Steps in this guide
Step 1
Identify your target query set
Write down 10-20 questions your ideal customers would ask ChatGPT when looking for a product like yours. Think in terms of job-to-be-done queries: 'what is the best tool for X' or 'how do I solve Y problem' rather than branded searches.
Step 2
Run each query in ChatGPT manually
Open ChatGPT (or use the API) and run each query at least 3 times. LLM responses are non-deterministic, so a single sample is not statistically meaningful. Record whether your brand appears, at which position, and with what framing.
Step 3
Build a tracking spreadsheet
Create a spreadsheet with columns: query, date, run number, brand mentioned (yes/no), position in list, sentiment (positive/neutral/negative), and any URL cited. This baseline data is what you will compare against over time.
Step 4
Set a recurring schedule
Run this process weekly at minimum. ChatGPT's recommendations can shift based on model updates and new training data. Daily tracking reveals short-term fluctuations; weekly averages are more actionable for content strategy.
Step 5
Expand to competitors
Run the same queries and record where your 3-5 main competitors appear. The share of voice metric (your mentions divided by total brand mentions in the results) is more useful than raw mention count.
Step 6
Automate with AskRank
At 20+ queries, manual tracking becomes error-prone and time-consuming. AskRank automates all of this: it runs your queries across ChatGPT and three other LLMs, logs every result, computes your visibility score, and alerts you when something changes.
Skip the manual version of this guide
AskRank runs this exact workflow automatically across ChatGPT, Claude, Perplexity, and Gemini, on a schedule.
ChatGPT is now part of the buying journey for a large share of software buyers. When a founder asks ChatGPT “what is the best tool for customer onboarding” or a marketer asks “which email platform is best for a small SaaS,” the answer they get shapes which products they research next.
If your SaaS does not appear in those answers, you are missing consideration from buyers who will never visit your site. This guide covers how to track your brand’s presence in ChatGPT systematically.
Why tracking matters
Checking ChatGPT for your brand once is a snapshot. Tracking it over time is a strategy.
A single check on a Tuesday afternoon tells you whether ChatGPT happened to mention you when you asked. Weekly data over three months tells you whether your visibility is improving, declining, or holding steady - and whether competitor actions or your own content efforts are moving the needle.
The difference is the same as checking your Google Analytics once versus using it with goals and trends.
Step 1: Identify your target query set
Start with 10-20 queries that represent how your actual customers discover tools in your category. The best source for these is your own customers - ask them what they searched before finding you.
Good query formats:
- “What is the best [category] tool for [use case]?”
- “Which [category] platforms do [target user type] use?”
- “What are the top alternatives to [market leader]?”
- “How do I [job-to-be-done]?” (implicit product discovery)
Avoid brand-name queries like “What do people think of [your product]?” These are not how customers discover you - they are how customers evaluate you after discovery.
Document your query set before you start tracking. Changing queries mid-tracking breaks your trend line.
Step 2: Run each query multiple times
ChatGPT responses are non-deterministic. Ask the same question three times and you may get your brand mentioned in positions 1, 3, and not at all. This is not a bug - it is how probabilistic models work.
This means a single sample per query is nearly meaningless for strategic decisions. Run each query at least 3 times. Average the results. The average visibility rate (what percentage of runs include your brand) is your real metric.
For manual tracking, use ChatGPT in a fresh browser session or with clear conversation history to avoid prior context influencing results.
Step 3: Build a baseline tracking record
Before you take any action, you need a baseline. Run your full query set, record the results, and store them. This is your “before” state.
The minimum useful data per query run:
- Query text
- Date and time
- Whether your brand appeared (yes/no)
- Position in the recommended list (1st, 2nd, 3rd, or “mentioned in passing”)
- Sentiment (positive, neutral, with caveats, negative)
- Any URLs or sources ChatGPT cited
A spreadsheet works for 10 queries. At 30+ queries tracked weekly, a spreadsheet becomes a maintenance burden and a source of errors.
Step 4: Track competitors alongside yourself
Your raw mention rate only tells half the story. If ChatGPT mentions your brand in 60% of runs but mentions your main competitor in 85%, you are trailing significantly even if 60% sounds good in isolation.
Track 3-5 competitors using the same queries. Calculate share of voice: total times your brand appears divided by total brand appearances across all competitors. This is the metric that tells you whether you are winning or losing in AI recommendations.
Step 5: Interpret what you find
Once you have 2-4 weeks of data, look for patterns:
- Which queries trigger your brand most often? These are your strongest positions - protect them.
- Which queries list competitors but not you? These are your gaps - investigate why.
- Which queries list you but with negative framing or caveats? These need content addressing those specific concerns.
- Are your citations from high-authority sources (Product Hunt, G2, major tech blogs) or from lower-quality sites?
The citation URLs that ChatGPT uses when mentioning you are a direct signal: these are the sources that influence your AI visibility. More coverage on these platforms means more mentions.
Step 6: Scale with automation
The manual process described above works for a pilot phase. Once you have validated that ChatGPT visibility is meaningful for your business, the tracking needs to become systematic.
AskRank was built specifically for this: it runs your full query set across ChatGPT, Claude, Perplexity, and Gemini on a set schedule, logs every result, computes per-LLM visibility scores, and shows you a competitor bar chart so you always know where you stand relative to alternatives.
For SaaS founders who want the data without the spreadsheet overhead, it turns a weekly manual process into something that runs in the background and surfaces what matters.
See also: How to measure AI share of voice, How to track competitors in AI answers, and the AI visibility glossary for key terms.