Introduction

EVO Bank’s virtual assistant had grown unpredictable. With 535 intents and close to 5,000 training phrases feeding it, nobody could reliably tell where a conversation would end up, and when it failed, fallback was effectively random: most of the time it didn’t trigger at all, leaving the user with no way forward.

I owned conversational design for the rebuild, from intent architecture through the assistant’s tone, leading a team of Spanish-language linguists and working with EVO’s marketing team on a three-month deadline.

Framing

The core decision was to rebuild from zero rather than patch the existing model intent by intent. We started from the simplest compositions and worked up, confirming at every step that a change didn’t break a flow or leave the user at a dead end, with each intent’s purpose and coverage clear before adding the next layer of complexity on top of it.

The alternative we discarded was incremental cleanup of the existing 535-intent model, the same direction that had produced an assistant nobody could predict in the first place.

Crafting

Rebuilding intents and fallback

Every intent’s purpose and coverage had to be explicit before it shipped. The previous model let fallback trigger at random, or not at all, so flows were rebuilt to guarantee a failed match always resolved into a clear next step for the user, instead of leaving the conversation without an answer.

Tone and personality

Personality was built to match EVO’s own web copy, not a generic assistant voice, and validated directly with the marketing team so the tone stayed consistent with what customers already read on the site.

Scope and language

The assistant launched in Spanish, on EVO’s web channel, the surface where the rebuilt model and the marketing-aligned personality could be validated with real usage before expanding further.

Outcome

Training phrases dropped from roughly 5,000 to 586, and intents from 535 to 491, a smaller, more deliberate model. Despite doing less, the assistant got the right answer 98% of the time, up from 67% before the rebuild.

What started as a three-month, fixed-deadline engagement turned into a long-term service contract with a dedicated team, and EVO expanded the assistant beyond its original web channel.

The work was demanding, and ultimately rewarding: it produced a repeatable set of guidelines and a client workshop for auditing an assistant’s state and defining lines of action before a rebuild. That method made the analysis work accessible to the rest of the team, instead of depending on one person to run it.