Perplexity Comet for job search, evaluated

Perplexity made Comet free worldwide in October 2025 and the new Comet Assistant landed shortly after. One of Perplexity's own suggested use cases is "Organize your job search," and a wave of LinkedIn posts followed showing people using it to scrape listings, draft cover letters, and autofill applications. The pitch is tempting. One AI browser, your whole pipeline.

We spent time with Comet on a real job search. The verdict is more nuanced than the launch posts suggest. Comet is genuinely useful for parts of the search and genuinely thin on others. This post walks through both sides honestly so you can decide where it fits in your workflow.

What Comet actually is

Comet is a Chromium-based web browser with an AI sidebar that can read whatever page you're on and act on it. The assistant runs on a mix of GPT-5, Gemini, and Perplexity's own Sonar model. The free product covers the browser plus the assistant. A paid Comet Max tier adds an email assistant that surfaces recruiter messages and helps schedule interviews.

The mental model that works best is "AI agent that lives in your browser." Anywhere you're already browsing, Comet can summarize, organize, draft, and click. Job search is one application of that capability, alongside research, shopping, and general task automation.

Where Comet shines for job search

Three things stand out from real use.

Company research is the obvious win. Paste a company name, ask for a brief, and Comet returns a citation-backed summary covering recent news, funding, headcount, Glassdoor sentiment, and any relevant press. This is what Perplexity has always been good at, just routed through a job-search lens. If you research five companies before applying, the time savings compound fast.

Tab management across job listings actually works. Open ten Workday tabs from a recruiter dump and ask Comet to summarize which roles are worth a deeper look. It reads the postings in your open tabs and produces a ranked digest. For anyone who hoards browser tabs during a search (everyone), this is a real workflow improvement.

LinkedIn networking research is fast. Find a hiring manager's profile, ask Comet to identify second-degree connections you could ask for an intro, and you get a concrete list with reasoning. Same pattern for finding alumni at the company, prepping for an interview by reading the interviewer's blog posts and conference talks, or writing a connection request that cites something specific from the person's recent activity.

These three use cases lean on Comet's core strengths, which are reading the open web fast and producing structured summaries with citations. Anything that fits that shape works.

Where Comet falls short

Now the honest part.

Application autofill is hit or miss. Built In's review of Comet for job search reports "mixed luck" with autofill across applicant tracking systems, and that "it sometimes gets bogged down or disabled by difficult prompts or technical issues." Our experience matched. On LinkedIn Easy Apply listings the autofill works cleanly. On longer ATS forms with conditional fields, knockout questions, or non-standard work-history sections, the results are inconsistent. Comet often gets the basics right and then stalls or fills the wrong field for everything custom.

Comet cannot autonomously submit applications that require written components. Built In's review states this directly. "Comet cannot autonomously submit applications that require cover letters or other written elements." Fully automated submission is limited to Easy Apply roles. Anything beyond that, custom screening questions, written cover letter prompts, the "tell us in 150 words why you want to work here" box, still requires you to write and submit yourself. The marketing language around "automate your applications" papers over this gap.

Resume tailoring is keyword swapping, not structural rewriting. Ask Comet to tailor your resume to a posting and you get a chat-window output that adjusts language and surfaces keywords from the job description. It works for a quick polish. What it doesn't do is rebuild bullet points against the specific requirements with a structured before-and-after match score, the way dedicated tools do. There's no record of what changed and why, no ATS scoring against the posting, no version control across applications. You're left with a chat history rather than a tracked tailoring workflow.

Interview prep is question generation, not practice. Comet will generate role-specific interview questions from your resume and the job description. That's useful as a study aid. It will not run a voice mock interview, score your spoken answers on structure or specificity, or give you the kind of feedback that improves performance across reps. Reading questions silently and answering them in your head is not the same skill as speaking out loud under time pressure, and Comet treats them as if they are.

Generated content can drift from your real experience. Built In's review flags this directly, warning readers to "read through Comet's recommendations before adding its language to your LinkedIn page, as it may generate information that's inconsistent with your skills, experiences and accomplishments." When you ask Comet to write a cover letter or summarize your accomplishments, it sometimes invents specifics or recombines past roles in ways that read confidently but don't reflect what you actually did. Manual review for accuracy is non-optional, and the time you spend checking eats into the time savings.

