If you run visitor experience at a museum and you've been reading vendor decks for the past two years, you've seen the same slide repeatedly: a list of "visitor expectations in 2026," usually padded with verbs like "transform" and "reimagine." Most of it is true. A meaningful fraction of it is overstated. This piece is the reference I wish existed for that slide — what has actually shifted in visitor behavior since 2019, what the documented evidence supports, and where vendors (including, sometimes, mine) are getting out over their skis.
I'm Eric Duffy. I run Convo, a platform that lets museums publish multilingual, conversational audio tours visitors reach by scanning a QR code. I have a position on this category, and I've been honest about it elsewhere on this site. What follows is the underlying reference for that position — sourced, with the soft spots called out.
For a shorter editorial argument, the essay on the 2026 museum visitor is the right read. This hub is the longer, structured reference it points to.
What does "museum visitor experience in 2026" actually mean?
It's a phrase doing two different jobs at once, and that's part of why vendor copy lands flat. In one sense it means the physical and emotional shape of a visit — wayfinding, the docent at the door, the bench in front of the painting, the line at the cafe. That hasn't changed much. In the other sense it means the interpretation layer: how visitors get context, ask questions, and understand what they're looking at. That has changed a lot, because the device in the visitor's pocket changed, the visitor's expectations of that device changed, and the production tooling available to museums changed alongside them.
When this piece talks about "what's changed in 2026," it means the interpretation layer and the behavior around it. The bench is still the bench. What the visitor does on the way to the bench is different.
Which behavioral shifts are actually documented?
Five shifts have enough evidence behind them to plan around. I'm going to walk through each in detail below, but the short list is:
- BYOD has replaced the rented handset as the dominant audio-guide delivery mode.
- Native app downloads have collapsed as a visitor acquisition channel.
- QR codes have become a normal scan, not a pandemic novelty.
- Multilingual expectation has hardened from aspiration to floor.
- The broadcast-only format is losing ground to interactive, question-and-answer interpretation.
A sixth — accessibility — isn't a behavioral shift so much as a regulatory and ethical one. I'll cover it separately because the dynamics are different.
I'll also flag, at the end, three claims you'll see in vendor decks that I don't think are quite earned yet: "personalization," "immersive AR," and "every visitor wants a chatbot." Each has a kernel of truth that vendors have inflated into a slogan.
Why is BYOD the default for museum audio guides now?
The simplest answer is that the alternative stopped making sense. Per Pew Research's Mobile Fact Sheet, 91% of US adults now own a smartphone — and crucially, 78% of those age 65 and over do. The 50–64 bracket sits at 90%. The visitor segment without a smartphone is small and shrinking, and the argument for renting a piece of hardware to people who are already carrying a more capable one in their pocket has gotten harder to make every year.
The cost side decided it. A rental handset program has staffing implications (someone hands them out, someone takes them back, someone sanitizes them, someone replaces the lost ones), per-unit hardware costs, and a battery-charging operations layer. A BYOD program has a printed QR code and a web player. The trade-off used to be "but our visitors don't have phones with them." That trade-off is gone.
What changed in practice: the question for a 2026 visitor-experience director is no longer whether to deliver to the visitor's own phone. It's how. App, mobile-optimized web, native install via App Clip / Instant App, or a plain QR-to-web flow. The data on each of those has gotten very lopsided, which is the next section.
What happened to the museum app?
The museum app is, with rare exceptions, an artifact of the 2014–2019 era that didn't survive contact with how visitors actually behave. The cross-industry research on cultural-organisation apps — most consistently the long-running tracking by Frankly Green + Webb — converges on the same finding: the average museum app pulls visitor adoption in the low single-digit percentages, and the median app for a cultural organisation has under a thousand downloads and is opened less than once. The actual usage rate is lower than even the download counts suggest, because download counts include people who installed once and never opened.
The cause isn't museum-specific. It's broader app fatigue: the average smartphone user has about 80 apps installed but actively uses only around 30 a month, per industry usage data summarized by Buildfire. The barrier to adding another one — for a single visit to a single institution — is high enough that most visitors don't.
The museums that get higher app adoption tend to be either (a) very large institutions where the app does meaningful work beyond a single visit (Smithsonian-scale collections, multi-site networks), or (b) institutions where the app is bundled into a membership flow that's already happening. For everyone else, the math doesn't work — and even on heavily promoted, well-staffed apps, double-digit adoption is rare enough that the industry literature treats it as the outlier rather than the operating expectation.
The practical takeaway is that "we'll build an app" is no longer a defensible answer to "how will visitors access this." The web player launched from a QR code is. That's the channel BYOD has settled into.
Have QR codes actually been adopted, or is that vendor wishful thinking?
