OperStack vs Profound: weekly ChatGPT and Perplexity checks for $35

What are OperStack and Profound for?

OperStack offers a published $35 monthly price for 10 weekly buyer questions in ChatGPT and Perplexity, page checks and written fixes. Profound’s brand pricing page lists a 7-day trial and a quoted Enterprise contract. Watch suits an owner seeking a defined recurring task; Profound includes broader marketing and enterprise capabilities under a different scope.

1. A price before any sales call. Watch costs $35 a month, or $299 for a year, which works out at about $25 a month, and the first 7 days are free. On 3 October 2026, Profound’s pricing page offered brands just two things: a free 7-day Trial and an Enterprise plan behind a “Get a Demo” button. Agencies see Agency Growth at $99 a month, plus $399 a month for each full client workspace added to it.

2. A free check that asks for nothing but your address. The OperStack check at /ai-visibility/ needs no signup and no install. It samples up to ten pages from your sitemap, scores them, then puts three questions a buyer of your kind of business would ask to ChatGPT with web search switched on, and keeps the result at a permanent link. Profound’s free AEO Report asks for an email address before it runs, and its Trial needs an account.

3. Ten of your pages, measured again every week. Watch scores ten pages out of 100 and repeats that measurement in every letter, so an edit you made on Tuesday can show up as a moved number the following week. Profound’s help pages describe a readiness grade per page but give no figure for how many of your pages a crawl covers.

4. Names you can look up. Each letter lists the firms ChatGPT and Perplexity recommended in your place across the ten questions, with a count for each, including firms you may never have thought of as rivals, and scores three rivals on the same 100-point scale as you. In the report shown above, two firms were each named three times out of ten. A name is something you can open in a new tab and study this morning.

5. The paragraph, written, with no credits to spend. When a page has no short answer at the top, OperStack drafts that opening paragraph using only what your own pages already say, and the free check puts a “Copy the text” button beside it. Watch selects one priority page fix each week, with its page address and wording to review. Profound’s drafting runs through Agents that consume credits: “limited” on the Trial, 400 a month in an agency client workspace.

6. One letter, and it says what moved. Once a week, dated from your own subscription day rather than a fixed Monday, Watch writes to you with what changed against the previous week. Profound does not describe a built-in email report. Its help pages show how to assemble a weekly reporting Agent that posts to Slack, and an Outlook connection lets an Agent send mail.

7. No developer and no integration. Most Watch fixes are sentences to paste, and the few that need code are flagged as such. Profound’s prompt tracking is listed as “Instant” to set up, while its bot and referral analytics require a CDN or log connection (Cloudflare, AWS, Vercel, Fastly and others) and its Agents are built in a node-based editor.

8. You can leave in one step. Watch cancels whenever you choose; inside the first 7 days you pay nothing, and afterwards you are simply not charged again. Profound’s Master Subscription Agreement renews an order for another term unless notice arrives at least 30 days before renewal, and it calls fees non-refundable.

What is Profound, and what is OperStack?

Profound is an AI marketing platform for large brands; OperStack is a weekly set of ChatGPT and Perplexity answers and a priority page fix for businesses that live on inbound enquiries.

OperStack: the answer to one question, every week

OperStack tells you whether ChatGPT names your business when a buyer asks (Watch adds Perplexity), who is named instead, and what to change on which page. The free check gives a first reading at no cost. Watch, the subscription, repeats the work every seven days with ten questions, three rivals and a priority page fix. Maksim Shchegolev, who owns it, answers support himself at info@oper-stack.com.

How much does Profound cost in 2026?

On 3 October 2026, Profound published a 7-day brand trial and custom Enterprise pricing, so there is no current public brand subscription total to compare with Watch. OperStack costs $35 monthly or $299 annually.

The only brand-tracking figures Profound publishes are for agencies: Agency Growth at $99 a month, plus $399 a month for each full client workspace. A brand buying for itself has no public monthly figure to compare. Watch is $35 month to month, or $299 paid once for the year; both options sit on our pricing page.

A note on Profound’s Trial: it needs an account created with an email address, and the pricing page does not say whether a card is taken.

How does each tool check AI visibility?

