The Limits to Personal AI

At the start of the year, OpenClaw was the hottest AI project around. It became the first personal AI project to hit mainstream consciousness. Its advocates pitched it as an open source personal assistant straight out of science fiction until the underlying security flaws caught up with it. Now we're seeing the same moment happen again following the release of Meta Muse and ChatGPT Dots. Both are pitched as always on AI assistants with the capability to interact with the broader internet on behalf of the user.

Muse and Dots

The OpenClaw hype cycle only lasted a few weeks until its creator was hired by OpenAI. This one will last longer. Muse is probably the first Meta AI product to be commercially successful since they closed off their open source Llama AI models last year. OpenClaw was too technical for most average users, which limited its reach. Meanwhile, Muse has the Facebook and Instagram marketing engine behind it, allowing it to briefly top the app store download charts a few days after launch.

The capabilities of the new wave of personal AI agents are much the same as OpenClaw. However, they do open up new AI capabilities to the average user. They go beyond planning itineraries and drafting messages, marking the latest attempt by AI labs to become ecommerce intermediaries. The new wave of personal AI can buy products on behalf of users as well as make purchase recommendations. In doing so, it follows on from OpenAI’s failed browser efforts, which had the same objective.

Payment Restrictions

Amazon have been quick to react accordingly, banning Muse from their checkout process just as they banned AI browsers last year. It's these kinds of security controls which will ultimately limit the real world value of automated purchasing. Buying agents have been a hot topic in marketing circles for a while, but the hype has rarely been matched by actual practical experience. Automated purchasing is hardly a new concept. Smart fridges have been touting the same concept for years. Amazon has Rufus, its own recommendation chatbot, which can even complete the checkout process for you.

The risk is always the same. Who do you trust with your credit card? Buying low cost disposable widgets is one thing; a high ticket enterprise software purchase is quite another. The scope for AI-led purchase decisions is limited by risk tolerance, as well as a frequent need for joint decision making. Sure, AI can be used for creating shortlists, but then you don't need an agent for that. A basic ChatGPT query can research a purchase decision just as effectively as a custom deployed enterprise procurement agent.

The Marketing Context

In a business context, the real benefit of the agent comes after the purchase decision has been made. It can fill out the approval paperwork, support the pricing negotiations and generate the PO for you. Trouble is, technology already exists to do all of those things. AI can make them better (and based on my agency experience, it would be very difficult to make enterprise purchasing much worse). Enterprise procurement is a multi-million dollar software market ripe for AI-driven disruption, just as Clay has disrupted the marketing data world.

Ultimately, buying agents and agentic procurement are not the revolution people claim them to be. In a business context, they are classic enterprise automation tasks, where AI takes an assistive role to guide decision making and surface missing information. None of those tasks are relevant in a marketing context. They don't change the marketing message or the sales process. Sure, a moderate level of consumer eCommerce purchasing may become AI-driven, but that only changes the final checkout step. Otherwise, AI purchasing is merely another description for the army of AI assistants already disrupting corporate web traffic everywhere.

Clay Becomes Reusable

Clay has been the biggest martech success story of the AI era. However, its rapid growth has led to governance challenges for many users. At the user conference last week, they announced a new workflows capability, which attempts to address this issue and the related discoverability challenges with the platform.

Operationalising ABM in the Real World: Your Step-by-Step Guide

Finding it difficult to navigate the complexities of ABM. A new eGuide from CRMT Digital cuts through ABM hype, explaining how to implement and run an ABM program in practical terms. Boost ROI with practical approaches to account-based targeting, personalisation and engagement.

Marketing Operations Roadmap Matrix

A Marketing Operations Roadmap is an essential strategic tool that aligns marketing priorities with your overall business goals. Any roadmap begins with understanding where your marketing organisation is right now, and then defining where you want to get to. This free Marketing Operations Roadmap Matrix from CRMT Digital allows you to do exactly that.