This is one where I want to be careful, because vendor decks overstate it. The honest answer: QR codes crossed the awkwardness threshold during the pandemic and have stayed on the right side of it. By 2024, roughly 84% of US smartphone users had scanned a QR code at least once and around 72% scan one at least monthly, per industry tracking summarized by Electro IQ. The "I don't know how to do this" problem that defined QR codes from 2011 to 2019 is largely gone.
What's not settled is gallery-floor adoption inside the museum. Whether a visitor actually scans the code on the label depends on signage clarity, wall placement, lighting, and whether the museum has communicated that the tour exists before the visitor walked up to the object. The QR code isn't doing magic by itself — it removes the friction that used to kill conversion, but it doesn't create the intent. Several of the institutions I've talked with track "scans per visit" and find that the number is highly sensitive to how the first wall card sets up the experience.
So: yes, the technology is past the adoption tipping point. No, that doesn't mean every code on every wall gets scanned. Signage is still a visitor-experience design problem. For more on the production side of what gets delivered when the code is scanned, see the hub on AI audio guides for museums.
Why is multilingual interpretation now an expectation rather than a perk?
Two reasons, and they compound. The first is demographic. Per US Census American Community Survey data (2018–2022 5-year estimates), roughly 21.7% of the US population age five and older speaks a language other than English at home — about 68 million people. In tourist-destination cities the share of international visitors compounds that further. English-only interpretation was always a constraint; it's now a constraint visitors notice.
The second is technological. The reason museums historically shipped English-only or English-plus-one wasn't curatorial choice — it was production economics. Studio re-recording per language was prohibitive. Translation at ATA standard rates ran twelve to thirty cents per word before re-recording. AI translation and neural voice synthesis have collapsed both of those cost lines, which means the constraint has flipped: language coverage used to be a budget question; now it's almost entirely an editorial-review question.
What that means for visitor experience: in 2019, a visitor encountering an English-only audio tour read it as the default. In 2026, increasingly, they read it as a choice — and not a particularly visitor-friendly one. The family next to them might have access in their language. The museum's wall text might already be bilingual. The mismatch becomes visible.
The deeper treatment of this — how many languages a specific museum actually needs, how to handle editorial review across languages, and how to think about regional dialects — lives in Pillar 3 (Multilingual interpretation) when those pieces ship.
How are accessibility expectations changing under the DOJ Title II rule?
This is the shift that's least about visitor behavior and most about institutional obligation, but it changes what visitors expect because it changes what they encounter elsewhere.
On April 24, 2024, the Department of Justice finalized a rule under Title II of the ADA requiring state and local government entities to make their web content and mobile apps accessible per WCAG 2.1 Level AA. The compliance deadlines have since been extended: April 26, 2027 for entities serving 50,000 or more, and April 26, 2028 for smaller jurisdictions. The rule applies to public museums — including websites, mobile apps, digital signage, and interactive kiosks.
The downstream effect on visitor experience is twofold. First, public-sector museums are now under explicit regulatory pressure to bring their digital interpretation up to WCAG 2.1 AA — which covers captions for audio, audio description for video, screen-reader compatibility, keyboard navigation, and color contrast. Second, private museums that don't fall under Title II are nonetheless being held to a higher bar by visitors, foundations, and accreditation bodies, because the public-sector floor has moved up.
The practical implication for any audio interpretation layer being deployed in 2026: it should be reachable through the visitor's own accessibility settings (screen reader, larger text, captions on video), not a parallel "accessibility version" of the tour that requires the visitor to ask for it. The web-via-QR delivery pattern is, in most respects, structurally better for accessibility than a rented handset, because it inherits the visitor's own assistive configuration. Pillar 4 (Accessibility & inclusion) covers this in depth.
Why is the broadcast-only audio tour starting to feel dated?
This is the shift I have to be most careful about, because I'm not a neutral party — Convo's whole product is built around the idea that visitors should be able to ask questions, not just listen. So let me try to be honest about what the evidence actually supports.
What's documented: across categories, people have moved from broadcast-style media to interactive. Streaming replaced scheduled programming, chatbots replaced FAQ pages, social media replaced one-way publishing. The pattern is consistent enough that expecting museum interpretation to be exempt from it would be unusual. Visitors arrive at the gallery having spent the previous hour asking their phone for things, getting answers, and following their own questions. The expectation calibrates from there.
What's not yet fully documented: the magnitude of the engagement lift from interactive vs. broadcast interpretation. Early peer-reviewed work at venues like Centre Pompidou and the 2025 ACM IMX study on generative-AI chatbots in art museums report measurably higher engagement when visitors can ask, not just listen. The effect is directional. The numbers are still being settled. I'd treat any vendor that claims a precise "X% engagement lift from conversational tours" with skepticism — the studies are early, the methodologies vary, and the institutions doing them aren't yet representative of the field.
So: directionally real, magnitude TBD. The interpretation team that's building today should plan for visitors who want to ask follow-up questions; that pattern isn't going away. The team that's projecting board slides should be careful about citing engagement numbers that haven't been independently replicated.
