> ## Documentation Index
> Fetch the complete documentation index at: https://docs.oppin.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Tracking

> Let the assistant tune your setup. Prompts to paste into Claude or any assistant connected to Oppin.

Paste a prompt into an assistant connected to Oppin ([how to connect](https://docs.oppin.ai/mcp/overview.md)) and replace the parts in \[brackets] with your own. In Oppin, the **MCP** tab of the **Data & MCP** page has the same prompts with your brand and a competitor already filled in.

## Quick questions

One question, one answer. Paste it as it is.

### What should we work on next?

*Quick question*

Oppin's top recommendations, each with the evidence behind it.

```text wrap theme={null}
Using Oppin, what should we work on next?
```

### Which brands does AI mention that we aren't tracking?

*Quick question*

Untracked brands, most mentioned first.

```text wrap theme={null}
Using Oppin, which brands does AI mention that we aren't tracking?
```

### How much of our plan are we using?

*Quick question*

Brands, questions, answers this month and Claude seats vs your limits.

```text wrap theme={null}
Using Oppin, how much of our plan are we using?
```

## Quick changes

These change what Oppin tracks, so they need a connection with write access. The assistant lists exactly what it will change and waits for your yes.

### Stop tracking the questions about \[old product].

*Quick change*

Finds them, lists them, stops them after you confirm. History is kept.

```text wrap theme={null}
Using Oppin, stop tracking the questions about [old product].
```

### Add \[new competitor] as a competitor. Their site is \[domain].

*Quick change*

Added. Past answers count straight away, because matching happens when you read.

```text wrap theme={null}
Using Oppin, add [new competitor] as a competitor. Their site is [domain].
```

### AI sometimes calls us "\[variant]". Count that as us.

*Quick change*

Name added. Past answers recount.

```text wrap theme={null}
Using Oppin, AI sometimes calls us "[variant]". Count that as us.
```

### Put every pricing question in a "Pricing" topic and tag them "bottom funnel".

*Quick change*

Topic created, questions moved and tagged.

```text wrap theme={null}
Using Oppin, put every pricing question in a "Pricing" topic and tag them "bottom funnel".
```

### Mark the schema markup recommendation as done.

*Quick change*

Status updated.

```text wrap theme={null}
Using Oppin, mark the schema markup recommendation as done.
```

### What did AI assistants change in our tracking this week? Undo \[the change].

*Quick change*

A receipt for every change, and undo for most of them.

```text wrap theme={null}
Using Oppin, what did AI assistants change in our tracking this week? Undo [the change].
```

* **Take it further:** see [Monday tracking check, approved in Slack](https://docs.oppin.ai/mcp/use-cases/tracking.md#monday-tracking-check-approved-in-slack).

## Workflows

A few steps and one useful result.

### Website to new topics

*Workflow · Answer (3 proposed topics) + topics and questions added on approval*

Find what your website sells that your tracking doesn't cover yet, and start tracking it.

```text wrap theme={null}
Using Oppin, read our website and find the products, use cases and audiences we sell to that none of our tracked topics cover. Propose 3 new topics with 5 questions each, written the way a buyer would ask, and say which countries and AI assistants each should run on. Check it fits our plan, and add them after I approve.
```

* **You get:** Three topics with five buyer questions each, the countries and assistants for each, and the room left on your plan. The ones you approve are added, and their first answers arrive within minutes.
* **Why it helps:** Setup stops being a one-time guess made at onboarding. When the site changes, tracking follows.
* **Changes tracking:** needs a connection with write access. The assistant shows the changes and waits for your yes.
* **Take it further:** see [Refresh our brand profile](https://docs.oppin.ai/mcp/use-cases/tracking.md#refresh-our-brand-profile) and [Hidden questions to new prompts](https://docs.oppin.ai/mcp/use-cases/hidden-demand.md#hidden-questions-to-new-prompts).

### Find every name AI uses for us

*Workflow · Answer (name by name, with examples) + names added on approval*

Count short forms, product names and old names as the brand they belong to.

```text wrap theme={null}
Using Oppin, look through the brands AI names in our answers and find the ones that are really us or one of our competitors under another name (short forms, product names, old names, typos). Show me each with an example answer, then add the ones I confirm and show our share of voice before and after.
```

* **You get:** Every other name AI uses, who it belongs to and an example answer. The ones you confirm are added and past answers recount straight away, so share of voice updates in the same chat.
* **Why it helps:** Mentions split across several names undercount a brand. This is the one change whose effect you see immediately.
* **Changes tracking:** needs a connection with write access. The assistant shows the changes and waits for your yes.
* **Note:** A name that already belongs to another tracked brand is refused, so mentions never move to the wrong brand.
* **Take it further:** see [AI sometimes calls us "\[variant\]". Count that as us.](https://docs.oppin.ai/mcp/use-cases/tracking#ai-sometimes-calls-us-variant-count-that-as-us) and [Keep the competitor list current](https://docs.oppin.ai/mcp/use-cases/tracking.md#keep-the-competitor-list-current).

