Unlock ROI and Strengthen Website Security in the Agentic Era

As AI becomes deeply infused with business outcomes, leaders must look beyond mere prompt response visibility and toward a more complete picture of what it means to be AI ready. 

Understanding how AI is changing the very fabric of the internet can help leaders invest in the right tools while maintaining the security of their digital properties.

Key takeaways:

• Digital properties now serve two distinct audiences: humans and AI agents.
• Organizations are starting to implement autonomous AI agents that integrate with existing workflows to perform complex tasks.
• While enterprise investment in AI is high, 95% of pilots fail to deliver measurable financial returns, often due to insufficient business integration.
• AI-assisted cyberattacks have surged by 72% YoY, necessitating a transition toward unified security platforms.

The dual web experience

Regardless of industry, businesses are grappling with a major transition as we all move from a human-centric web to one built for both humans and agents. For over two decades, websites were built exclusively for the human experience. Today, we are seeing a massive leap in machine-to-machine interactions.

Visibility is moving away from the classic gateway model of search engines toward an “answer engine” model. According to McKinsey’s AI Discovery survey, half of consumers already intentionally seek out AI-powered search engines like Perplexity, Gemini, and ChatGPT to guide their buying decisions. They projected that by 2028, over $750 billion in U.S. revenue will funnel through these AI-driven discovery platforms.

This creates two distinct audiences for every web property:

• The human: Seeking experience, brand connection, and decisive answers.
• The agent: Seeking raw data, structured information, and API-like interactions to feed generative answer engines.

For any business with an online presence, preparing for the future means investing in tools to ensure your digital stack can serve both audiences without compromising data security.

From simple chatbots to agent-powered decision making

We are moving quickly into the agentic era of AI. Leaders are going beyond mere experimentation and surface-level investments, focusing on better integrating agents with existing systems to execute tasks faster. Gartner projects that by the end of this year, 40% of enterprise applications will include task-specific AI agents.

Agentic AI goes much further than simple prompt response. It is a system that can accomplish tasks with little supervision. Agents can make decisions, write code on the fly, analyze behavioral patterns, and interact with existing workflows or other agents to solve problems. While a generative AI system, like ChatGPT, creates content based on speech patterns and predicting the next best word, agents can take proactive, specific actions on behalf of their managers.

“Many people think of AI as a chatbot. It’s much, much more than that. We’re seeing now that AI is integrating into existing applications and workflows,” said Cloudflare Field CTO Daniel Kent in a recent webinar. “It’s not just to provide a prompt in response. It is how can it make decisions for me? We’re starting to see how AI can make your applications become more effective and have more benefit to the way you deal with your customers.”

404: ROI not found (yet)

While exciting, these changes aren’t yet providing the ROI many early adopters had imagined. In fact, a study from MIT revealed that despite $30–40 billion in enterprise investment, only 5% of organizations report any measurable returns. This may come down to leaders investing in tools that enhance individual productivity, like ChatGPT, instead of implementing unique systems intended to create structural change. 

“I think a shift we all have to make is thinking of AI as an execution accelerator. A lot of roles today focus on execution. With AI tools and LLMs, we’re shifting into this world of orchestration,” said WP Engine CTO Ramadass Prabhakaer.

“We are describing the outcomes we want, what problems we’re trying to solve, and then as the results come back, we are having to inspect and adapt them. It’s a shift in expectations and behavior that is going to take some adjustment, but it allows us to think more about broader problems and use AI to accelerate the time to value.”

Leaders must create a culture where experimentation, testing, and learning are celebrated for AI adoption to result in true organizational change. Fear of failure, fear of causing additional problems, and fear of losing investment in tools that are deemed not worthwhile make it impossible to harness the true power of AI tools.

“How do we measure success? It has to shift from ‘how many articles did you publish,’ or ‘how many lines of code did you return,’ and instead focus on ‘what outcome were you able to achieve,’” Prabhakar said.

The same MIT study also found that companies that implement AI tools with the help of external experts are seeing twice the success rate of those building new systems internally. So, it seems that building an internal culture in which failure is not the enemy and investing in quality partnerships may be the keys to unlocking real revenue growth with AI.

Evolving web security to defend against AI threats

With the rise of machine-driven interactions, security is about managing the sheer volume and sophistication of automated threats. AI allows bad actors to scale attacks with unprecedented speed. Total Assure reports that AI-assisted attacks increased by 72% since 2024. They attribute generative tools to a 1,265% surge in phishing attacks.

To stay resilient, organizations must evolve their security strategies:

• Traffic analysis and management: Moving beyond signature-based security to identify anomalies in how agents and humans interact with your platform, preferably by eliminating threats at the network edge.
• Implementing modern best practices: Despite the risks, only 25% to 32% of organizations have invested in modern cybersecurity practices like identity management, federated security, or zero trust architecture in the past year, according to Deloitte.
• Unified security platforms: Palo Alto Networks predicts a shift toward unifying on platforms that enable organizations to be more resilient in the face of evolving threats instead of relying on multiple vendors for ad hoc security measures.

How WP Engine supports security on the dual web

By leveraging a managed environment, businesses can automate complex web security measures behind the scenes, ensuring their tech stack is resilient against AI agents. WP Engine’s Managed Platform provides a secure foundation for business websites built on WordPress^®[1] software, and additional tools, like Global Edge Security, can harden edge security even further.

Through partnerships with industry leaders like Cloudflare, WP Engine ensures our customers’ sites are equipped with enterprise-grade protection, so they can focus on building breakthrough digital experiences while we manage the evolving threat landscape.

Want to learn more about how organizations can stay resilient, secure, and competitive amid the rapid changes brought about by the dual web? Check out our recent webinar, Future-Proof Your Digital Stack: Security in the Era of AI-Driven Threats featuring WP Engine CTO Ramadass Prabhakar and Cloudflare Field CTO Daniel Kent.