Firing My First Bot
Written by: Jules Flesner
I’ve never held a traditional corporate management role with direct reports. Have I led teams with dozens of support and collaboration roles? Yes. Have I ever been the manager who signs the approval of someone clocking in late for a personal doctor’s appointment? No.
Instead, in recent years, I’ve had chatbot assistants “reporting” to me. And it only took a matter of weeks before I fired my first bot.
In recent years, conversational AI agents have been THE staff to ‘hire’. Having a chatbot on your website, according to flashy sales brochures, meant being a hero by saving your staff a bazillion hours of task work and covering after-hours customer inquiries.
These conversational AI agents run on a new technology (vs. earlier chatbots who hit dead ends very quickly). Newer bots can actually learn by studying every word your customers and sales staff type inside the platform. In doing this, bots get smarter over time.
Keywords: over time.
The reason I fired my first bots was simple: they were entry-level, “hired” to fill jobs way above their pay grade. The software companies behind them were basically using clients as free training grounds. (Companies have always done this with humans too, so the concept is nothing new!)
But here’s the problem: my client was paying top dollar for bots that couldn’t handle sensitive or legally risky customer conversations. Responses weren’t naturally timed or phrased. Users sniffed out the AI quickly and disengaged.
An audit discovered the website chatbot feature (a.k.a. that annoying pop-up window) had practically no engagement. An overwhelming 90% of users were actually closing or minimizing the pop-up chat pane on the website, engaging via automated email instead.
After reviewing engagement data, I found that by the time customers reached out to sales staff, they’d already consumed all the digital + automated content they wanted, and were ready for a human.
Another key takeaway of the audit was that my client was double-spending for these automated email campaigns. This was a feature their existing software already handled. (I see this too often- sales reps taking advantage of owners’ and marketing directors’ limited tech knowledge just to hit a commission.)
By firing the bot, I saved the client $30,000 a year in contracted software services they didn't need.
Don't get me wrong! I am all for using AI efficiencies in the modern business era, what we here at VD + Co. have coined Survival of the Wittest℠.
Some bots can truly perform for a business and deliver value. Agentic AI tech is opening a whole new level of productivity and salary expense savings.
But here’s the reality check: managing AI tech can take just as much bandwidth as managing and training a human, if not more. And traditionally staffed businesses are not accounting for (or, perhaps are not aware of) this workload.
I’ve accepted that I’m not cut out for the classic corporate manager gig. My skill set isn’t handling a complex human resource workload. Turns out, I’m better at managing “artificial” staff, which for now means working with product developers to teach chatbots the contextual nuances needed to talk with a potential customer who may have just left a personal doctor’s appointment.
Equal Opportunity Bot Policy: Vermillion Design is committed to fair treatment of all bots, regardless of coding language, platform, or version number. We do not discriminate based on operating system, service provider, a bot’s background or where it was trained. All bots are given equal opportunity to annoy, misfire, or genuinely assist in customer engagement moments.
Disclaimer: This post is for informational and entertainment purposes only. It is certainly not legal advice. Always consult a qualified attorney for guidance on compliance related to emerging AI technologies and HR policies.