The sales process will evolve from human-to-human or human-to-agent interactions to a world where company 'buyer agents' and 'seller agents' negotiate directly. Humans will only step in for the 'final mile' to provide the ultimate sign-off after the AI has conducted the research and presented the optimal solution.

A significant portion of B2B contracts will soon be negotiated and executed by autonomous AI agents. This shift will create an entirely new class of disputes when agents err, necessitating automated, potentially on-chain, systems to resolve conflicts efficiently without human intervention.

Today's AI agents can connect but can't collaborate effectively because they lack a shared understanding of meaning. Semantic protocols are needed to enable true collaboration through grounding, conflict resolution, and negotiation, moving beyond simple message passing.

While price is a zero-sum outcome, it follows thousands of preceding tasks like information sharing and compliance checks. In these areas, both buyer and seller agents share the same incentive for speed and reduced friction. This shared ground is key to productive agent-to-agent transactions.

The conflict between Amazon and Muse validates the theory that independent agents won't succeed by executing direct commerce. Platforms are successfully forcing agents into an affiliate or advertising model, where the platform retains control and monetization, rather than allowing agents to become a new transactional layer.

Switching between AI agents is effectively costless, as users can simply feed context into a new system. This lack of a technical or network-effect moat means the primary competitive strategy becomes securing exclusive, high-cost business development deals with major distribution platforms.

MCP is not a universal solution. For single-purpose agents with a fixed, small toolset, the protocol's overhead adds unnecessary complexity. Direct, handcrafted integrations are more efficient for such focused projects, reserving MCP for multi-agent ecosystems or reusable tooling platforms where its value is maximized.

The speakers predict a divergence between personal and enterprise agents. Workplaces favor a unified corporate agent that accumulates company-wide institutional context and workflows, making isolated personal bots inefficient. Meanwhile, consumer personal agents thrive on deep personal attachment, identity, and tailored domestic tasks.

For agents to buy on users' behalf, merchants need a shared technical language to expose catalogs and process payments securely. Protocols like the one Stripe co-created with OpenAI allow merchants to sell through new AI channels without ceding the customer relationship or control over fraud.