AI agents will start by aiding us like colleagues; wild thought: they may become our customers & competitors.
What are AI agents?
They are autonomous applications designed to complete tasks and solve problems on the web and apps, acting on behalf of users with minimal intervention.
Agents: emerging tech in startups and enterprise software
At my firm, we’ve already invested in a few, and my next Llama Lounge 14n event will feature multiple agent startups, and Salesforce and Hubspot have announced their versions.
This was my closing slide at the enterprise AI-focused conference in SF, AI Realized (thanks Daren McKelvey for intro). I gave real examples of how AI agents are currently aiding productivity through simple tasks, and how, as they become more complex, they will evolve into autonomous organizations we’ve yet to fully envision.
Future Relationships: AI Agents and Companies
- Human workers increase productivity via Agents: AI agents will handle tasks, freeing humans to focus on higher-value work. Both Microsoft’s 365 Copilot powered by OpenAI and Google’s Gemini products are demonstrating how these technologies can assist workers like an AI assistant or copilot—this is just the precursor of what’s to come.
- New AI Agent colleagues emerge in the workforce: Virtual colleagues, collaborating with humans to execute tasks, are just being rolled out. Salesforce’s Agentforce platform and HubSpot are automating tasks traditionally done by customer care, marketing, and account development reps. Level one customer care roles are being automated, and lower-level marketing and sales tasks are also being handled by startups.
- Your company will interact with customer AI Agents: We’ll be dealing with AI agents acting on behalf of customers. We’ve invested in AI agent startups that enable consumers to conduct web-based shopping tasks online, making decisions on behalf of humans. I’ve written about how AI agents will make decisions for consumers.
- Agent-run organizations will be your customer and even competitor: Fully autonomous, agent-run entities may someday become both your customers and competitors. The logic path: autonomous agents could offer services, generate revenue (from humans or others agents), self-improve, and even replicate themselves by hiring or building more agents or event humans. Recently, we invested in a payment protocol that enables agents to hire other agents to complete tasks.
Seven year ago, I gave a TED talk at the Frankfurt Auto Show on behalf of BMW, about how self driving cars could become autonomous organizations, as the self-driving cars generated revenue (from driving humans, delivering packages), they would earn enough to upgrade themselves and eventually buy additional vehicles, a form of reproduction.
