Recipes / Competitor customers: find companies on a competitor's product and open with what its own reviewers complain about

Competitor customers: find companies on a competitor's product and open with what its own reviewers complain about

PredictLeads + G2 + Apollo

IntermediateSDRsMarketing

Who it's for

SDRs and marketing teams who want a short list of target accounts running a competitor's product, the problems that competitor's own G2 reviewers keep raising, and the people at each account who feel those problems. Everything stays read-only and no credits are spent in Apollo. Apollo can be swapped for ZoomInfo, Cognism or Lusha.

What you get

Accounts on the competitor, its review themes and one draft opener. Sample output: every company and person in it is fictional.

Competitor: Ledgerline (fictional). Accounts checked with PredictLeads: 40. Ledgerline detected at 4.

| Company | Website | What PredictLeads returned |
|---|---|---|
| Fernhill Software (fictional) | fernhill.example | Ledgerline listed among detected technologies |
| Larkspur Freightworks (fictional) | larkspur.example | Ledgerline listed among detected technologies |
| Mossgate Labs (fictional) | mossgate.example | Ledgerline listed among detected technologies |
| Quarry Lane Analytics (fictional) | quarrylane.example | Ledgerline listed among detected technologies |

No data for 6 accounts. They're left out, not counted as non-users.

G2 reviews of Ledgerline, last 12 months: 62 read. Top complaint themes:
1. Slow month-end reports (11 reviews). "Our close reports take twenty minutes to load once the quarter fills up." Controller, mid-market, Aug 2026
2. Support replies take days (8 reviews). "Tickets sit for three or four days before anyone answers." Finance Manager, Jun 2026
3. Price jumps at renewal (6 reviews). "Renewal came in 30% higher with no new features." VP Finance, Sep 2026

None of these reviews name any of the 4 companies above. The themes describe Ledgerline reviewers in general.

Apollo people search (free, no credits used):
- Fernhill Software: Dana Okafor, VP Finance
- Larkspur Freightworks: Sam Whitlock, Controller
- Mossgate Labs: Priya Raman, Head of FP&A

Opener for Sam (draft, text only):
"A lot of finance teams on Ledgerline tell us month-end reports slow right down as the quarter fills up. If that's ever been a problem at Larkspur, I can show you how other controllers got their close reports back under a minute."

Nothing was sent, enriched or added to Apollo.

Time and cost

Estimate About 20–30 minutes for 30–50 accounts, most of it reading the review themes and openers.

Assumptions: One run over a list of about 40 target accounts, finding 3–5 on the competitor, with 2–3 people found per account.

Exact credit costs depend on your plan with each vendor. Check the tool pages before a large run.

MCP servers

Every server is confirmed in: ClaudeChatGPT

Set up in a chat app

Every server in this recipe is confirmed for Claude and ChatGPT. Connect each one, then paste the prompts below into one conversation.

Developer setup

Auth setup and mcp.json for Claude Code and Cursor
  • PredictLeads: Sign in with OAuth in your client (run /mcp in Claude Code). Do it once interactively, and the client stores and refreshes the token.
  • G2: Sign in with OAuth in your client (run /mcp in Claude Code). Do it once interactively, and the client stores and refreshes the token.
  • Apollo: Use OAuth sign-in for interactive use. For headless runs, use an API key in the X-Api-Key header.

Put this in .mcp.json (Claude Code, project root) or ~/.cursor/mcp.json (Cursor). Swap the example keys for your own, and don't commit real ones. Remove the // comment lines if your client requires strict JSON.

{
  "mcpServers": {
    "predictleads": {
      "url": "None"
    },
    "g2": {
      "url": "https://mcp.g2.com/mcp"
    },
    "apollo": {
      "url": "https://mcp.apollo.io/mcp"
      // headless alternative: "headers": { "X-Api-Key": "YOUR_APOLLO_API_KEY" }
    }
  }
}

Step-by-step prompts

  1. Using PredictLeads, check the technologies detected for each of these companies: {account list}. Keep the ones where {competitor product} shows up. For each, show the company, website and exactly what PredictLeads returned for that technology. List companies with no data separately, and don't guess for them.
  2. In G2, read the product reviews for {competitor product} from the last {months} months. Group the complaints into 3–5 themes, with a count and one or two exact quotes per theme, plus the review date and the reviewer's role if shown. Only link a review to a company if the review itself names that company. Don't create or change any research boards.
  3. For each company I keep, use Apollo people search to find 2–3 people with titles like {titles}. People search only: no enrichment and no company search. Show name, title and which review theme fits their role best.
  4. Write a two-sentence opener for each person that starts from one theme, framed as what {competitor product} users often say, not as something their company said. Show me the text only. Don't add anyone to Apollo and don't send anything.

Run it on a schedule (optional)

Where this goes wrong

Gotchas

Download

Get the prompts, guardrails and expected output as one Markdown file. Paste it into a Claude Project or ChatGPT project's instructions, or save it as CLAUDE.md for Claude Code.

Download competitor-customers.md

Vendor sites: Apollo. If we join an affiliate program, these links will be labeled as affiliate links.