My AI Advisers: Lessons from a Year of Expert Digital Assistants

In January, Ezra Klein reflected on how, despite recognizing AI’s enormous potential, he struggled to find a practical place for it in his daily routine. He echoed what I’ve heard from many people: Even after trying various GenAI tools, they didn’t see a strong reason to keep returning to them.

This hesitation is understandable—unlocking the genuine value of AI requires a fundamental shift in mindset. Rather than viewing tools like ChatGPT as search engines, it’s far more effective to think of them as having access to specialized assistants capable of an ever-expanding range of tasks and expertise.

Klein’s article led me to reflect on my own experience using AI tools over the past year, and what I’ve learned from having access to on-demand expertise. Here are some of the most meaningful ways I've integrated AI into my workflow and life:

Personalized AI Assistants:

Personalized Tutor: I built an AI tutor using OpenAI’s GPTs that adapts lessons to individual interests, breaks down complex concepts, and provides immediate feedback. Now, learners can explore fractions through Star Wars scenarios or understand Chain Weighted CPI via examples inspired by Taylor Swift.
Healthcare Startup Advisor: For a first-time founder, I created an AI assistant that breaks down complex topics like venture funding, provides practical business advice, and supports tasks from reviewing term sheets to drafting job descriptions. The founder described this assistance as pivotal in boosting her confidence to launch her startup.
Compassionate Medical Assistant: For a cancer patient, I developed a GPT that summarizes pathology and treatment reports, explains medical jargon in plain language, and suggests actionable follow-up questions for doctors.
SommeliAI: I created a virtual sommelier that provides tailored wine recommendations, analyzes labels, and offers information about vineyards, grape varietals, and winemaking processes.

Research and Data Analysis

Summarization: I’ve frequently used AI tools to distill complex legislation, regulatory documents, research studies, and dense health insurance plans into clear, manageable summaries. What is perhaps most amazing is giving AI a complicated piece of legislation and then having a “conversation” with it about the different sections, the implications of a provision, or explaining another legislative reference.
Organizing Research: I use Google’s NotebookLM to organize research into notebooks focused on topics like AI studies, healthcare research, and exit polls from the 2024 US election. Within the notebooks, The ability to query specific sources within these notebooks has dramatically streamlined my workflow—and the podcasting feature has added an engaging, enjoyable dimension.

Data Analysis: Leveraging ChatGPT, I’ve analyzed financial forecasts, Census data, public opinion polling, NAEP results, and state healthcare data. Instead of writing complicated Excel formulas or pivotal tables, I simply pose questions conversationally, as if consulting a seasoned data analyst.

Legislative Drafting Assistance: Claude assisted me by analyzing the Loper Bright Enterprises v. Raimondo decision, reviewing draft legislation for vulnerabilities related to executive authority, and proposing language to address those weaknesses. The suggestions provided were so good that only minimal refinements were needed by the Legislative Counsel before incorporating them into the introduced bill.
Research: Perplexity has been my go-to for deep analysis and synthesis of search results. I’ve been blown away by Google’s Gemini Advanced 1.5 Pro with Deep Research, which takes a set of questions, puts together a research plan, and then explores hundreds of websites before producing a 15-page brief with citations.

Everyday Assistance

Meta AI with Ray-Ban Glasses: These smart glasses combine stylish Ray-Ban design with advanced Meta AI, enabling real-time language translations, image analysis, and conversational features. They've allowed me to effortlessly translate signs from French to English, gain insights into art and landmarks, and capture memorable experiences through photos and videos. These are incredible particularly for families who want to capture moments with their kids.

Recipe Brainstorming: Activating ChatGPT’s Advanced Voice with Vision, I wandered through a farmers market, letting it identify ingredients from various stalls and brainstorm recipes on the spot.

Graphics: I regularly use Midjourney to design graphics or stock art. I even created a personalized graphic for wrapping paper, which I uploaded to Zazzle.
Household Repairs: After sharing a photo of a malfunctioning smart light switch with ChatGPT, I received step-by-step instructions on how to fix it.
Public Speaking Practice: I used ChatGPT’s Advanced Voice to rehearse a speech for XPrize, receiving constructive feedback on tone and delivery.
SimpleAI: It has been incredible calling restaurants for reservations and canceling subscriptions. It’s like having a personal agent concierge to execute tasks for you.
Waymo: I was surprised to learn that I was in the top six percent of Waymo riders last year. Riding in a Waymo nothing short of a magical experience - all powered by AI.

AI has become an invaluable partner for me, giving me instant access to a wide array of “expert assistants.” I now have a data analyst, a driver with Waymo, a brainstorming partner, a legislative analyst, a medical assistant, a start-up advisor, a graphics designer, and a researcher at my fingertips, ready to help whenever I need specialized skills or help.

Yet, what’s most exciting about these AI capabilities is that we’re still in the early days. Many of the technologies remain at the start of their maturity curve, meaning they’ll only grow smarter, more capable, and easier to use over time. As these systems continue to evolve, so too will the ways we integrate them into our daily routines.