You are weighing two products and you need a decision you can defend by Friday, not a browser full of 40 open tabs. If you already use Perplexity, you can hand it both URLs and get back a sourced, dated matrix with an overall lean. No code, no prompt engineering, no researcher on staff. Here is the whole flow, one step at a time.
Step 1: Connect Cobalt to Perplexity once
Cobalt runs through an MCP connector, the standard way an assistant plugs into an outside tool. In Perplexity, add the custom connector and paste https://cobaltcapture.com/mcp. The full setup, with the exact menu path, is on the Perplexity connector setup page. You do this one time. After that, Perplexity can see every Cobalt study and run them on request.
Outcome: Perplexity now has the comparison study available. You never touch it again for setup.
Step 2: Confirm the study exists before you ask
Type /workflows in Perplexity. It lists every study Cobalt offers and the exact words that trigger each one. You will see the comparison study described as two or three products compared for one decision. This step takes five seconds and saves you from guessing at phrasing.
Outcome: you know the study is live and you know what to say next.
Step 3: Ask in plain words
Write the request the way you would say it out loud. Name the two products, name the pages, and name the decision. For example:
Run a comparison study of acme.com and rival.com for a 10-person team picking a CRM. Compare pricing, onboarding, and reporting.
You do not need to write a spec. The more specific the decision, the sharper the lean at the end. "Which is cheaper for a solo user" produces a different matrix than "which scales to 50 seats." Pick the question you actually have.
Outcome: Perplexity kicks off the study. Comparison studies run on public pages plus web search, so there is nothing for you to click through and nothing to log into. Cobalt renders the pages itself.
Step 4: Read the dated matrix
The study comes back as a review you can share. It leads with the verdict, an overall lean toward one product for the decision you named, then backs it with a matrix. Each row is a thing you asked about. Each cell is sourced and dated, so when a competitor changes a price next month you know your numbers were current as of the day the study ran.
Because the study pulls public pages and web search, the matrix cites where each claim came from. That matters when a stakeholder pushes back: you point at the source, not at a hunch. This is the same backbone as a hand-built competitive analysis, except you did not spend a morning tab-hopping to build it.
Outcome: a decision with a reason attached, and a dated record of why.
Step 5: Share the review or pin what matters
The study publishes to a short review link anyone can open, no login on their end. Send it to the person who has to sign off. If you want to flag one cell, say a pricing tier that changes the whole call, you can open the review and drop a numbered pin and a comment on the exact spot before you share it. For a worked version of pinning the sources that decide the call, see pin the Perplexity citation that matters.
Whoever opens the link can comment on any item, and you can mark each comment resolved as you settle it. That turns a one-shot study into a short thread instead of a reply-all email chain.
Outcome: the decision moves to the person who signs off, with the evidence attached and a place to argue with it.
When to run the study on three products, or on the same screens
A comparison study takes up to three products, not just two. If you are deciding between three, name all three and the matrix widens. If your decision really hinges on one specific screen, the checkout, the signup, the empty state, a different approach fits better: walk all of them on the same screens, which the compare two competitors on the same three screens template covers. The comparison study answers "which product, overall, for this decision." A screen-by-screen pass answers "which experience is better at this one moment."
What it costs and what stays open
Running the study is free to try, no credit card. Cobalt asks for an email when the review is ready so it can send you the link. The honest catch worth knowing up front: after 7 days a review goes read-only for you, the owner. The shared link keeps working for everyone you sent it to, comments included, and nothing gets deleted. You just lose the ability to edit or re-export until you reopen it. Pro is $5 a month or $50 a year, keeps every review editable with no cutoff, reopens read-only reviews instantly, and cancels any time. The full breakdown is on the pricing page.
So the next time two competitors land on your shortlist, open Perplexity, name the pair and the decision in one sentence, and let the study build the matrix you would have built by hand.