Cover illustration created with AI for PIN Agency Insights. The scene and interface elements are conceptual, not a real product demonstration or client work.
When leaving a product is difficult, continued use can reflect switching costs as well as satisfaction. As AI lowers some customization and switching costs, customers gain more ways to change how they use a service.
In July 2025, developer Simon Willison heard friends complain that the Open Sauce conference schedule was difficult to use on a phone. He used AI coding tools to extract the schedule and build a mobile-friendly alternative, including calendar export, while working on his own phone.
His firsthand account also describes the repairs: incorrect dates supplied in his prompt, oversized speaker images, and accessibility problems. He explicitly credits more than 25 years of web-development experience. His expertise and the repairs were part of the result.
The conference supplied the event and its information. An independent developer supplied a different way to navigate that information.
For businesses, this separates the value of a service from the value of its interface. Customers may want the information or capability while preferring a different way to access it.
Switching costs can conceal dissatisfaction
A service can bundle useful information, access to other people, transactions, and the interface through which everything happens. The customer may value only some of that bundle.
Someone can want a publication’s reporting without its autoplay video. A creator can need a particular editing function without wanting a larger subscription. A professional may keep paying because moving files, retraining colleagues, and replacing integrations would be expensive.
Continued usage alone gives an incomplete picture of customer satisfaction. Files, habits, integrations, and access to other users can keep people with a service despite dissatisfaction with promotions, restrictive defaults, or unwanted features.
That is where “slopification” becomes a commercial risk: the useful service is increasingly surrounded by interchangeable content and demands on attention. AI can worsen the experience through unwanted summaries and generic recommendations, while also helping users build ways around them.
Some interface steps support the service, such as authentication and transaction checks. Others serve advertising, promotion, or retention goals. As workarounds become easier to build, customers have more options to avoid interruptions that add little value to their task.
AI changes the cost of a workaround
Users have been changing interfaces for years. Firefox’s desktop Reader View appeared in 2015, well before today’s generative-AI coding tools. Mozilla’s current documentation describes removing clutter such as ads, buttons, and background media from compatible pages, with controls for reading presentation.
The newer possibility is more specific customization. A general-purpose reading mode offers a prepared alternative. AI-assisted coding can help someone attempt a view built around a particular task: selected event sessions, a comparison using chosen fields, or a small tool that performs one repetitive operation.
The inference from Willison’s project is that some adaptations are becoming cheaper to attempt, especially for technically capable users. One builder can supply an alternative that many people use, extending its reach beyond people who can program.
Competition could begin with a rival taking over particular tasks: smaller projects moving elsewhere, teams adding fewer paid seats, or newcomers choosing another tool. Each is a possible route by which customers could need less of an established suite.
Replacing presentation involves different work from replacing the capabilities behind it. Authoritative information, rights to use it, dependable transactions, and professional file fidelity are harder to substitute than a layout. For a designer comparing tools, a useful test would be editing a representative client file and handing it back with layers and fonts intact. Reliability on the finished task determines whether the alternative is a practical substitute.
Adobe shows why trust matters before an alternative is complete
In June 2024, a request to accept updated terms brought the company's access to customer work into focus. Adobe's initial explanation described changes around content moderation, including human review of material processed or stored on its servers. Alongside licensing language about operating and improving its services, this prompted questions about what Adobe could do with creators' work.
For professionals handling unreleased campaigns or confidential client images, upload permissions, access to files, and use of the underlying work affect whether a cloud service fits their client obligations. Uncertainty about those terms can therefore affect trust even before a customer has a practical replacement for the software.
On June 10, 2024, Adobe acknowledged the concerns and promised narrower, clearer explanations. Its current general terms, effective October 3, 2025, do not give it blanket permission to train generative AI on private customer work. Material submitted to Adobe Stock is governed separately, and training a custom model at a customer's specific request is also distinguished. Adobe’s subsequent clarifications addressed the wording, but the backlash exposed a deeper problem: customers were questioning whether they could still trust the company with their work.
Subscription terms and switching costs
The terms controversy followed other changes to Adobe's commercial model. Adobe's 2013 regulatory filing records that new creative-product features would go exclusively to Creative Cloud subscribers, with CS6 the last major update for perpetual-license customers. For a customer who preferred paying for a version and deciding when to upgrade, ongoing subscription access was a materially different proposition.
