Communication And Collaboration: The Next Frontier Of AI

Forbes

4 days ago

Anil Pantangi is an award-winning AI and product leader who has driven impactful initiatives across Fortune 500 firms.
Currently, AI is at a stage similar to where the internet was in the early 1990s. The building blocks are here—powerful models, real business use cases and innovative startups. However, none of it is connected in a meaningful way yet.
Think back to those early days of the web. Websites were clunky. Browsers didn't agree on how to display them. Developers worked in isolation. It wasn't until shared languages like HTML and protocols like HTTP arrived that the web truly took off. Those standards transformed the internet from a scattered experiment into the global system we depend on today.
AI is facing that same turning point. Every company is building its own agents, copilots or chat interfaces. SaaS vendors are adding GenAI to their products. Enterprises are launching internal AI platforms to automate tasks, summarize reports and generate content. What we're not building is a way for these agents to communicate with each other coherently, securely or responsibly.
Right now, Salesforce's assistant doesn't meaningfully collaborate with Microsoft Copilot or your finance team's enterprise resource planning system. Each operates in its bubble, optimized for a single workflow, tool or department. Instead of unlocking intelligence across businesses, we're creating intelligent silos. This fragmentation is a significant obstacle to productivity, profitability and long-term operational success.
I've seen this firsthand in industries like telecom, HR and enterprise platforms. One of the biggest challenges in telecom wasn't network speed or device compatibility—it was systems that couldn't coordinate. IoT sounded like the future, but connected vehicles and smart grids struggled because carriers didn't share consistent standards for data or messaging. It took years before initiatives like the GSMA's eSIM brought common ground. Until then, every integration was a fragile and expensive one-off effort.
Now, we have an opportunity to solve that same coordination challenge at a much larger scale. AI agents will soon be everywhere—managing workflows, guiding customers, assisting product teams, even advising on compliance. Without shared frameworks, each of these agents will operate with a narrow view of the world. They won't be able to collaborate or build on each other's work.
The same pattern shows up in HR technology. Enterprises are investing millions into employee learning platforms, internal mobility programs and skills development systems. Employees often find themselves stuck between disconnected systems, wondering why their work in one part of the organization doesn't translate into meaningful growth elsewhere. GenAI could exacerbate this issue if each system generates its own independent recommendations. Without coordination, companies will soon have competing AI-generated guidance for the same person.
Enterprise SaaS is heading in the same direction. Today's tools are filled with isolated AI features. One tool drafts marketing copy, another predicts revenue, and another generates sales materials—but they don't share context or decisions. Teams won't know which tool has the correct version or whether critical steps are missing. Instead of making work easier, these systems risk adding new layers of confusion.
However, this is also the moment of possibility. We don't just need more intelligent agents. We need better foundations.
The next wave of AI innovation will depend on the development of shared protocols. Not just APIs or vendor integrations. We need true frameworks for agents to exchange context, share reasoning and work together to achieve outcomes, not just isolated outputs. Imagine something like TCP/IP but designed for intelligent systems—a standard protocol for communication, security and traceability that's built with global business in mind.
This concept is moving from theory to practice. Treating LLMs as a new kind of operating system can allow your specialized agents to run on them like applications. It's a decisive step, but it only addresses part of the problem. Enterprises don't just need individual operating systems—they need a universal way for agents across industries to collaborate. It's time to build something like an agent collaboration protocol, with open standards supported by both the public and private sectors.
Of course, protocols alone aren't enough. These systems also need embedded ethical frameworks, permission controls and security models. Otherwise, we risk intelligent agents working at cross purposes, leaking sensitive data or making decisions that don't align with business strategy. Governance should be an integral part of the design, not added as an afterthought.
Enterprise product leaders, CIOs and technology executives should lead this. It's time to stop focusing only on features and start focusing on foundations. This is what product leadership looks like in the next era. Those who help build these protocols will not only lead their industries—they will shape the future of global business.
The urgency is real. IDC projected in 2023 that the IoT market alone would surpass $1 trillion by 2026. GenAI is already becoming standard in every major enterprise software suite. Without shared frameworks, this next phase of growth will slow under the weight of duplicated effort and uncoordinated decision making. Enterprises that lead in shaping these foundations will define how industries operate for years to come.
We need a common language for intelligent agents—not just for convenience but for resilience, growth and trust. We've built the agents. Now, it's time to build the roads that connect them.
Intelligent agents won't change the world alone, but if we help them work together, they just might.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?