Detailed Analysis
A photography business owner and multi-platform AI user has reported a significant and persistent performance degradation with Claude when Google Drive source documents are attached to projects, describing response times exceeding one minute for simple queries that previously resolved in seconds. The issue emerged several months ago without any corresponding change to the user's document setup, which consists of five lean, carefully optimized knowledge-base files covering business operations, rates, contracts, and communication protocols. Critically, the same documents continue to function normally in both ChatGPT and Gemini, isolating the problem to Claude's handling of linked Drive sources rather than any issue with the documents themselves. The user's attempts to reproduce or circumvent the problem through varied configurations — different document combinations, new threads, different models, and different platforms — have all yielded the same slow behavior, suggesting a systemic change on Anthropic's infrastructure side.
The practical consequences for the user's workflow are substantial. The workaround they identified — consolidating all five documents into a single compiled file — resolved the speed issue but destroyed the operational elegance of their original system. They must now maintain two parallel versions of their knowledge base: the original five-document structure used by ChatGPT and Gemini, and a separately compiled monolithic file for Claude. Any amendment to business information must be manually mirrored across both formats, introducing a meaningful risk of version drift and inconsistency — a serious concern for a business relying on accurate, up-to-date operational documents in client-facing contexts.
The support dimension of this situation is notable in its own right. The user reports sending multiple support tickets to Anthropic over six weeks without receiving a single reply, a pattern that points to either systemic gaps in Anthropic's customer support infrastructure or a triage process that deprioritizes non-critical performance complaints from individual subscribers. For a paying annual subscriber, the absence of any acknowledgment compounds the frustration of the underlying technical issue and reflects a customer experience gap that stands in contrast to the company's public positioning around responsible and user-centered AI development.
This report fits into a broader pattern of friction that power users encounter when AI platforms silently alter integration behavior. As users increasingly build multi-tool workflows that depend on consistent, predictable API and file-system integrations, undocumented backend changes can cascade into significant operational disruptions. The user's setup — maintaining synchronized knowledge bases across Claude, ChatGPT, and Gemini — represents a sophisticated and increasingly common approach to AI-augmented business operations, and the fragility exposed here illustrates a maturity gap between how advanced users deploy these tools and the stability guarantees platforms currently provide.
The incident also highlights a competitive dimension in the enterprise and prosumer AI market. When one platform introduces friction through unreliable integrations and unresponsive support, users with cross-platform workflows are positioned to deprioritize or replace that tool with minimal disruption. The user in this case has effectively already done so in functional terms, reducing Claude to a secondary, manually-maintained system while ChatGPT and Gemini continue to anchor the primary workflow. For Anthropic, which has invested heavily in positioning Claude as a serious productivity and business tool, consistent integration reliability and responsive support infrastructure are not peripheral concerns — they are core to retaining sophisticated users whose workflows increasingly demand interoperability across the AI ecosystem.
Read original article →