Rules of Eng-Agent
Rethinking common courtesy in a world filled with robots
As AI agents transition from theory to practice, there has been a shift in the discourse. Across many domains, there are groups celebrating the massive increase in efficiencies and scale that are possible through the use of agents. At the same time, other groups are decrying the use of agents as a source of noise that is destroying existing processes. In some situations, a single group is simultaneously extoling the virtues of their use of agents while protesting the use of agents by the party on the other side.
At the heart of these issues are unwritten informal agreements about the way things ought to be done that are quickly falling to the wayside.
Consider three examples:
Recruiting
In the past, candidates had to invest effort to apply for a job by writing a unique cover letter and filling out forms and so they would only apply for the jobs that interested them. On the other side of the equation, employers had to invest time to perform interviews and so they would carefully review resumes and only ask candidates for an interview if they made a short list. Today, both sides of the equation have changed. Candidates can use tools to automatically bulk apply to vast quantities of jobs and employers are using AI to review applications and even to perform the interviews. With AI interviews being significantly cheaper for the employer, candidates may be asked to interview for an unreasonable number of positions and there may be an exceedingly small chance of any one interview leading to a job offer. The logical next step is for candidates to use AI agents to perform the interviews with the AI interviewers. If that is not already happening, it will be soon.
Sales and Procurement
Have you received a call from an AI sales agent yet? Are you sure? I received my first one a few weeks ago and it was an interesting experience. It almost had me fooled but I caught a quirk and asked if I was speaking to a robot. The AI agent was honest enough to confirm that it was. From the perspective of many companies, these agents are an important step forward that will allow them to introduce their valuable offerings to many more potential customers. From my perspective, I was annoyed that the company had the gall to interrupt me with a phone call and did not have the decency to disclose to me that I was speaking to AI. Sales calls are expensive when they are made by real people and that has kept the situation under control as companies take care to prioritize who they are calling. What might happen to this communication channel when outbound sales calls are an order of magnitude more affordable?
Similar to the situation with recruiting, this is a problem that will run in both directions. What if I want to buy a new pickup truck and so I ask my agent to place phone calls to every dealership within a 600 mile radius and to go back and forth in negotiations to get me the best price? What if every salesperson at every dealer starts receiving dozens or hundreds of these calls every day and only a tiny fraction lead to new business?
Customer Service
It is no secret that companies are moving aggressively to deploy AI to replace customer support chat and phone agents but the other side of the equation is equally interesting. I read a story recently of a customer that was using an AI agent to act on their behalf to close their account. The company has a policy against working with agents and so the request was denied. However, the AI agent went on to make over 250 repeated attempts. Clearly there can be challenges for both customers and companies when agents are only being used by one side.
The Common Challenge
The potential for AI agents to simplify our businesses and our lives is significant. I look forward to the day where my agent is able to communicate to agents from a wide variety of businesses to save me time and ultimately to generate better outcomes. However, these three examples highlight a common challenge during this transition phase. Time is the most value commodity for living breathing human beings. There is no amount of wealth and no technology that can buy us more time. Across society, we have many processes that function today because they require both sides to make an investment of that finite valuable resource. AI agents don’t face that constraint. Their ability to invest time is limitless and that can cause existing processes to break down.
Rules of Eng-Agent
With that in mind, I propose the following “Rules of Eng-Agent:”
If AI agents are used, it must always be disclosed. No person should be mislead into thinking they are dealing with another person if they are actually dealing with an AI agent.
In any process that a person or corporation can not reasonably avoid, if one side uses an AI agent, the other side must be allowed to use an AI agent as well.
With reasonable exceptions, any person or corporation should be able to opt-out of communication from AI agents.
What do you think? Are you already using agents or interacting with agents in your day to day?
Cheers,
Steve