Lukas Amacher Is Building a Chatbot for the Art World

The next time you scan a QR code at an opening, you may find yourself conversing with a digital docent, compliments of the new art‑tech platform CONTXT. The idea behind the customizable software, which is essentially an exhibition-specific A.I. chatbot, was inspired by its founders’ shared belief that art appreciation should be a conversation. Every artwork has a story, but the challenge has always been that there may not be anyone immediately available to tell it. Wall text can’t answer specific questions, and by the time a gallery visitor or museum-goer gets home, they may have forgotten what it was they wanted to know. “People in exhibitions often have questions about what they’re seeing but don’t have access to curatorial insight or background knowledge in real time,” Lukas Amacher, the curator, art collector and entrepreneur who launched the project with developer David Simon, told Observer. “With CONTXT, visitors can ask questions about an artwork and get answers sourced directly from the institution’s knowledge base rather than generic internet results.”

I was surprised to find out that Amacher, the former managing director of the 1OF1 digital art collection, never learned to code. But “the distinction between computer user and developer is slowly evaporating,” he said in an interview with Seth Goldstein for the Let’s Vibe! Podcast. With Simon building the foundational structure of the platform, Amacher could experiment with its functionality, with the goal of building “the richest citation‑ready context layer for art and culture.”

And that’s CONTXT, an artificially intelligent system that’s trained, project by project, on every catalogue entry, essay, curator note and piece of explanatory text for an exhibition. On the visitor side, CONTXT provides, via a familiar chat interface, real-time answers to questions, compressing reams of material into meaningful conversations about artists and artworks that can happen while the user is still at the exhibition. For institutions, it provides a comprehensive digitized knowledge repository and analytics tracking direct visitor engagement, helping curators understand what pieces and ideas resonate most. “It helps bridge the gap between the insights a curator might provide in person and the self‑guided experience of visitors who might otherwise struggle to access deeper interpretive context,” Amacher explained.

While CONTXT is still in its early stages, it’s already attracted interest from museums (especially smaller ones with limited mediation resources), galleries, corporate collections and art fairs. Earlier this month, the company launched a public research preview of the platform in collaboration with Steve Sacks’ bitforms gallery that includes the current show, Maya Man’s “StarPower” and exhibitions going all the way back to 2002. We took it for a spin, QA tester style, initially posing questions about the artists and shows, then veering off into weirder territory: asking it to solve physics equations, inputting 900 words of Lorem Ipsum text, requesting recipes for soup and then for artist Alexander Reben. We sat down with Amacher—serious questions in hand this time—to find out what CONTXT might mean for the future of how institutions share art and how people experience it.

How did CONTXT come about?

Both my co-founder, David, and I have been in the art world for most of our lives. David’s a collector, an artist, and a developer, and I’ve always been a collector as well. My first major acquisition was a Fred Sandback drawing. We both realized that we view art as a grand conversation. Our project came about when we saw how limited access to knowledge could be in the art world, especially in museums and galleries. If you visit a show with a curator, you have access to their insights, which enriches the experience, but without that, it can be hard to understand the deeper context. With the emergence of LLMs, we saw an opportunity to use this technology to compress the time it takes to dive into an artwork’s context—from hours of research to seconds of conversation. So, we started experimenting with how we could enhance in-person art experiences using LLMs.

What does CONTXT implementation look like for galleries and museums? What do they need to do to opt in?

We’re building libraries of context. The first layer is the knowledge base, where you input everything—from curator interviews to artwork images and reference materials like exhibition guides, articles, and even multimedia like videos. This allows visitors to interact with a comprehensive corpus of information. The second layer involves tailoring this knowledge to different audience types, such as museum visitors or corporate collectors, allowing specific voices to communicate with different groups. The third layer focuses on analytics, where we track what works are asked about the most, what types of questions visitors are asking and what ideas are resonating the most. This provides curators with valuable data to understand what engages audiences.

And what’s the experience like for a visitor using the tool?

For the visitor, it’s a chat interface, like the ones we all use, whether on ChatGPT or other platforms. The visitor scans a QR code at the exhibition, which brings up the chat on their phone. They can ask questions about the exhibition or specific works, and the tool will provide answers sourced from the curated knowledge base, ensuring it’s citation-ready. This gives visitors the ability to have real-time, meaningful conversations about the artwork they’re engaging with without waiting hours to do research or running into misinformation. The tool even allows them to share insights directly to social media, which is a fun twist on the “Instagrammable art” trend. Additionally, visitors can save their interactions, creating a personal archive of their experience at the show.

Can galleries and museums use CONTXT for more than just exhibitions?

Yes, absolutely. We’re also developing a tool for galleries that allows them to create a mobile-friendly, interactive experience where visitors can engage directly with artwork descriptions and ideas. Instead of just scrolling through a PDF or reading a basic label, visitors can have this dynamic conversation, asking questions and diving deeper into the work. This feature is particularly useful for galleries that want to present their works in a more engaging and modern way, merging the tactile with the digital.

What kind of feedback have you received from artists? Are they concerned about losing control of their narratives?

I love that you asked this! It’s something we’re planning to address in the future. We want to give artists and curators the ability to take control of the narrative. Artists could create their own knowledge bases—essentially mood boards for their practice—where they can share their statements, inspirations and ideas. Curators could then connect their exhibition’s context with the artist’s own knowledge base, allowing them to better represent the artist’s vision. This tool isn’t just for visitors; it’s for the whole ecosystem to represent ideas authentically and allow artists to guide the conversation.

Are there any concerns on the data privacy side, especially as you’re collecting interaction data and people’s questions might veer into the personal?

The information we gather is more of a general overview. For example, galleries can see which works are asked about the most, what types of questions people are asking, and how artworks are being interacted with. We collect insights into question types—whether people are asking about biographical details, conceptual meanings or curatorial narratives. But the actual data is anonymized, so it’s more about helping curators understand what ideas are resonating with the audience rather than tracking individuals.

What’s the institutional interest been like so far? Are the challenges museums and galleries are looking to solve the ones you anticipated?

We’ve spoken to several museums, galleries, and corporate collections. Interestingly, many smaller and mid-sized museums have reached out because they have limited mediation budgets and need a cost-effective way to help visitors engage with their exhibitions. The ability to provide multilingual support and make the experience more accessible is a big draw. We’ve also seen interest from art fairs, which could benefit from offering deeper engagement for visitors who aren’t necessarily there to make a purchase but want to learn more.

What’s next for the project?

We’re rolling out pilots, and soon, we’ll be ready to launch. We’re excited to see how it’s received and how visitors engage with the artwork when they can interact with a corpus of ideas in real-time. The overall goal is to give both exhibitors and visitors the tools to facilitate deeper, more meaningful conversations about art.

More Arts Interviews

• Sotheby’s Evelyn Lin On Asia’s Maturing Market, Young Collectors and the New Rules of the Hong Kong Sale

• Under Director Kate Sierzputowski, EXPO CHICAGO Bets On Renewed Institutional Depth

• The Louche, Romantic, Feral World of Sébastien Tellier

• As Venice Nears, Arch Hades Traces Her Shift from Verse to Visuals

• Curators Dieter Buchhart and Anna Karina Hofbauer Want to Reintroduce You to Keith Haring