Tyler Peppel’s interest in artificial intelligence began with a small act of academic trespass. In the mid-1980s, he slipped into Marvin Minsky’s lectures at MIT. What caught him was the possibility of “machines that can understand stories.” Long before a chatbot could turn a dinner invitation into a sonnet, here was a proposition worth sneaking into a classroom for.
Peppel was studying the passage from analog media to digital media. Minsky was describing another passage: from machinery that processes information to machinery that might make something of it. Between those two ideas lies much of Peppel’s working life. He has helped bring moving images to the Macintosh, founded an email company, built enterprise software, and taught students to question the tools they use to communicate.
Today he leads RiskWise by TICKR, which applies AI to emerging business risks. He also teaches at New York University. Those jobs sit comfortably beside each other. Both involve an awkward question that survives every improvement in computing: once a machine has produced an answer, what should a person do with it?
A Macintosh acquires a sense of time
At Apple, Peppel worked on the transition from a computer of documents and still pictures to one that could handle video, sound, and animation. His five years in new product development put him at a busy intersection of engineering, product planning, and communication. The machine’s expressive vocabulary was expanding. Someone had to help turn that expansion into a product.
In 1990, that meant QuickTime. Peppel pushed the multimedia proposal toward Don Casey, Apple’s networking and communications chief. Casey wanted a plan he could announce at the developers conference. The distance between an intriguing experiment and a public commitment suddenly became rather short. A team would have to make the promise work.
Even the name required negotiation. Tektronix already held QuickTime; QuickStream was briefly considered while the companies worked things out. Peppel disliked the alternative. A product concerned with the coordination of events in time deserved a name that did not suggest a plumbing emergency. Eventually, the QuickTime name was secured.
There is a useful glimpse of him here: advocating for a technical possibility, handling its practical packaging, and retaining an opinion about what it ought to be called. Product development includes the future, the lawyers, and occasionally the bathroom joke.
- 1990QuickTime announced at Apple’s developers conference
- 2001Electronic trading card patent granted
- 2026AI for Creatives listed at NYU for fall
Email, with something to say
After Apple came more ways to put digital media into people’s hands. Peppel founded and led MightyMail, a rich email services platform. The description is a small time capsule. Email could carry more than a plain message; it could become a place for designed, expressive communication. The inbox was acquiring ambitions beyond the office memo.
His later consulting work included digital data strategies for Goldman Sachs and Cisco. The settings changed, from personal computing to messages to enterprise information, while the work kept returning to how people receive and use what a computer provides. A bank and a creative software team have different vocabularies. Both can suffer from an impressive system whose meaning has failed to reach its intended audience.
His education helps explain that range. He earned a bachelor’s degree at Ohio University and a Master of Science in Visual Studies at MIT. Visual studies is a telling qualification in a career full of software: it leaves room to ask what a person perceives, rather than stopping at what a system can calculate. In Peppel’s career, the display and the message keep getting as much attention as the machinery behind them.
The trading card inside the computer
One of his inventions makes the point in miniature. In March 1995, Peppel filed an application for an electronic trading card system. US patent 6,200,216 was granted on March 13, 2001. Its premise brought the familiar practices of collecting, trading, creating, and playing with cards into digital media.
The interesting choice was the metaphor. A trading card gives a person something recognizable to hold in mind. There are individual objects, collections, exchanges, and games. Peppel’s design applied that vocabulary to computer programs and the hardware and software that supported them. It took a social activity people already understood and explored what could happen when it became electronic.
A patent is a particularly formal home for an imaginative habit. The language is technical; the underlying move is approachable. Start with a recognizable human practice. Give it new capabilities. Work through the system required to support it. That reading of the invention fits the move from static Macintosh documents to media that could unfold over time. The old form supplies a way into the new one.
The same complaint, in different offices
Tickr grew out of Peppel’s consulting practice, InfoMotor. Across industries, he kept hearing related complaints about enterprise data. Information existed, but bringing it together and making it useful to a business user remained difficult. Cloud services and APIs offered a way to assemble a more coherent picture.