Cancellation became a separate controversy. The US government alleged that Adobe inadequately disclosed early-termination charges and made cancellation unnecessarily difficult. A settlement announced in March 2026 provided for $75 million in civil penalties and $75 million in services, alongside clearer disclosures and simpler cancellation. Adobe denied wrongdoing; its April benefit FAQ says eligible US customers had been contacted. The settlement concerned subscription and cancellation practices; the June 2024 controversy concerned access to customer content. Both brought questions of customer control into focus.
Switching costs remain substantial for some professional workflows. A studio may have years of editable client files, established plug-ins, trained staff, and collaborators expecting particular formats. A lower-priced application becomes a practical substitute when the studio can finish and hand over its work reliably. These requirements can keep a customer using a service despite dissatisfaction with its terms.
ArtCraft's Crafting Apps initiative illustrates the ambition to build independent alternatives. Its PhotoCraft repository describes an independent, clean-room Photoshop reimplementation. The project remains early alpha and describes itself as unsuitable for daily professional use; its parity documentation distinguishes visible commands from dependable behavior.
An alternative can become useful for a narrower job before it covers an entire professional workflow. That creates a potential opening for challengers among customers already looking for different terms or more control.
PlayStation shows what reducing choice puts at stake
PlayStation offers a related example of customer choice: the shift from physical discs toward digital distribution.
On July 1, 2026, Sony announced that physical-disc production for new PlayStation console games would end from January 2028. It cited changing preferences toward digital media. Existing disc releases, including games released on disc before that date, are unaffected. Sony says new games will remain available digitally through PlayStation Store and retailers.
Players who buy used games, lend discs, or resell them after finishing use options tied to physical copies. Digital access changes those options. Sony’s decision therefore changes the offer for this group, even as the company reports a broader preference for downloads.
A company can simplify its product around typical usage while removing options valued by other customers. Aggregate usage figures describe adoption, but give a less complete picture of how different groups value particular forms of access and control.
Overdrive and reactions to the disc decision
In October 2026, GamesRadar+ reported on Overdrive, an unofficial storefront listing more than 1,600 games for modified PS5 consoles, and replacements that appeared after it disappeared. TechSpot reported that some supporters linked their enthusiasm to resentment over the disc decision. It also described compatibility limits and unreliable downloads; the reported Sony legal notice was not publicly confirmed.
The storefront involved unauthorized game distribution. Its usage and commercial impact remain uncertain in the available reporting.
The reported support for Overdrive shows how dissatisfaction with a provider can become part of a challenger’s appeal. In this case, some supporters framed a piracy project as a way to regain choice.
An unfinished product can become a powerful story
ArtCraft attracted specialist media attention while its software was still in early alpha: Creative Bloq tested the apps, while Digital Camera World discussed the challenge to Adobe. The reviews included substantial reservations. The coverage also shows how a technically unfinished alternative can enter the conversation before it becomes a dependable replacement.
Cost and control are also part of the project’s appeal. ArtCraft's developer announcement emphasizes freedom from subscriptions and access to the source code. TechSpot reported that cancellation fees helped motivate the project. In the Hacker News discussion, enthusiasm about leaving Adobe sits alongside detailed criticism of the software. People can recognize flaws and still want the challenger to succeed.
A similar resistance narrative appeared around Overdrive. Tom's Hardware reported millions of views for a demonstration and comments connecting support to anger over Sony's disc decision. Independent reimplementation and unauthorized game distribution raise different legal and technical issues. The reactions share a David-versus-Goliath appeal, with the challenger cast as giving users control that a powerful company has taken away.
The reputational contest can begin while the practical alternative is still limited. In these discussions, support for the challenger also expresses a view about who represents users’ interests. Our expectation is that AI-assisted development will encourage more attempts to build narrower, user-controlled alternatives to established digital products. Which ones endure will depend on reliability, maintenance, rights, and a workable business model. For established providers, trust and customer control can influence how these alternatives are received, well before their long-term prospects are clear.
User control can be part of the service
Control can be part of the official product. Bluesky’s custom-feed announcement described users selecting feeds built by third-party developers. A provider can support different experiences without abandoning the underlying service.
Kagi takes a paid approach to a cleaner interface. It offers subscription plans for ad-free search, drawing on its own indexes and external search sources. Its model puts the ad-free experience within the paid offer.