Profound lists 50 prompts on its trial, with daily checks across 3 engines, and custom coverage on Enterprise. Watch asks ChatGPT and Perplexity 10 questions weekly and separately scores up to 10 pages. Their page and answer measurements serve different purposes; neither a higher score nor more prompts proves a commercial result.

The OperStack score has five parts with fixed weights, so a change is easy to trace. The first 25 points go to access: whether your robots.txt lets in each of 14 named AI crawlers and fetchers. The agent map, worth 15, checks for a working llms.txt. Entity, worth 20, looks for Organization or business schema, FAQPage, the page type, a canonical tag and Open Graph. Quotable content, worth 25, wants at least 300 words, a short answer with a figure right after the H1, three or more H2 sections and a table. Dates and sources, worth 15, look for dates in meta or schema, phrases that name a source, and lastmod in the sitemap. An SEO basics score on the same pages sits beside it.

The names matter as much as the number. When the free check asks ChatGPT its three questions, it files each name it finds under one of three labels: a genuine rival, a name that is no rival at all (a portal, say, or a law firm the answer happened to cite), or a name the check could not place. Watch keeps that count across ten questions, so a rival who turns up three times out of ten stands out from one who appeared once.

Why ten questions, and why weekly? ChatGPT answers can vary between runs: put one buyer question to it in the morning and again after lunch and the firms it lists can differ. A single check is therefore a sample. Ten fixed questions, asked again every seven days, turn that noise into a trend you can act on. An illustrative move from 5 of 10 to 7 of 10 would appear in the record, but that alone would not establish that an edit caused it.

Whether assistants visit your pages is a different question from whether they name you. For that we run a free AI visit counter: a line on your pages counts visits from AI systems, and the MCP server can read your own count back with a private token. It is a counter, not an analytics suite, and it costs nothing.

Profound’s Enterprise engines are ChatGPT, Perplexity, Google AI Mode, Gemini, Microsoft Copilot, DeepSeek, Claude, Google AI Overviews and Exa Search. Watch asks ChatGPT and Perplexity only, and we would rather say that plainly than blur it.

What lands in your inbox on a Monday morning?

A Watch subscriber receives 1 weekly email with findings and suggested edits, while Profound describes dashboards and configurable reporting Agents. The sample above shows our email format. Receiving a draft does not publish it: an owner must confirm the wording, apply the chosen change and preserve a record for later review.

OperStack Watch follows the same 10 buyer questions through recurring ChatGPT and Perplexity observations and records who is named in the answers. It measures 3 rivals alongside the business, checks a sample of pages and supplies suggested edits tied to findings. The next report compares the observations with the previous week, giving an owner a continuing record rather than a single score. The purpose is to make the next action visible: which page needs review, what wording is proposed and who can approve it? A business with an Enterprise reporting requirement may keep a wider platform for that separate scope. The Watch record should retain its limits when shared internally, particularly the question set, sampled pages and two-engine coverage. Review the draft against the current offer before publication. A later change in an answer is useful evidence to investigate, but it does not isolate the effect of that edit from other changes.

1. Mark the limited question and page scope in shared reports.
2. Approve a draft before handing it to the site editor.
3. Review the following observations without promising a particular outcome.

When you need a different kind of tool

Profound Enterprise covers organisational requirements such as access controls, support commitments and up to 9 answer engines. Those can matter to a procurement team managing a large brand. Watch instead gives an owner a defined weekly ChatGPT and Perplexity report for $35, without requiring the purchase of a broader marketing platform.

1. Keep required access controls and contractual service levels in the purchasing specification. Do not infer them from a monitoring feature list.
2. Assess content Agents and CMS staging when several editors must approve work before publication. That workflow is outside Watch’s email delivery.
3. Preserve additional engine or shopping coverage if it supports existing reporting. Watch checks buyer questions in ChatGPT and Perplexity and does not reproduce Profound’s full Enterprise scope.

Keyword and backlink investigation still require their own data sources. For those jobs, our Semrush comparison explains where an SEO suite fits next to a smaller recurring report.

Which one fits a clinic, an agency with ten clients, a SaaS company or an online shop?

Across 4 business scenarios, a published scope helps a buyer decide what requires a larger platform. Watch covers a limited question set in ChatGPT and Perplexity and sampled pages. A clinic, a client agency, a software firm and a store should separately assess permissions, engine coverage, catalogue size and the person approving edits.

