Rethinking Quality

A Conversation with Dror Elad, CEO of Provengo

As a software CTO, I’ve often found myself frustrated by how broken the software development process can feel. Misaligned teams, unclear requirements, and endless bugs seem like inevitable parts of the job. A large portion of my role involves bridging the communication gap between business, developers, and product teams - an exhausting but necessary task. But does it really have to be this way?

I recently spoke with Dror Elad, CEO of Provengo, whose platform claims to ensure that software products are 100% aligned with business requirements. Provengo tackles many of the challenges I’ve faced by aiming to streamline communication, automate testing, and improve the overall quality of software delivery. In our conversation, I wanted to dig into how Provengo actually works and whether it can deliver on its ambitious promises.

Aron: Over the years, I’ve been frustrated by how manual and inefficient test creation and maintenance can be. Can you give examples of how Provengo automates these processes for customers and whether it truly replaces the need for manual effort?

A: Provengo takes the hassle out of testing by automatically generating a complete, executable test plan from a formal description of your requirements. We don’t just create the test plan for you: we also visualize it, check your requirements against regulations or company policies, keep the plan up-to-date as your system evolves, and even execute it directly. If you prefer, we can also generate test scripts that integrate seamlessly with your automation framework.

While we don’t claim to eliminate all manual work, we make a huge dent in the effort required. On average, our customers report cutting the time spent writing test plans by about 50% and reducing maintenance effort by an impressive 80%.

Aron: I want to understand the nuts and bolts of Provengo’s technology. How does it transform something as abstract as a user story into actionable test scenarios? What’s happening behind the scenes?

A: Provengo introduces a new approach to Formal Modeling, by taking inspiration from nature's problem-solving methods. Using evolutionary algorithms, we mimic natural selection: trying out various possibilities, selecting the most effective ones, and refining them iteratively to create optimized solutions.

We enable users to define system behaviors, conditions, rules, and regulations, encompassing elements like user stories, data points, actionable steps, and business rules. These inputs become "active" elements, capable of interacting with one another, adapting to relevant rules, and responding to personalization data points. From there, we generate all possible pathways the system might follow, mapping out a comprehensive range of scenarios.

Once the pathways are mapped, we apply our evolutionary algorithms to design a test plan tailored to the user’s needs. We even provide a performance grade for the test plan, giving users the option to implement it as is or further refine it with our assistance. This unique blend of natural inspiration and advanced modeling turns system requirements into dynamic, actionable solutions.

Aron: Many software teams struggle with balancing speed with quality. What advice would you give to a manager aiming to reduce defects without sacrificing timelines, within an existing software product?

A: One of the best ways to boost development speed is through automation - the more you automate, the faster each cycle can run. But there’s a catch: automation takes time to set up, and the more you automate, the harder it becomes to maintain.

That’s where we come in. We eliminate most of the maintenance headaches, solving one of the biggest challenges to scaling automation. Plus, since our automation cycles are shorter (the scripts are built as soon as the requirements are entered), you can easily incorporate automation into even the shortest sprints.

Aron: Amazing. Now, a question for the entrepreneurs among our readers: Building B2D (business-to-developer) products comes with its own unique challenges. What advice would you give to those looking to create tools that truly connect with developers?

A: The most important thing is to really understand your users: how they work, what they need, and what frustrates them. Developers are incredibly specific in their requirements. They want tools that solve their problems, not generic solutions. Focus on addressing a clear, well-defined pain point, and make sure your solution is straightforward to implement. Most importantly, listen to their feedback. When developers point out issues or suggest improvements, take that seriously. It’s your best guide for building something they’ll truly value.

Aron: With the growth of LLM-based code generation, where do you see your space evolving in the next 5–10 years? How does Provengo plan to stay ahead?

A: The rise of large language models (LLMs) and Agentic AI is transforming how users build systems and handle everyday tasks. While these technologies are impressive, they fall short when it comes to supporting dynamic processes. They often introduce hallucinations, statistical errors, and lack explainability or proper audit trails.

Agentic AI works well for small, well-defined tasks in structured environments but struggles with complex, AI-driven business processes in more dynamic or less organized settings. Code generators have definitely boosted developer productivity, but studies show the code they produce isn’t always great. Part of the issue lies with the LLMs themselves, but another part is that developers don’t always communicate their requirements clearly or thoroughly review the generated code. And let’s face it, no one enjoys diving into someone else’s messy, unclear code, right?

This is where Provengo shines. We’re built for complex systems, highly regulated environments, and continuous evolution. By enabling users to create and maintain comprehensive test plans with ease, we catch bad code and problematic implementations during testing - long before they can cause issues in production.

That’s how we stay ahead of the game and help our users do the same.