Marcelo Lombardi

1994

Natural language processing

I started working with AI in 1994, in natural language processing. We used tokenization, word counts and statistical models on transcribed interviews tagged with demographics to identify consumer preferences. It bridged the gap between quantitative and qualitative market research. By 1998 I was building a chatbot to interview customer service reps, classify their answers, and identify their strengths and areas for improvement.

Neural networks were there all along, and I experimented with them, but the computing power wasn’t. The real challenge was explaining what was going on under the hood. With my statistics background, I knew the assumptions wouldn’t hold, so I’d describe it as an engineering solution, not a scientific one. Even with plain statistics, the more variables you add, the harder the results are to interpret. With neural networks it’s harder still, and research on it continues today.

2000s

Adapting

By the early 2000s, after the dot-com crash, AI had become a dirty word, and I had to adapt. Since then I’ve worked as a software architect and developer on medical software, SaaS development, performance optimization, e-commerce, security and SEO across a variety of industries.

Today

Strategy and self-reliance

Today I build AI automations and AI-era visibility strategies through Strategy Voilà and AIO Presence.

As an AI Strategist, my goal with every client is self-reliance. AI is becoming as basic as a word processor, and I want you running it on your own, with your data and your privacy in your hands.

As a writer and content creator, I write about AI, public policy, privacy and competitiveness, sometimes with a satirical edge.