Published
Being on a fast-moving team, in an accelerated environment, in the age of AI is a challenge for both the designer and the stakeholders. Having immediacy means that the decisions that need to be made, at any level, must be backed by data.
I saw this in my current job. What happens after implementing a conversational AI agent in a marketing team’s acquisition system? The key (to sum it up, haha) is judgment.
Challenges that leave the visual behind
First, what we used to know as the classic form now lives in a non-deterministic way. If before a lead arrived with defined data and was attended to and advised by a sales rep, now a lead with no data is advised by an AI. The only way in for that prospect is a single input. What happens with that input? How do we validate the minimum amount of data to ask leads for in a first instance? How do we ensure a good user experience when the conversation starts with an LLM?
User culture is decisive
The first discovery when we incorporated this new acquisition method was seeing how user preferences emerge according to their culture. Chilean or Mexican users have no problem talking to a virtual agent, but, for example, we saw how the lead-to-demo conversion rate suffered in markets like Argentina or Colombia, where users clearly prefer more human attention. This opens up a range of possibilities: Why not have a mixed acquisition model? Do we gain something if, in certain countries, we mix consultative selling with AI agents? How do we continue the user experience at a digital level?
This is the first time I’ve come across this kind of paradigm, and it leaves me with more questions than answers. But what I have learned is to appreciate the value of a mixed approach. Today we can fine-tune things with as much detail as we need, and that is precisely what we’ve been testing, to the point of editing each agent and each digital experience to fit each business.
The importance of judgment in user experience
I think it’s important not to become biased in this era of AI. Now that we have the channels at the cutting edge, what’s left are the decisions. Being able to adjust a knob at a global level, whether in a design system or in a knowledge base, should always be done through an experiment, a hypothesis, a case study. I’ve been surprised by how many times very important stakeholders don’t know their own data and make biased decisions, or decide because “someone told them.” A colleague in the industry used to laugh at this paradigm. He referred to it as “guata driven” (or gut-driven, in English) instead of “data driven”. That kind of behavior distorts the entire discovery process and diminishes the final work. Within that, the challenge for the designer, the analyst, and the marketer lies in moving forward with quantifiable experiences.
Well, that’s it for now. More questions than answers, but that’s half the fun. Less gut, more data. See you!