This interview is with Alix Gallardo, Co-founder & CPO at Invent.

Alix Gallardo, Co-founder & CPO, Invent

Can you introduce yourself and tell us about your current role working with AI assistants and intelligent automation? What specific problems are you solving for organizations today?

Hi, I’m Alix Gallardo, co-founder and CPO at Invent. I’m a passionate builder and product strategist with a strong focus on UX and user empowerment. At Invent my work centers on building AI assistants, chatbots, and voice agents that make advanced automation accessible and truly helpful for organizations of any size.

What path led you to specialize in AI assistants and smart automation? Was there a pivotal moment or project that shifted your focus to this area?

The pivotal moment for us was seeing just how much friction people face when they have to leave their favorite app, open a different channel, or try to figure out a support process on their own. It became clear that the future isn’t just digital, it’s conversational. People expect to engage in natural, guided conversations, whether that’s through a familiar WhatsApp chat, a bubble floating on a website, or directly within the app they’re already using.This vision is what drives us at Invent. We believe support and automation should come to the user, seamlessly blending into their flow, whether they’re browsing a site, messaging on WhatsApp, or exploring a new product. Our platform is designed so organizations can instantly meet their users wherever they are and offer help, onboarding, and powerful automation through friendly, conversational assistants.

You’ve mentioned that companies often jump to training custom models before validating the problem. When a company approaches you wanting to implement an AI assistant, what’s the first question you ask them to ensure they’re starting on the right foot?

Definitely, it’s rarely about just asking a single question. To design an effective AI assistant, we need to understand the whole context of the business, the pain, the expectation, the current user flow, the roles or areas who are facing the friction/challenge/pain currently.

We’ll start with: “Help me see the big picture. What’s really happening on the ground?” And then we’ll dive deeper with questions like:

• What department or area are we focusing on, support, sales, onboarding, operations, something else?
• How many people or processes are currently involved in this workflow? What do these processes look like day-to-day?
• What’s your operational stack? Are you using CRMs, ticketing systems, channels, what tools are in play, and where are things breaking down?
• How does the challenge you’re facing affect your ability to scale or handle more volume?
• What’s your team’s experience, are they overwhelmed, frustrated, or bottlenecked?
• Can you walk me through your ideal user flow, what would a frictionless, guided experience look like for you and your customers?

Ultimately, I want to step into your shoes—seeing not only where the pain is, but the entire journey, the systems and people involved, and the outcomes you actually need. Only then can we figure out if AI is the right solution, and if so, design something that truly delivers results right where it matters.

What’s one AI assistant implementation you’ve seen fail or underperform, and what specific lesson did you take away from that experience that changed how you approach projects now?

Sure. A long time ago we built an assistant to help a company automate paystub analysis. The company gave us a bunch of separate statements about how numbers should be calculated, like “add these fields together,” or “subtract this amount”, but they never provided a single, fully worked-out example on an actual paystub. There wasn’t one concrete formula, just disconnected elements.

When we rolled out a demo of the assistant, we ran into a challenge: the results weren’t matching what their human agents were doing. After sitting down with the agents, we realized why, each agent had invented their own version of the formula, filling in the gaps however made sense to them.

The big lesson for me was, before automating, always get everyone on the same page first. Now I always ask things like:

• How many people actually do this process?
• Do they all use the exact same guide/formula/template, or is everyone doing things differently? Can we see one, end-to-end example applied step-by-step on a real case?

Until we have that common ground and clarity, we don’t start building, otherwise, you risk just automating confusion.

From your experience, what’s the most underrated factor in getting employees to actually adopt and trust AI assistants in their daily workflow? What’s one tactic that’s worked surprisingly well for you?

Honestly, the most underrated factor is simple: people don’t adopt what they don’t help build. In our experience, real buy-in happens when employees, especially those in key operational roles, are actively involved in shaping how the AI assistant works for them. We always start by building the right team (usually in operations) and running a quick benchmark to identify any perception gaps between leadership’s expectations and employee reality.

Make employees co-designers, inviting them to tweak the assistant’s instructions in natural language and see instant impact on AI adoption. Collaboration is the best approach for adoption: add your team to the platform and co-design together. This not only improves adoption but builds trust.

It’s also one of the biggest advantages that Invent offers: bridging the gap between business and development teams by enabling both to collaborate and contribute directly on the same platform. We recently shared more about this in our latest article on employee-centricity and AI adoption read here, drawing on insights from Harvard Business Review, BCG, and Columbia Business School. The evidence is clear: putting employees at the center isn’t just good for morale, it’s ESSENTIAL for AI success.

Thanks for sharing your knowledge and expertise. Is there anything else you’d like to add?

The teams and companies that focus on building with their employees, and really embrace co-design, are going to see the biggest benefits and the strongest adoption. And honestly, that’s what motivates me: making sure AI is something people want to use, not just something they have to use. Let’s have the best of both worlds: Human expertise and AI efficiency.  Just that this is a moving space, and the most important thing is to stay curious, experiment in small, safe ways, and keep humans at the center of every AI decision. If anyone listening wants to explore concrete use cases or needs a sanity check on their AI roadmap, I’m always happy to chat.