Yonathan Cohen Explains How to Build AI Systems by Connecting MCPs

In a recent LinkedIn post, Yonathan Cohen discusses the strategic advantage of connecting Modular Cognitive Processes (MCPs) to create powerful AI systems, rather than viewing them as standalone features. Cohen emphasizes that the true potential of these tools lies in their interoperability, enabling them to work together to automate complex workflows.

Cohen challenges the common perception that MCPs are primarily linked to specific AI models like Claude. Instead, he advocates for a more integrated approach. As Cohen states:

“Most people connect MCPs to Claude. The power is connecting them to each other.”

This perspective shifts the focus from individual tool capabilities to the emergent power of their synergy. Cohen argues that by pairing MCPs, businesses can create automated processes that handle tasks end-to-end, significantly boosting efficiency.

Building End-to-End Workflows with MCPs

Yonathan Cohen outlines ten specific examples of how different MCPs can be integrated to achieve distinct business outcomes. These examples illustrate a clear path from raw data or initial contact to a completed, enriched, or analyzed output. Cohen’s approach suggests a systematic way to leverage AI for practical business applications.

One key integration highlighted by Cohen is the combination of Granola and FullEnrich. He explains the practical application:

“Granola + FullEnrich → names from a call, straight to pipeline”

This example demonstrates how an MCP that captures call data can be seamlessly linked with a data enrichment tool to directly feed qualified leads into a sales pipeline. This eliminates manual data entry and speeds up the sales cycle.

From Calls to Pipeline and Pitches

Cohen further elaborates on the potential of connecting MCPs for sales and marketing enablement. For instance, the combination of Granola and Gamma is presented as a way to automate the creation of call recaps.

“Granola + Gamma → a call to a recap deck on its own”

According to Cohen, this integration transforms a completed call into a ready-to-use summary deck without manual intervention. Similarly, pairing FullEnrich with Gamma aims to identify and research a buyer, then construct a tailored pitch deck.

Enhancing CRM and Data Enrichment

The post also delves into how MCPs can significantly improve Customer Relationship Management (CRM) data. Yonathan Cohen points out the utility of connecting FullEnrich with Attio:

“FullEnrich + Attio → new contact, enriched inside your CRM”

This integration ensures that when a new contact is identified, their information is automatically enriched and updated within the CRM system. Cohen also suggests that combining Attio with Bright Data can lead to comprehensive company intelligence being added directly to a contact or account record.

The Systemic Advantage of Integrated MCPs

Throughout his post, Cohen consistently returns to the theme that individual MCPs are merely features, but when interconnected, they form powerful systems. This systemic view is crucial for businesses looking to implement AI effectively.

Yonathan Cohen argues that by understanding these connections, organizations can move beyond using AI tools in isolation and instead build robust, automated workflows that drive significant business value. The ability to wire together these modular processes is, in his view, the key to unlocking the full potential of modern AI capabilities.