Claude and ChatGPT have quietly replaced Perplexity in my daily workflow

At a glance:

• The author gradually dropped Perplexity from her daily routine, keeping only Claude and ChatGPT as her primary AI tabs.
• Claude and ChatGPT complement each other: one catches structural gaps and vagueness while the other handles image generation and voice dictation in-browser.
• Perplexity still wins when inline citations and visible sources are needed immediately, but the author now asks Claude and ChatGPT for sources on demand.

How the workflow shifted

Writer Judy, a long-time contributor at XDA with over a decade covering digital media, describes a surprisingly common experience: the gradual displacement of one tool by two others that turned out to work better together. For months she opened three browser tabs every morning — Claude, ChatGPT, and Perplexity — convinced she needed all of them. Over time, however, Perplexity drifted to the background. The Windows app would still auto-launch on boot, but she found herself closing its window before doing anything else. What replaced it was a two-model routine: paste the same prompt into Claude and ChatGPT, compare the answers, and let each fill the gaps the other missed.

The shift wasn't dramatic or planned. It happened "little by little," she writes. One day she realized she couldn't even remember why she still had Perplexity installed. The two remaining tools had quietly absorbed her entire workflow — writing, research, coding, image requests, and even voice-to-text brainstorming. The experience highlights how the AI assistant market is settling into overlapping but distinct roles rather than a single universal tool.

Claude and ChatGPT catch what the other misses

The core of the new workflow is cross-checking. When working on a project, Judy pastes the same prompt into both models and compares outputs side by side. One model might deliver the strongest structure, while the other flags vagueness, missing context, or information that should be added to strengthen the result.

A concrete example stands out. While working on a side project about how ancient governments handled power — focusing mostly on the Roman Republic and Greek city-states — she pasted ChatGPT's answer into Claude expecting a quick approval. Instead, Claude flagged two problems: she had explained the Roman Senate twice in different paragraphs, and she had barely mentioned how Athens actually ran its assembly. She then pasted that feedback back to ChatGPT, which admitted Claude had a point. The back-and-forth continued until one model conceded the other's observation because it had no counterargument. This iterative cross-examination, she says, produces noticeably better output than either model alone.

ChatGPT handles images and dictation in ways Claude can't (yet)

One practical differentiator Judy highlights is image generation. She frequently asks ChatGPT to create images directly inside a chat. In one instance, she uploaded a photo of her cat and prompted ChatGPT to change the background to a paradise-type setting. The result came back in minutes, and after a couple of follow-up prompts to adjust the background and overall look, she had a finished image without leaving the conversation. Claude can help with visual ideas and has its own Claude Design workspace, but Judy says it doesn't serve the quick-turnaround image requests she needs — like making a "Good Morning" banner with a bunny.

The second differentiator is voice dictation. On desktop, Judy doesn't always want to type out long, messy thoughts. Sometimes she switches between English and Spanish, or she just wants to jot down an idea before she forgets it. ChatGPT's dictation button in the browser lets her speak her thoughts, review the transcribed text, make changes, and then send the prompt — all without leaving the tab. At the time of writing, Claude on Chrome offers voice mode but not the dictation button; the Claude Windows 11 app has the dictation button but requires leaving the browser. For her in-browser workflow, ChatGPT's dictation feature is the smoother option.

Perplexity still wins for visible sources

Despite the workflow shift, Judy hasn't abandoned Perplexity entirely. She still opens it when she needs inline citations and visible sources immediately. Perplexity's answers come with citations scattered throughout, and hovering over any citation reveals a popup showing the source. At the bottom of each answer, Perplexity typically provides about five follow-up questions, and clicking the sources at the bottom opens a side panel with the reference material.

That automatic sourcing is a real time-saver when she's fact-checking something quickly. She always opens Claude and ChatGPT first, but when sources need to be part of the answer from the start, Perplexity wins on convenience. The trade-off is that Claude and ChatGPT don't show sources by default — but they can produce them on request. For example, she asked ChatGPT to run a live search on the different price plans for Claude, ChatGPT, and Gemini and to show the sources. When she clicked a source, a new tab opened with the full information. The extra step is the only difference; once she remembers to ask, she gets the proof she needs.

What this says about the AI assistant landscape

Judy's story is a small but telling data point in a crowded field. As of May 2026, the major players — Anthropic's Claude, OpenAI's ChatGPT, and Perplexity — each occupy different niches. Claude excels at catching structural issues and nuanced feedback. ChatGPT leads on multimodal tasks like image generation and in-browser dictation. Perplexity leads on source transparency and automatic citation. None of them is a one-size-fits-all replacement for the others, but the author's experience suggests that two tools can cover most of the ground that previously required three.

The pricing context matters too. Claude's Pro plan runs $17 per month on top of a free tier, and ChatGPT similarly offers a free plan with a paid tier. Both are available on Windows and macOS. For a writer and researcher who toggles between English and Spanish and manages multiple side projects, the cost of maintaining two subscriptions is offset by the quality gains from cross-checking outputs. Perplexity remains in the rotation, but it's no longer the first tab she opens each morning.

Tags

• Claude
• ChatGPT
• Perplexity
• AI workflow
• Anthropic
• OpenAI