### Measure what you ship

*Workflow · Answer (the question list) + questions added and recommendations completed on approval*

Turn "we did it" into "it worked" for every recommendation you finish.

```text wrap theme={null}
Using Oppin, for each recommendation we finished this month, suggest 2 or 3 questions that should move if it worked (a comparison page against [competitor] should win "[brand] vs [competitor]"). Show me the list first, then add them tagged with the recommendation and mark each recommendation completed. Next quarter, show me whether those questions moved.
```

* **You get:** Two or three tagged questions per finished recommendation, the recommendations marked completed, and next quarter a before and after on exactly those questions.
* **Why it helps:** Your own work, shown paying off question by question, instead of hoping a headline number moves. It is also what the quarterly review reads.
* **Changes tracking:** needs a connection with write access. The assistant shows the changes and waits for your yes.
* **Note:** Add the questions as soon as the work ships, so there is a baseline. The check next quarter is a second request: ask again then, or set it as a scheduled task in your assistant.
* **Take it further:** see [Quarterly review: did our work move anything?](https://docs.oppin.ai/mcp/use-cases/reporting.md#quarterly-review-did-our-work-move-anything) and [Recommendation triage](https://docs.oppin.ai/mcp/use-cases/tracking.md#recommendation-triage).

### Launch kit

*Workflow · Answer (the question list) + topic and questions added on approval*

Start tracking a launch before it happens, so the before and after is real.

```text wrap theme={null}
Using Oppin, we're launching [product] on [date]. Here is the landing page: [URL]. Set up a "[product] launch" topic with 10 questions buyers will ask (finding it, comparing it with [competitor], pricing), on ChatGPT, Perplexity and Google AI Mode in [countries], tagged "launch". If [product] has its own name, count it as us. Show me the list first.
```

* **You get:** A launch topic with 10 tagged buyer questions, and the product's name counted as yours, in place before launch day.
* **Why it helps:** "Did the launch show up in AI answers?" is the question every launch review asks now, and it only has an answer if tracking started before the launch.
* **Changes tracking:** needs a connection with write access. The assistant shows the changes and waits for your yes.
* **Note:** Set it up a week or more ahead, so the before and after has a baseline. After launch, ask "Using Oppin, how did the launch do? Compare the launch questions before and after \[date]."
* **Take it further:** see [Campaign or launch impact](https://docs.oppin.ai/mcp/use-cases/pulse.md#campaign-or-launch-impact).

### Get more from the plan

*Workflow · Answer (a swap list), applied on approval*

Swap the questions that tell you least for gaps you don't cover yet, same total.

```text wrap theme={null}
Using Oppin, we're running out of room on our plan. Find the questions that tell us least (every assistant gives the same answer, no brands named, or they duplicate another question) and the biggest gaps we don't cover. Propose a swap: stop these, add those, same total. Apply it after I approve.
```

* **You get:** A stop list and an add list of the same size, a reason for each, applied once you approve. Stopped questions keep their history.
* **Why it helps:** Every slot on the plan goes to a question that tells you something. And when the gaps outnumber the slots, you have the case for more room, in your own data.
* **Changes tracking:** needs a connection with write access. The assistant shows the changes and waits for your yes.
* **Take it further:** see [Prompt set health check](https://docs.oppin.ai/mcp/use-cases/tracking.md#prompt-set-health-check) and [How much of our plan are we using?](https://docs.oppin.ai/mcp/use-cases/tracking.md#how-much-of-our-plan-are-we-using) and [Hidden questions to new prompts](https://docs.oppin.ai/mcp/use-cases/hidden-demand.md#hidden-questions-to-new-prompts).

