally yields lower performance and higher costs. For code generation, developers have the freedom to choose tools like Cursor, Claude Code, or Codex, supported by internal training and “hack weeks” to foster innovation.
The Founder’s Playbook: Theory of Constraints and 50-50 Rule
Tank champions the Theory of Constraints, asserting that business success hinges on identifying and relentlessly addressing the single biggest bottleneck. He warns against “local optimization” of non-critical areas, which yield minimal impact. This bottleneck could be a weakness or a significant opportunity, as AI was for Jotform. Founders often get bogged down in administrative “busy work” or seemingly productive tasks that don’t address core constraints. He introduces the 50-50 Rule: founders should dedicate half their time to business growth (sales, marketing) and half to product or service improvement. This balance is vital because an excellent product unnoticed fails, and aggressive marketing for a poor product leads to high churn. Jotform’s early success stemmed from offering a free product, meticulously gathering user feedback, and rapidly iterating, building trust before monetization. Scaling involved slow, deliberate team growth while maintaining focus on quality and user feedback.
Ethical AI and Democratization
Tank advocates for empowering employees with AI to enhance productivity, rather than viewing AI as a job displacement threat. Jotform’s achievement of 80% AI-handled support with zero layoffs exemplifies this philosophy. As a bootstrapped company, Jotform can prioritize a 5-year strategic outlook over short-term quarterly gains, enabling this employee-centric AI adoption. He describes AI as a “democratizing force” for founders, simplifying tasks across coding, accounting, and HR, allowing them to focus on core problem-solving and value creation. He emphasizes that AI security practices align with traditional software, and AI can even be leveraged internally for security bug detection. Tank advises caution with certain open-source or local models (potentially referring to “Open Assistant” or similar, as “Open Claw” was a likely transcription error), preferring to build custom agents atop stable, enterprise-grade LLMs.
Concluding Thoughts
Tank’s primary advice for struggling businesses is to identify and resolve their fundamental bottleneck. He urges founders to strategically manage their time, allocating significant portions to product and growth, while delegating or automating administrative tasks. He references the “Eisenhower Matrix,” noting that “important but not urgent” tasks (often bottleneck resolution) are frequently postponed despite being game-changers. The “vibe coding” philosophy extends to management: focus on the ultimate objective rather than micromanagement, as both AI and empowered employees can deliver surprisingly superior solutions.