Salary negotiation is missing. Comet can research comp benchmarks (Levels.fyi, Glassdoor, Blind threads). What it does not do is coach you through the actual negotiation conversation, prep you for the specific objections you'll hear, or rehearse the moment when the recruiter says "what are you looking for" and you need a non-disastrous answer. The negotiation conversation is where most candidates lose money. A research summary doesn't fix that.

The honest read on AI browser agents for job search

Comet is the first widely-distributed AI browser agent that takes job search seriously as a use case. ChatGPT Atlas, Dia, and Arc Search are circling the same territory. The category is real and it will keep improving fast.

Where these agents win is the parts of job search that look like generic browsing tasks. Reading a posting, summarizing a company, organizing tabs, drafting a first pass at a cover letter, networking research. Anything that's basically "look at this web page and tell me about it" or "fill out this form with what you already know about me" is what they were built for.

Where they break down is the parts of job search that are workflow-specific rather than browsing-specific. Per-application resume tailoring with structured ATS scoring. Voice-based interview reps with role-specific feedback. Salary negotiation rehearsal. Application tracking with state transitions. These tasks have specialist shapes that don't reduce cleanly to "agent in a sidebar."

Generalist agents will keep improving on the workflow tasks. Specialist tools will keep improving on the browsing tasks. For now and the next year or two, the right answer is using both, not picking one.

When to add a specialist tool to your Comet setup

If you're already using Comet for company research and tab management, the gap to fill is the workflow side. You want a tool that does three things Comet doesn't.

The first is per-application resume tailoring with a structured ATS match score and a recorded before-and-after, so you can see exactly what changed for each posting and which keywords the job description actually weighted. Chat-window keyword swapping isn't enough when you're applying to thirty roles and need to track which version of your resume went where.

The second is voice mock interviews with role-specific scoring. Reading questions doesn't build the skill. Speaking out loud under time pressure does. The platforms that handle this well run multi-turn voice conversations with adaptive follow-ups and feedback on content quality, structure, specificity, and clarity. Comet doesn't, and there's no roadmap signal that it will.

The third is salary negotiation prep. The parts of negotiation that determine the outcome are the rehearsal and the specific moves (anchoring, silence, the script for "what are you looking for"). Comet can pull benchmarks. It can't run the conversation with you.

Four-Leaf covers all three at $20 a month with a 7-day free trial. Voice mock interviews across 20+ roles with structured scoring rubrics. Per-application resume tailoring with ATS match scores and before-and-after views. A salary negotiation coach that walks through the actual conversation, not just the comp data. AI cover letter generation tied to the same per-application workflow. AI job discovery that ranks postings by skill overlap. An email assistant for recruiter messages. A LinkedIn optimizer.

The split that works in practice is using Comet for the browsing tasks (research, summarization, networking, autofill on Easy Apply) and using a specialist platform for the workflow tasks (tailoring, voice practice, negotiation prep). Both run in parallel. Both earn their cost back on different parts of the search.

(Quick disclosure. Four-Leaf is our product. We wrote this post because Comet is genuinely useful for parts of job search and the launch coverage is glossing over the gaps. Same evaluation we'd write for any other tool.)

Bottom line

Perplexity Comet earns a place in an active job search. Use it for company research, tab management, networking, and Easy Apply autofill. Don't expect it to replace per-application resume tailoring, structured interview practice, or negotiation coaching. Those are workflow-specific tasks that benefit from purpose-built tools.

The most expensive mistake right now is reading the launch posts, signing up for Comet, and assuming you've covered the search. You haven't. You've covered the browsing layer. The work that decides whether you get the offer happens below that, and that's where you still need specialists.

If you want to see what specialist coverage looks like, Four-Leaf covers tailoring, voice mocks, and negotiation in one product. If you want to dig deeper on any of these tasks individually, the related reading below covers each one in detail.

---

Related reading

• Best AI interview prep tools in 2026, 10 compared covers the dedicated interview-practice category in depth.
• Best resume tailoring services in 2026, 6 compared covers the resume-tailoring category, including how Comet's chat-window approach compares to dedicated tools.
• Best mock interview platforms in 2026, 8 compared by role covers voice-based and live mock interview options across software engineering, product management, and data science.
• Best job application software in 2026, 5 compared covers trackers and pipeline tools for organizing an active search.
• How AI is changing the job search in 2026 covers the broader category trend across every stage of the search.