### Pick where Claude runs

*Workflow · Answer (the question list) + Claude on for the ones you approve*

Spend your Claude seats on the questions where Claude's answer matters most.

```text wrap theme={null}
Using Oppin, check how many Claude seats we have free, then pick that many questions where Claude's answer matters most (high buying intent, and questions we lose on other assistants) and switch Claude on for them. Show me the list first.
```

* **You get:** The questions worth a Claude seat, a reason for each, switched on once you approve.
* **Why it helps:** A seat nobody uses is paid for and tells you nothing. This fills them with the questions that matter, in one message.
* **Changes tracking:** needs a connection with write access. The assistant shows the changes and waits for your yes.
* **Note:** Needs the Claude add-on. Claude is only switched on for questions you approve, and a country Claude can't search from is refused with the reason.
* **Take it further:** see [How much of our plan are we using?](https://docs.oppin.ai/mcp/use-cases/tracking.md#how-much-of-our-plan-are-we-using) and [Which questions are we losing to competitors?](https://docs.oppin.ai/mcp/use-cases/competition.md#which-questions-are-we-losing-to-competitors).

### Right assistant for each market

*Workflow · Answer (country by country) + assistants switched on approval*

Run each country's questions on the AI assistants that matter there.

```text wrap theme={null}
Using Oppin, check which AI assistants give us useful answers in each of our countries. Where one rarely names any brand, or always agrees with another, suggest moving those questions to one that matters more there. Apply it after I approve.
```

* **You get:** For each country, the assistants worth keeping and the ones to swap, with the evidence, applied to those questions once you approve.
* **Why it helps:** Each question runs on up to three assistants. This spends them where your buyers actually ask.
* **Changes tracking:** needs a connection with write access. The assistant shows the changes and waits for your yes.
* **Note:** Only assistants that can search from a country are offered for it.
* **Take it further:** see [Where in the world are we strongest and weakest?](https://docs.oppin.ai/mcp/use-cases/pulse.md#where-in-the-world-are-we-strongest-and-weakest) and [Which AI assistant likes us most, and which least?](https://docs.oppin.ai/mcp/use-cases/pulse.md#which-ai-assistant-likes-us-most-and-which-least).

### Report by buying stage

*Workflow · Answer (visibility by stage) + tags applied on approval*

Tag every question with its buying stage and audience, then see your visibility at each stage.

```text wrap theme={null}
Using Oppin, tag every tracked question with its buying stage (learning, comparing, ready to buy) and who is asking ([our personas]). Show me the tags first, then apply them and show our visibility by stage.
```

* **You get:** Every question tagged by stage and persona, then your visibility at each stage. From then on the stage is one filter on every report.
* **Why it helps:** "Are we visible when people are ready to buy?" becomes one filter, for good.
* **Changes tracking:** needs a connection with write access. The assistant shows the changes and waits for your yes.
* **Take it further:** see [Buying-journey map](https://docs.oppin.ai/mcp/use-cases/hidden-demand.md#buying-journey-map) and [Put every pricing question in a "Pricing" topic and tag them "bottom funnel".](https://docs.oppin.ai/mcp/use-cases/tracking.md#put-every-pricing-question-in-a-pricing-topic-and-tag-them-bottom-funnel).

### Refresh our brand profile

*Workflow · Answer (current and proposed, field by field) + profile updated on approval*

Keep the description, industry and domains Oppin knows you by in step with your website.

```text wrap theme={null}
Using Oppin, read our website and compare it with our brand profile in Oppin: the description, industry and extra domains. Suggest updates so drafted questions and recommendations fit what we sell today, and apply them after I approve.
```

* **You get:** Your current profile next to a proposed one, field by field, applied once you approve.
* **Why it helps:** Drafted questions and recommendations read the profile, so a stale one quietly skews both.
* **Changes tracking:** needs a connection with write access. The assistant shows the changes and waits for your yes.
* **Note:** Extra domains count as yours in mentions and sources, so list only sites you own.
* **Take it further:** see [Website to new topics](https://docs.oppin.ai/mcp/use-cases/tracking.md#website-to-new-topics) and [AI sometimes calls us "\[variant\]". Count that as us.](https://docs.oppin.ai/mcp/use-cases/tracking#ai-sometimes-calls-us-variant-count-that-as-us).

### New brand from its website

*Workflow · Answer (the setup to approve) + brand created · For agencies*

Set up a new brand from its website: profile, competitors and starter questions, in one approval.

```text wrap theme={null}
Using Oppin, set up [new brand] as a new brand from its website [URL]. Read the site, draft the profile (what it sells, industry, country), suggest up to 5 competitors with their real websites and 10 starter questions buyers would ask, and create the brand once I approve.
```

* **You get:** A drafted profile, up to 5 competitors with their websites and 10 starter questions. Once you approve, the brand is created and its first answers arrive within minutes.
* **Why it helps:** A new client or product line is tracked the same day, set up from its own site rather than from memory.
* **Changes tracking:** needs a connection with write access. The assistant shows the changes and waits for your yes.
* **Note:** Only workspace admins (or a write key an admin created) can add a brand, and your plan's brand limit applies.
* **Take it further:** see [Copy a setup to another brand](https://docs.oppin.ai/mcp/use-cases/tracking.md#copy-a-setup-to-another-brand) and [Website to new topics](https://docs.oppin.ai/mcp/use-cases/tracking.md#website-to-new-topics).

