Q&A with Omar Torresola

Can you give us a quick background on yourself? And what are you passionate about outside of work?

I’m a product builder with 10+ years spanning fintech, e-commerce, health tech, and payments; I recently managed Upgrade’s chatbot initiative across its multimillion-MAU platform. Once work is over, my evenings are for building Star Wars LEGOs with my 6-year-old, Nintendo-ing with my 4-year-old, and date nights with my wife whenever we can get  a babysitter.

When and how did you discover chatbots and Gen AI?

I first dove in during a 2018 Galvanize meetup as part of the downtown Phoenix community. I saw a demo of Botco.ai and rolled the tech into my company at the time, STC Health, to handle after-hours support.

When did you realize Gen AI could solve real problems at your company?

The moment I realized very simple inquiries, which are basically Help Center articles, can quickly be spun up in a different medium that tailors to a consumer’s question. This frees up phone queues and slashes agent load. So for a while I’ve known that this technology was the right lever for cost and customer satisfaction.

How was the process of integrating Gen AI technologies like Botco.ai into your systems and workflows?

Because these technologies can tackle three types of problems for us — saving time, optimizing resources, and reducing risk — it is very easy to get momentum toward integration. You share certain data and a response comes back that is refined to your workflow. The detection of a forged document for identity verification purposes is a great example. You pass a file with a picture and ask: What is the likelihood that this picture is fake and what is your confidence? The response tackles all three problems, and it is up to a human to review. That is a powerful assistant.

How have you incorporated Gen AI chatbots into customer engagement, marketing, or operations?

I spearheaded a customer-facing bot launch for payment flows, and general account information, and have some plans for a marketing bot that assists with leads on our website 24/7. The bot offers guidance while we still make live help an option, thus improving application completion. Even a modest uptick in conversions translates to substantial growth across products.

How have Gen AI chatbots helped you understand your customers better?

For Product Managers it’s a well-known challenge to review anecdotal feedback at scale. Because delegating this work is a sub-optimal choice, I spent some time using this tool to help bridge the gap. For example, in Google Sheets, you can use a function called AI. I’ve used it to tag labels of every customer survey with a word or two, allowing me to spot themes and patterns and deep dive as necessary. To be clear, I was doing this without AI because it’s part of the job. Now I’m simply getting more of my time back.

Where do you see the next application of Gen AI chatbots in your company/industry?

For customers, the industry will flood with multilingual Gen AI outbound voice agents. People will get used to this as a replacement to the traditional phone system: “Press 1 for this, Press 2 for that.” For organizations, I think co-pilots that listen to calls, review account information, and leverage a user interface can speed up training for anyone that is focused on providing a great customer experience. The assistant will be able to evaluate and surface the next best action onscreen on a real-time basis.

What results or ROI can you share?

Containment holds at around 33%, bot-only CSAT is 4.5/5, and average chat handle time fell.

What advice would you give leaders considering Gen AI chatbot implementation?

Start small, pick a use case, and layer in voice early. During my 25-minute commute, I open ChatGPT’s voice mode and outline a single high-volume use case along with its key success metric. I just talk normally as if I’m talking with an engineer or an executive, and it makes the drive to work a bit less of a routine. Once I’m at my desk, I refine the auto-saved transcript into a one-pager specification for the team. This turns my travel time into planning time and lets me plan other initiatives sooner.

How is Gen AI transforming your industry, and what improvements do you want to see?

I’d like to see a one-click trust badge. Banks need an out-of-the-box toolkit that stress-tests an AI for accuracy, fairness, and privacy, then prints a regulator-ready “seal of approval” without weeks of custom work. Additionally, I’d like to see more clear and scalable rules. We need state legislatures to pass consistent AI standards that federal agencies can adopt. Without that harmony, I think rolling the same experience across all 50 states is a compliance maze. And finally, I’d like to see more privacy baked in. Sensitive data such as card numbers and SSNs should be automatically masked before it even reaches the model, not protected by extra layers bolted on later.