For the underlying argument about why visitors expect to ask, see the 2026 museum visitor essay. For the harder argument about whether AI-assisted interpretation undermines institutional authenticity, see authenticity and AI.
Where is the visitor-experience category overclaiming in 2026?
Three claims to watch for. Each has a real kernel; each is being inflated.
"Personalized interpretation." The kernel: visitors do have different background knowledge, different time budgets, and different interests. A conversational tour that lets a visitor follow up on what interests them is, in a meaningful sense, more personalized than a fixed broadcast script. The over-claim: nothing in this category is doing Netflix-style behavioral profiling on visitors, and probably shouldn't be. "Personalization" in vendor copy usually means "the visitor can choose what to ask about." That's a real capability; calling it personalization sets the wrong expectation.
"Immersive AR overlays." The kernel: AR has gotten meaningfully better and a handful of high-budget institutions are doing interesting work with it. The over-claim: AR is still a niche delivery channel for a niche fraction of visitors. The infrastructure cost, the device limitations, the inconsistent visitor experience across phones, and the lack of a stable standard mean that AR is not yet the default visitor-experience layer for any institution outside the top tier. Vendors that lead with AR are usually selling to the funding cycle, not the visitor.
"Every visitor wants to chat with the artwork." The kernel: a meaningful share of visitors do want to ask follow-up questions, and the broadcast-only format misses them. The over-claim: not every visit is, or should be, a conversation. Plenty of visitors want to listen, look, and move on. The right framing isn't "every visitor will chat"; it's "the visitors who want more should be able to get it without the museum staffing another channel." That's a more honest pitch and a more accurate description of what happens in real galleries.
For the deeper argument about why the audio guide itself isn't the product, see the audio guide is not the product.
What about visitors who explicitly don't want a digital layer?
A reasonable share of museum-goers come specifically to disconnect from their phone, and they should be served. The shift isn't that wall text and docent tours have been deprecated — they haven't. It's that the digital layer is no longer the only alternative when those channels don't cover what a visitor wants to know. The 3pm docent tour is still the 3pm docent tour. The wall card is still the floor. The audio layer is the optional channel for visitors who want more.
Any vendor (including mine) that frames a digital interpretation layer as a replacement for human interpretation is misreading the room. The right framing is additive: most visits don't get a docent, most galleries don't get a dedicated wall essay, most institutions don't have the staff time for what their curators actually want to say. The phone-delivered layer is what fills the gap for the visitors who reach for it. The visitors who don't reach for it still have everything they always had.
Which shifts are most relevant to a small or mid-size museum?
Not all of them apply equally. If you run a 15,000-visitor regional history museum with no rental program and a single docent corps, the "death of the rented handset" shift isn't operative — you never had handsets. What's most likely to apply, in rough order:
- The multilingual gap is becoming visible. Even in non-tourist regions, the local population's language profile has changed. Spanish-language interpretation is the most common immediate need.
- Visitors arrive expecting digital interpretation to exist. Not necessarily to use it, but to find it when they want it. Its absence reads as institutional choice now in a way it didn't a decade ago.
- Accessibility is becoming both an obligation and a visible signal. Whether or not your institution is under Title II, the public-sector floor is moving up.
- The broadcast-only audio guide, if you have one, is losing relevance faster than its production cost would suggest. A single English tour produced in 2017 doesn't read as "thoroughly considered" to a 2026 visitor the way it did in 2019.
The institutions where I've seen the shifts land hardest are the ones that didn't have a strong digital interpretation layer to begin with — because for them, every shift is additive rather than disruptive. For institutions that invested heavily in a 2014-era app or a fleet of handsets, the shifts feel more like sunk-cost decisions to unwind. Both are working with the same forward picture; they're starting from different places.
For pricing context on what a small or mid-size institution should expect to budget, see Convo's pricing page and the buying-and-cost pillar when it ships.
Frequently asked questions
Continue reading
For the editorial version of this argument, the essay on the 2026 museum visitor covers the same ground in a more compressed form. For the production side — how the audio that visitors hear actually gets made — start with the hub on AI audio guides for museums and the spoke on AI audio guide vs traditional audio guide.
For the harder question of whether machine-assisted interpretation undermines the institutional voice, the note on authenticity and AI is the place to start. For the argument that the audio guide itself isn't really what visitors come for, see the audio guide is not the product.
When the rest of the visitor-experience pillar ships, this hub will link out to the spokes covering BYOD signage, dwell-time data, multi-site visitor flows, and accessibility implementation. For pricing context, Convo's pricing is the source of truth.
About the author
Eric Duffy is the founder of Convo, a platform that lets museums and cultural institutions publish multilingual audio tours their visitors can have a conversation with. He writes about how museums could afford to be more ambitious with interpretation, drawing on discovery conversations with curators, directors, and education leads at small and mid-size US museums. Reach him at eric@convo.app or on LinkedIn.