### Copy a setup to another brand

*Workflow · Answer (the question list) + questions added on approval · For agencies*

Give a new brand the questions that already work for another one, the same day.

```text wrap theme={null}
Using Oppin, copy [Brand A]'s pricing and comparison questions to [Brand B], swapping the names and competitors, in [Brand B]'s country. Show me the list first.
```

* **You get:** The questions rewritten for the other brand, with its own name, competitors and country, added once you approve.
* **Why it helps:** A new brand is useful the same day instead of after weeks of setup.
* **Changes tracking:** needs a connection with write access. The assistant shows the changes and waits for your yes.
* **Note:** Both brands must already be in Oppin. To add one, set it up from its website first.
* **Take it further:** see [New brand from its website](https://docs.oppin.ai/mcp/use-cases/tracking.md#new-brand-from-its-website) and [One structure across every brand](https://docs.oppin.ai/mcp/use-cases/tracking.md#one-structure-across-every-brand).

### Prompt set health check

*Workflow · Answer (a graded scorecard)*

Make sure the numbers measure the market, not your own marketing.

```text wrap theme={null}
Using Oppin, grade our tracked questions. Are they balanced across topics and buying stages, or are some worded to flatter us? Quote the worst ones and suggest replacements.
```

* **You get:** An A to F grade on topic balance, funnel coverage, bias and blind spots, the questions to swap, and optional stop/add on approval.
* **Why it helps:** A self-flattering prompt set makes every other number meaningless.
* **Shortcut:** the [`oppin_prompt_audit`](https://docs.oppin.ai/mcp/commands#prompt-audit) command runs the same analysis.

### Grow the prompt set

*Workflow · Structured data (CSV in the bulk-import format)*

Add a batch of well-written questions, reviewed by the team first.

```text wrap theme={null}
Using Oppin, draft 20 new questions for our [topic] topic, check they fit our plan, and give me a CSV in Oppin's import format (prompt, models, country, language, topic, tags) so the team can review it. I'll upload the ones we keep.
```

* **You get:** A review-ready CSV that uploads straight into the Prompts page. With write access, the assistant can instead have Oppin draft them and accept the approved ones in the chat.
* **Why it helps:** It works without write access: the team reviews in a sheet, then uploads what it keeps.

### Keep the competitor list current

*Workflow · Answer + competitors added on approval*

Make sure the scoreboard compares you with the brands AI actually names.

```text wrap theme={null}
Using Oppin, which brands does AI keep naming alongside us that we don't track? Recommend which to track and which to just watch, and add them once I confirm.
```

* **You get:** Candidates with mention counts and example contexts, added on approval with their history already counted.
* **Why it helps:** A scoreboard is only as honest as its competitor list.
* **Changes tracking:** needs a connection with write access. The assistant shows the changes and waits for your yes.

### Recommendation triage

*Workflow · Answer + statuses updated*

Keep Oppin's recommendations a to-do list, not a pile.

```text wrap theme={null}
Using Oppin, go through our new recommendations. For each one, tell me what it is, the evidence behind it, and the rough effort. Suggest do now, later or dismiss, and update them once I agree.
```

* **You get:** A short do / later / dismiss list with evidence, statuses updated.
* **Why it helps:** Recommendations only help when someone decides on them. This keeps the list short and current, which the quarterly review relies on.
* **Changes tracking:** needs a connection with write access. The assistant shows the changes and waits for your yes.
* **Shortcut:** the [`oppin_opportunity_briefing`](https://docs.oppin.ai/mcp/commands#opportunity-briefing) command runs the same analysis.

### Recommendation board, both ways

*Workflow · Tasks in your project tool, statuses back in Oppin*

Turn recommendations into tasks, then mark them done in Oppin when the work ships.

```text wrap theme={null}
Using Oppin, turn our high-impact recommendations into tasks in [Linear / Asana / Notion], with the evidence and steps in each. Later: "Check which of those tasks are done and mark the matching recommendations completed in Oppin."
```

* **Why it helps:** Oppin's recommendations live where work gets done, and Oppin knows what was acted on. That record is what the quarterly review reads.
* **Needs:** Linear, Asana or Notion, connected in your assistant.
* **Changes tracking:** needs a connection with write access. The assistant shows the changes and waits for your yes.
* **Take it further:** see [Quarterly review: did our work move anything?](https://docs.oppin.ai/mcp/use-cases/reporting.md#quarterly-review-did-our-work-move-anything).