The same principle could guide data sharing. One possible design is an optional program that offers credits or discounts in exchange for specifically defined contributions. A customer could choose eligible material for a stated improvement or training purpose, instead of discovering that a broad permission is bundled into continued use.
The program’s design would specify purposes, recipients, retention and withdrawal limits, and the benefit offered. Customers who declined would retain ordinary service at the standard price. Eligibility would depend on the customer’s rights and permissions to share the material, including any client work or personal data.
The commercial question is whether customers find that exchange worthwhile. Its reception would depend on the terms, the value of the benefit, and how freely customers can decline.
Reliability and rights remain part of the service
A custom interface needs current data, correct behavior, security updates, and accessible controls. Someone must maintain it when the source changes. Its privacy protections also depend on how it handles data and who can access it.
An official provider can justify its place through dependable file handling, support, accessibility, and recovery when something goes wrong. Those capabilities give customers concrete reasons to pay and stay.
Public information, private account data, licensed assets, and protected software carry different access and use rights. Interface changes also affect attribution, context, and the funding of original work. These considerations remain part of evaluating an alternative alongside its usability.
What gives customers a reason to stay
For a publisher, the enduring value may be reporting readers cannot find elsewhere. For a creative-software company, it may be reliable handling of demanding work. For a marketplace, it may be trustworthy transactions and effective dispute resolution.
Engagement metrics can obscure those priorities. An upsell may produce a click without improving satisfaction; a recommendation may increase time on a screen without helping someone finish a task. Retention figures are also affected by switching costs.
As alternatives become more practical, dependable operation, clear terms, and meaningful choice may become more important reasons to stay. The competitive risk is losing that preference even while retaining a feature advantage.
Sources and notes
Official announcements establish what providers said or changed. Firsthand accounts and independent reporting provide additional context; they do not by themselves establish market-wide behavior, causation or revenue effects.
- Simon Willison: Vibe scraping — A firsthand account of rebuilding an event-agenda interface.
- Mozilla: Firefox 38.0.5 release notes — Documents the 2015 introduction of Reader View.
- Mozilla: Reader View documentation — An established example of an alternative reading interface, not an AI invention.
- Adobe: 2013 quarterly filing — Background on Adobe’s move toward Creative Cloud subscriptions.
- Adobe: June 6, 2024 terms clarification — Adobe’s explanation during the terms controversy.
- Adobe: June 10, 2024 response — The company acknowledged the need for clearer commitments.
- Adobe: General Terms of Use — Consult the applicable current terms separately from historical commentary.
- U.S. Department of Justice: March 13, 2026 settlement announcement — Official account of the settlement over alleged subscription-law violations.
- Adobe: March 2026 settlement statement — Adobe’s response, including its denial of wrongdoing.
- Adobe: settlement benefit FAQ — The company’s explanation of the settlement benefit.
- ArtCraft: Crafting Apps — The developer’s description of its independent applications.
- PhotoCraft: public repository — Project status and source code; not evidence of full Adobe-suite parity.
- PhotoCraft: parity documentation — The project’s own notes on scope and limitations.
- PlayStation: July 1, 2026 disc-production announcement — The official announcement, including its stated scope and timing.
- GamesRadar+: reporting on Overdrive — Reporting on unauthorized distribution, not a usage guide or proof of a measured business impact.
- TechSpot: reporting on Overdrive — Additional reporting and limitations; an alleged Sony notice was not confirmed.
- Bluesky: custom feeds — A provider-supported example of user choice.
- Kagi: subscription plans — The provider’s plan information, not a recommendation based on a comparative test.
- Kagi: ad-free search — The provider’s explanation of its advertising model.
- Kagi: search sources — Clarifies the service’s use of its own and external search sources.
- Creative Bloq: hands-on ArtCraft review
- Digital Camera World: ArtCraft coverage
- ArtCraft developer announcement on Reddit
- TechSpot: ArtCraft developer motivation
- Hacker News: ArtCraft discussion
- Tom’s Hardware: Overdrive coverage
Further reading
- Before Asking for a Meeting, Make the B2B Pitch Useful — a practical companion on demonstrating value before asking for a customer’s time.
Editorial note: This article draws on official announcements, developer accounts and reporting, with AI assistance in research and writing. Business implications are analysis, not measured claims about customer behavior or revenue. Illustrations are AI-generated explanatory scenes, not documentary evidence of the cases discussed.