1. A clinic owner. A dental or aesthetics clinic needs to know whether ChatGPT and Perplexity suggest it when a patient asks for a clinic nearby, and which clinic they suggest instead. That is ten questions a week and a paragraph to paste. Profound’s brand route starts with a demo and an Enterprise contract. Pick OperStack, and run the free check first to see where you stand today.

2. An agency with ten clients. On Profound’s list prices, ten full client workspaces plus the $99 Agency Growth base come to $4,089 a month. Ten Watch subscriptions, one per client, come to $350 a month; our pricing page lists no agency bundle, so that is the plain sum. Each client gets its own weekly letter with names and ready text your team can ship the same day. For pitches, our free MCP server compares two to four sites on one scale from inside Claude, ChatGPT or Cursor, without an account.

3. A SaaS company. A software company whose buyers ask ChatGPT or Perplexity for tools in its category gets the names of the tools recommended instead and the text to change on the product pages. A founder can keep part of the loop inside Cursor or Claude Code: the MCP server scores the site, and a local npm package runs our sixteen content gates over a folder of Markdown before anything is published.

4. An online shop. A shop’s category, about and delivery pages decide whether an assistant has anything to quote. OperStack reads ten pages of the site, scores them and writes the missing opening paragraphs, which is the part of an AI answer a small shop controls. ChatGPT Shopping results, which Profound lists on Enterprise, are not among what Watch checks.

For other trackers in this space, see OperStack vs Peec AI and OperStack vs Otterly.AI.

How do you move from Profound to OperStack?

Before moving from Profound, preserve useful exports and inspect the renewal clause: its published agreement requires at least 30 days’ notice of non-renewal. Watch’s first 10-question report starts a separate record. Check that its weekly ChatGPT and Perplexity scope covers your current decisions before allowing any required monitoring to lapse.

1. Find the notice date. Section 10.1 of Profound’s Master Subscription Agreement, updated 5 June 2026, renews an order for the renewal term unless notice of non-renewal is given at least 30 days before it starts. Put that date in a calendar with a week to spare. Section 8.1 says fees are non-refundable.
2. Export what you have. Enterprise keeps all-time history and exports as CSV or JSON. Save your prompt list and competitor data before access ends; it is your record of the months you paid for.
3. Take a baseline. Run the free check the day you give notice. It returns a score for ten pages and three ChatGPT answers with the names given, which becomes the starting line for Watch.
4. Start Watch. The first 7 days cost nothing, and letters then go every 7 days from the day you subscribe. If you keep Profound until the term ends, both can sit in the same Claude Code or Cursor session, since each has an MCP server.

Leaving OperStack is simpler: cancel whenever you choose. A period already paid is not refunded, as our refunds page says, and nothing is charged inside the first week.

Operator note: request a scope you can compare

Before a sales conversation, write down the decisions the report must support: which customer questions, which market and which person will approve page changes. Ask for the smallest paid scope that covers them, including the contract term and any usage limits. Profound’s unpublished brand price cannot be inferred from an agency workspace price. Keep the written quote with the specification, then compare it with Watch’s published weekly scope rather than with an assumed enterprise bill.

What a one-page result can actually establish

The saved OperStack run read 1 page on tryprofound.com on 26 September 2026. It produced a readiness score, but that sample cannot describe Profound’s entire website or validate its product. The useful lesson for a buyer is to preserve the observed page count alongside every score and recommendation.

Our run recorded full crawler-access points but no agent-map points. The content and entity observations belong to that same limited sample. They are included here as our own measurements, with the page coverage stated rather than implied.

The table deliberately does not turn these areas into a comparison of software quality. A marketing homepage can score differently from a help centre built by the same company, and a failed discovery step can limit the pages available to the checker. A reviewer needs the URLs and findings before making a wider claim.

For an owner evaluating a report, ask whether the specific commercial page you want to improve appears in the sample. If it does not, request or run the relevant check rather than borrowing the homepage result. For a larger catalogue, a broader audit still needs an explicit coverage limit.

Watch’s value is the recurring task and proposed wording on the pages it examines. It is not certification of the rest of the site. Keep that boundary with any exported report, and review the actual passage before publishing an edit. Our free check makes the method visible before a subscription; the MCP page describes the available free inspection tools.