Gatorade became the first customer. Hallmark supplied another concrete example: a brand wanted to understand what was happening around its products while it could still respond. Yesterday’s beautifully assembled report could be perfectly accurate and arrive too late to help. Timing mattered as much as presentation.
Peppel also encountered a less romantic part of invention: deciding how to sell it. Early investor Don Hutchison pressed the questions of pricing, packaging, and turning the idea into a product. The attraction of gathering information in one place did not automatically specify a business model.
A channel partner program launched in June 2017 became an important route to growth. Other organizations could bring their sales relationships to the product. Peppel had spent years thinking about how information travels. Now the company had to think just as carefully about how software reaches a customer.
“No shortage of ideas right? That’s the easy part.”
Tyler Peppel
From reading the room to reading the risk
RiskWise gives the information problem a specific destination: identifying developing threats before they become costly events. Its intended users include insurers, financial services teams, and consumer product businesses. The company describes monitoring signals across companies, industries, and risk topics, including regulatory developments and supply chain changes.
That work adds a demanding layer to the old dashboard question. A business can observe a change and still struggle to decide what caused it, whom it affects, or whether it deserves attention. RiskWise’s stated approach connects monitoring with driver discovery, causal reasoning, entity attribution, and risk indices. The aim is to move from a mass of observations toward an explanation someone can use.
Peppel works with a leadership team that includes chief technology officer Sam Kahn and chief operating and financial officer Alex McNealey. The company emphasizes evidence trails and clear communication of uncertainty. Its own account dates the dedicated AI team to 2016 and the public RiskWise launch to 2025.
There is something refreshingly exacting about uncertainty as a product requirement. A screen can be visually confident long before its contents deserve confidence. For a person responsible for a consequential decision, knowing the limits of an answer is part of receiving the answer.
A very particular kind of AI enthusiasm
Peppel’s public updates also show an interest in the unglamorous details that make enterprise AI useful. In November 2024, he shared TICKR’s results on hierarchical product categorization. The company reported accuracy of up to 98.1 percent in its experiments, comparing its models with GPT-4o and Claude Sonnet 3.5 under the same prompts and conditions.
That is a bounded company-reported result on a particular task. The task itself is revealing. Retailers and consumer goods companies need products placed into meaningful categories. A system that gets the hierarchy wrong can create confusion farther down the chain of analysis. The work rewards domain knowledge and careful evaluation, even when it produces little that would enliven a dinner party.
His other updates have covered identifying substitute products from shopping basket data and using language models to help find useful forecasting variables. Here, his enthusiasm takes the form of practical questions about classification, behavior, and prediction. The recurring concern is whether a technical method helps someone understand a specific situation. A model has to earn its place in the workflow.
Back in the classroom, with better tools
For Fall 2026, NYU lists Peppel as the instructor for AI for Creatives, a two-credit undergraduate course. Students work with AI tools across photography, design, writing, illustration, user experience, and video. They produce work, critique it, and exhibit it. Making something is part of learning how to judge it.
The questions go beyond operating the software. How does AI change the creative process? Who makes these tools? What happens to a creator’s role when a machine participates? Those questions give the course its tension. Greater capability brings choices about authorship, ownership, and the direction of a creative profession.
On his course website, Peppel proposes conversations with toolmakers, creative practitioners, and editorial and legal experts. Learning the technology includes asking about the incentives behind it. Students are invited to have a voice in what gets built, as well as to gain skill with what is already available.
His wider projects pursue related questions. Wordloop explores adapting learning material to an individual student’s language. Prosocial Partners gathers researchers, designers, artists, and technologists interested in beneficial uses of AI. These interests connect the enterprise founder with the educator: the intended recipient matters.
The student who slipped into Minsky’s lectures now has a course of his own. Machines can produce stories in quantities nobody could read. Peppel’s work keeps returning to the people at the receiving end: the student making an image, the manager interpreting a signal, the colleague deciding what to do next. The abundance has arrived. Understanding still takes work.