William Jones
William Jones

Across industries, professionals spend countless hours on repetitive tasks across multiple websites. From data entry and lead generation to content management and market research, these mundane activities consume valuable time that could be better spent on strategic initiatives. While automation tools exist, they've historically been complex tools wielded primarily by technical specialists—until now.

According to tech innovator Raj Jaiswal, founder and CTO of Neyon Labs where he is developing the AI automation platform Witrium, artificial intelligence is fundamentally changing who can create and maintain web automations, potentially transforming operational workflows for businesses of all sizes. With an MS from Brown University and a decade of software engineering experience—including senior engineering roles involving leading projects and teams at companies like Gopuff and Modivcare and serving as co-founder and CTO for his previous last-mile logistics startup Fetchum—Jaiswal's insights are grounded in practical application.

"Traditional web automation has long been plagued by fragility," Jaiswal explains. "The conventional approach requires programming skills and patience. Writing scripts to interact with websites is complex, painstaking, and time-consuming. Worse, these scripts break when websites make even minor interface updates."

This observation stems from direct, hands-on experience. While scaling an e-commerce startup he was previously working on, his team wrote hundreds of specialized scripts to monitor product and stock availability across retailer websites. "Keeping those automations alive was a constant game of cat and mouse," he recalls. "A minor page change could break dozens of jobs overnight, forcing developers back to debugging – an equally painstaking process. This frustration was a key driver in founding Witrium, seeking a more resilient approach."

"The existing paradigm requires specialized code that's difficult to create and maintain," Jaiswal notes. "This reliance on hard-coding specific element selectors and hoping nothing shifts simply doesn't scale in today's dynamic digital landscape. The resulting technical debt kept powerful automation out of reach for many non-technical users."

The AI Difference

The fundamental shift happening now, Jaiswal argues, is that AI can interpret natural language instructions and adapt to changing web environments unlike traditional automation.

"We're witnessing a profound transformation in human-computer interaction," he says. "For decades, people needed to learn programming languages to instruct computers precisely. Now, AI, particularly large language models (LLMs), allows computers to understand human language with accuracy. It can translate an everyday sentence—like 'Go to the pricing page and capture the current prices'—into the low-level actions a browser must execute."

This capability enables a fundamentally different approach to web automation. Instead of brittle code that maps to specific element identifiers on a webpage, AI-powered tools can understand the context and purpose of page elements, making them more resilient to website changes.

The implications extend beyond technical convenience. For enterprises struggling to fill software development vacancies, enabling non-engineers to automate their own tasks is attractive. Marketing teams can gather competitive intelligence more efficiently. Finance can streamline reporting. Organizations facing labor shortages can multiply workforce effectiveness.

Practical Applications and Implementation Challenges

Democratized web automation applications span nearly every industry: HR automating candidate sourcing; sales synchronizing CRM data; operations monitoring inventory across suppliers without manual checks. What makes these use cases powerful is they may no longer require dedicated technical resources.

"When non-technical users can create their own automations through natural language, you remove the bottleneck," Jaiswal points out. "The person who understands the business process can now automate it directly."

This shift mirrors previous technology democratizations. Spreadsheet software gave financial analysis capabilities to non-mathematicians. Website builders allowed non-designers to create professional-looking sites. Now, AI puts sophisticated web automation within reach of non-programmers.

Despite the promise, Jaiswal acknowledges hurdles. Trust remains significant. He cautions against viewing all AI tools as monolithic, highlighting a pitfall with some end-to-end solutions functioning as opaque 'black boxes.'

"When automation runs entirely behind the scenes without clear visibility, diagnosing failures becomes difficult," he observes. "Furthermore, many purely end-to-end AI solutions lack the robustness required for reliable workflows. The underlying technology often works probabilistically, making it hard to guarantee it will perform the same action consistently. This unpredictability can erode trust, especially when used for critical business processes."

Instead, Jaiswal advocates for an approach centered on reliability and user control through iterative development – a philosophy central to his work with Witrium. "The most effective AI automation allows users to build, test, and refine workflows step-by-step, in real time," he suggests. "Observing the AI execute each instruction and having the ability to make immediate adjustments, like we enable with Witrium's natural language interface alongside a live browser, is crucial for building robust automations. This provides necessary guardrails and improves success rates compared to both fragile scripts and end-to-end AI agents."

Data security presents another challenge, requiring proper governance around automation tools. Finally, there's the cultural shift. "Companies need to reimagine workflows," Jaiswal advises. "This isn't about replacing jobs but elevating human work by removing mundane tasks."

The Future Landscape

Looking ahead, Jaiswal sees AI-powered automation continuing to evolve. First, these tools will become increasingly context-aware, understanding not just website structure but also business processes and compliance requirements.

"As these systems gain more context, they'll become partners in optimization," he predicts. "They'll suggest improvements, not just perform tasks."

Second, cross-platform integration will deepen. While current automations can work across different websites, future systems will more seamlessly bridge proprietary platforms that don't have formal integrations.

Third, collective intelligence will accelerate. "When thousands of users are creating automations for similar tasks, the underlying systems learn from all these examples," Jaiswal explains. "This creates a virtuous cycle where automation capabilities improve for everyone."

Broader Implications

The democratization of web automation signals a broader trend: making sophisticated digital capabilities accessible. "We're lowering the barrier between human intention and digital execution," Jaiswal concludes. "It's empowering non-programmers to interact with digital systems precisely. When everyday users can effectively 'converse' with digital systems, we unlock tremendous potential. Millions of these micro-automations add up to macro-level productivity gains."

For organizations, embracing these paradigms isn't just about efficiency—it's about empowering their workforce to focus on higher-value work. In this vision, technology truly serves as a force multiplier for human creativity and problem-solving.