As you can see from the following image, that's a behind-the-curtains view of my current WhatsApp bot. It’s just a chat, and I am “the bot” on the other end. So I had to make that decision, had to do that intervention. As an administrator user, all the chats coming through the WhatsApp bot, I can reply to them. I am the bot.
And before I explain why, or on my way, we need to talk about the before that. I had bots. Not unlike many builders, I really wanted to start with the use case, go straight to the point. The bot is in place, bot functions are in place, chat flows to support users in many ways, all good. Supposedly. Supposedly, the best use of my time as a builder, too. But then, you are sensing, tic tic burst.
So, let us talk about the step back. It started to smell bad in parallel with “I can’t see what's going on.” But then, a problem for a real builder who claims to be “now on fire.” Many bugs, yes. But yeah, I can deal with that. And the bots, I see, they sometimes start doing something, and wow, let’s fix that, let’s catch up, let’s work like a founder. Energy is high. Apologies, I am here for you. The next day, we fix. It’s the break and fix rule. Yet, all that seemed to be always around that idea:
I don’t know what's happening with the end user.
So, two points for us today here, a celebration, right? But a word of caution, too. We will always fix in retrospect, when we are energized (meaning cash flow, AI cycles, human cycles, resources). While fixing is great, watch out. And second, we can’t lose touch with the user.
While Fixing is Great, Watch Out.
And from the development, the traditional development standpoint, of course, doing lean, we can be quite lean, indeed, iterative, and non-stop: we create test cases, end-to-end testing, the end-to-end experience really has a weight on our heads. It’s end-to-end, see, we care about the user.
So we measure the box, its dimensions, right? We know all about the thing; we put the box in the weather, under military-grade heating conditions, we force success with metrics we can come up with, and triangulate things. That is it. And it solves, we know, it solves to some extent.
But see, or you may feel when it’s real life, that reality is different. It’s a trip, interstellar; it was all calculated. But in the end of the day, with all that we do, we may too find ourselves working while drifting. I can explain. The job was done, we worked. The outcome-based work was done. See? I was going for outcome my friend, my manager. But it also fails. Because, again, reality is different.
You may have missed one little thing; you thought it was a mountain, but it was a wave. All can happen when you are crazy, going for outcomes while missing the point, the point you can’t see. So, to close this point, yeah, go for it, fixing is great, but watch out if you start sensing that the more you do, the ideal looks like a mirage.
We Can’t Lose Touch
The thing is that while we do a lot when we try to make “the system,” it’s not guaranteed that we can carry on a level of sensibility. With my startup chatbot, that was my case. I think I can put it, the complication, that way.
I would say that going straight towards the solution, and even if I had energy to look at all the edges, the real problem is that I was not focused, or better, I was focusing too much on the right things, but running into missing the point:
Sensibility
So that's what led me to make that tough decision. I went there, and I made what I call “the interceptor”, where all the conversations with all my users that come through. So now that it's all come through that screen, that I can see, of course, now I can see.
To recap: it is a service, my clients still know that it's a “bot service”. But for now, it's all manual. I had to take that step back. When your spaceship is not operating the way you want, when you have too many systems to look at, when you're trying to automate too many things but you keep seeing things failing, it's time to recover that sensibility, because that gives you:
Connection
It's time to connect to the people, and that's why sometimes we need to enter that mode. It’s when you become “the bot” again. Which is about “the bot” becoming “the human.”
And here, I remind ourselves of the Mechanical Turk, that old idea, which was an inspiring anchor for Amazon’s Mechanical Turk service. From the outside, it looks like a robot, it looks like a machine, but inside the box is a human.
That something, many founders know. Or better, successful founders know more, they know that that is key for growth. That is the thing we have to be, for growth, before growth. So first, I say thanks to that good old idea.
In addition to that, and to close these thoughts so you can reflect on it, I say thanks too to “Do things that don't scale” from Paul Graham. That's essentially the thing that, in my case, I had to do. I had to reconnect with the users, so to start again, from there, without losing the touch, without losing the connection, without losing the chat, without losing the whole experience.