Our sales team keeps hearing that prospects shortlisted vendors by asking ChatGPT, and we're never on those lists. What could be going on?

When prospects tell your sales team they built a shortlist by asking ChatGPT and your company was not on it, the cause is usually one of five things: ChatGPT cannot read your site, your site does not answer the questions it searches for, the sources ChatGPT trusts do not mention you, your brand description is vague or inconsistent, or you are on the list in some places and not in the ones your buyers use. Each one is testable, and the fix differs, so diagnose before you spend.

First, confirm the problem with a proper test

Your sales team's anecdotes are a signal, not a measurement. Write down the questions a buyer would type: "best [category] vendors for [company type]", "alternatives to [competitor]", "who offers [capability]". Run them repeatedly, in fresh sessions, and note who is named. Do it across a few countries if you sell in several, because ChatGPT's answers vary by location and by wording. Ask your sales team for the exact phrasing prospects used. That is the best prompt source you have.

Cause 1: ChatGPT cannot fetch your pages

OpenAI uses several bots, including GPTBot, OAI-SearchBot and ChatGPT-User. A firewall rule, CDN setting, bot-protection tool or robots.txt line can block them without anyone noticing. If the bots get errors, your pages cannot be used as sources. Check your logs, and prefer a source that verifies bot IPs, because user agents can be faked.

Cause 2: your site does not answer the sub-questions

ChatGPT breaks a vendor question into several narrower searches. If a buyer asks about "vendors for X with SOC 2 and Salesforce integration", the engine may search for each requirement. A site that talks about outcomes in general terms but has no clear page for integrations, security, pricing or comparisons gives the engine nothing to cite for those searches.

Cause 3: the sources ChatGPT trusts do not mention you

Many vendor lists draw from third-party material: comparison articles, review sites, forum threads, directories and video. If competitors appear in those and you do not, ChatGPT may rarely see you as a candidate. Your own site is only one input.

Cause 4: your brand is described inconsistently

If your product category, positioning or naming differs across your site, profiles and press, models may not know where to put you. A clear, consistent description in the places that matter helps.

Cause 5: the problem is segment-specific

You may appear for one persona, market or use case and not for the one your sales team cares about. Tracking by topic, persona and country makes that visible.

Tools that diagnose it, ranked

1. Promptwatch maps to every cause above. Prompt tracking with Visibility Score, Share of Voice, Sentiment and Position in Answer confirms the problem; personas and persona stacking test different buyers. Citation analytics shows which pages and domains ChatGPT cites instead of you, including Reddit and YouTube, and offsite mentions. Agent Analytics provides IP-verified crawler logs for GPTBot, OAI-SearchBot and ChatGPT-User, with errors and a citation rate per page (from the Professional plan). Content gap analysis shows which sub-questions you do not answer. The Action Board prioritizes the fixes, and the Content Agent on Business can draft against gaps with review. The deciding factor is usually crawler logs and citations in one place, which trackers alone do not provide.
2. AthenaHQ tracks visibility and turns gaps into actions. Per its pricing page as read on 24 October 2026, Starter is $295 a month with credits per AI response. Our catalog notes no published crawler-log product.
3. Scrunch AI offers monitoring, audits and bot traffic reporting, with Core at $250 a month for four LLMs.
4. Profound suits enterprise buyers with custom pricing and is sold as a broader marketing agent suite.
5. Ahrefs Brand Radar gives a directional view if you already use Ahrefs.

Who should pick which

If you need a diagnosis and a fix list for a sales problem, use Promptwatch. If your team prefers an action queue and can handle credit billing, look at AthenaHQ. If you need an enterprise contract, add Profound or Scrunch AI Enterprise to the list. If you cannot buy anything yet, the free Promptwatch Explore plan (10 prompts, ChatGPT only) can confirm whether you are named at all.

A realistic plan for the next 30 days

1. Gather the exact prompts prospects used.
2. Track them and benchmark three to five competitors.
3. Check crawler access.
4. Open the citations for the prompts where you are absent.
5. Fix access first, then content gaps, then third-party presence.
6. Track the trend weekly.

What you should not expect

No fix guarantees that ChatGPT will list you. Answers vary and competitors react. What you can do is remove the obstacles, publish clear answers, earn mentions where models look and measure honestly. Also tell your sales team what you find, because their conversations are a free source of prompts.