### Battlecards to competitor list

*Workflow · Answer (a change list) + competitors updated on approval*

Make the scoreboard match the rivals your sales team actually meets.

```text wrap theme={null}
Using Oppin and our battlecards in [Notion / Google Drive], compare the competitors we pitch against with the ones Oppin tracks. Add the missing ones with the website from the battlecard and the product names AI might use for them. Move the ones sales no longer sees to the watchlist. Show me the changes first.
```

* **You get:** Rivals to add and rivals to move to the watchlist, with websites taken from your battlecards. Applied on approval, with past answers counted straight away.
* **Why it helps:** "Rank 2 of 7" only means something if they are the right 7.
* **Needs:** Notion or Google Drive, connected in your assistant.
* **Changes tracking:** needs a connection with write access. The assistant shows the changes and waits for your yes.
* **Take it further:** see [Keep the competitor list current](https://docs.oppin.ai/mcp/use-cases/tracking.md#keep-the-competitor-list-current).

### Monday tracking check, approved in Slack

*Workflow · Slack message + approved changes applied*

Proposed tracking changes in Slack every Monday, applied only when someone approves.

```text wrap theme={null}
Using Oppin, every Monday find the brands AI named alongside us 10 or more times that we don't track, new fan-out queries we never show up in, and questions that stopped producing answers. Post the proposed changes in #[channel]. Apply only the ones someone approves in the thread.
```

* **You get:** One Monday post with the proposed additions and stops, a reason for each. Only what gets approved in the thread is applied.
* **Why it helps:** The setup keeps itself current without anyone having to remember.
* **Needs:** Scheduled tasks, Slack, connected in your assistant.
* **Changes tracking:** needs a connection with write access. The assistant shows the changes and waits for your yes.
* **Note:** Applying the approvals is a second run that reads the thread, so schedule one for later in the day or ask for it. To review what was applied, ask what AI assistants changed in your tracking this week.
* **Take it further:** see [Keep the competitor list current](https://docs.oppin.ai/mcp/use-cases/tracking.md#keep-the-competitor-list-current) and [What did AI assistants change in our tracking this week? Undo \[the change\].](https://docs.oppin.ai/mcp/use-cases/tracking#what-did-ai-assistants-change-in-our-tracking-this-week-undo-the-change).

## Agent tasks

Longer tasks: the assistant investigates and reasons before it writes anything up.

### Tracking strategy rebuild

*Agent · Answer (a change list), applied on approval*

A better setup for your questions, topics and competitors, applied once you approve.

```text wrap theme={null}
Using Oppin, review how we track [brand]: our questions, topics, tags and competitors. Use the fan-out queries, the brands it names that we don't track, and our plan limits. Propose a better setup: questions to add, reword or stop, a clean topic and tag structure, and competitors to track or watch. Show the full change list and wait for my approval, then apply it. Two weeks later I'll ask you to check whether the new questions are producing useful data.
```

* **You get:** A reviewed change list, applied on approval, sized to the plan.
* **Why it helps:** Setup is usually a one-time guess. This makes it a maintained system.
* **Changes tracking:** needs a connection with write access. The assistant shows the changes and waits for your yes.

### One structure across every brand

*Agent · Answer (a plan per brand), applied on approval · For agencies*

The same topics and tags on every brand in the workspace, so client reports compare.

```text wrap theme={null}
Using Oppin, apply the topic structure and tags from [Brand A] to every brand in the workspace: create the missing topics, move each brand's questions into them, tag them by buying stage, and remove topics left empty. Show me the plan per brand first.
```

* **You get:** A plan per brand (topics to create, questions to move, tags to add, empty topics to remove), applied once you approve.
* **Why it helps:** Client reports become comparable, and hours of setup become one message.
* **Changes tracking:** needs a connection with write access. The assistant shows the changes and waits for your yes.
* **Take it further:** see [Report by buying stage](https://docs.oppin.ai/mcp/use-cases/tracking.md#report-by-buying-stage) and [Copy a setup to another brand](https://docs.oppin.ai/mcp/use-cases/tracking.md#copy-a-setup-to-another-brand) and [Portfolio check](https://docs.oppin.ai/mcp/use-cases/pulse.md#portfolio-check).


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.