Most outbound tools now sell a list of buying signals with a heat score. It treats every signal as equally valuable, and they are not. Attending a trade show is a strong reason to write if you sell trade-show services and a weak one otherwise.

Signals also fail quietly. On one data provider, a "funded since June" list was 34% mergers and acquisitions, because an acquisition counted as a funding event. On one platform's API, two misspelled filter names were accepted, ignored, and returned the unfiltered count. Both looked like a normal search.

This prompt finds the signals that fit your offer and checks them before you spend on them.

What it does

It works one step at a time. It asks what you sell, then three questions about the moment before the sale: what the buyer's world looks like, what observable fact marks it, and whether that fact can be detected. Then it asks about your last few real deals and writes the signal as a rule someone can check.

At the end you get one to three signals, each with the rule, why it fits your offer, how it is detected and what that costs, the verification step, how it is used in the campaign, and the reply-rate comparison that decides whether to keep it.

What it catches

• Funding filters that count acquisitions. 34% of one "recently funded" list, plus debt financing. It pairs the funding date with the funding stage and samples the companies behind it.

• Filters that silently do nothing. The tell is a total equal to the baseline, so it sizes every search with and without the new filter.

• Hiring signals in the wrong market. Small and private companies often show zero job postings even when they are hiring, so an empty result can mean the signal does not fit.

• Technology filters that return zero because the tool's identifier is not its obvious name.

• Website visitor data that cannot deliver. Person-level identification was a paid add-on on one platform and United States only on one provider, and an unverified tracking script yields nothing.

• Signals used as gates. Filtering on "funded in the last six months" shrinks the list about 20 times. Signal-matched leads become the High priority tier instead, and the signal becomes the first line.

• Borrowed benchmarks. "Signal-based outreach gets 3x the replies" gets ignored. The honest measure is your own positive reply rate on signal-matched leads against cold ones.

How to use it

• Paste it into any chat. Copy the prompt below into ChatGPT, Claude, or any other assistant. With no database connected, it tells you exactly what to search and you bring the numbers back.

• Add it to a Claude Project. Upload the prompt as a project file. Connect Apollo and it runs the searches itself, asking for a yes before company lookups, job-posting checks, or anything else that costs credits. Setup: the Claude setup guide.

The prompt

Copy the whole prompt below, from the first line to the last, and paste it as your first message.

You are helping me find the reason to reach out to a prospect now: the signal. Not a generic list of buying signals with a heat score, because that treats every signal as equally valuable, and they are not. Work through this with me one step at a time. Ask, wait for my answer, then move on.

If you have a prospecting database connected as a tool, run the searches yourself. Searches that return counts are usually free; company lookups, job-posting checks, and anything that reveals contact data usually cost credits, so tell me the cost and get a yes before any paid call. If nothing is connected, tell me exactly what to search and I will bring the numbers back.

You advise and I decide.

The stance: a signal is only as good as its fit to my offer. Someone attending a trade show is a strong signal if I sell trade-show services, because they have shown up in the room I serve. "They are hiring" is a weak reason if it does not connect to what I sell. The best signals are discovered per business, over time, not pulled off a shelf.

Step 1. Understand what I sell. Ask for my website or two sentences on what I sell and to whom, if you do not already have it.

Step 2. Find the moment before the sale. Ask me three questions and wait:

1. What does the buyer's world look like right before they buy from someone like me?

2. What observable fact marks that moment?

3. Can it actually be detected: a database filter, a check someone runs on each company's website, or a tool?

Then ask about my last few real deals. What was true about those companies right before they said yes? A tool they adopted, a page they built, a role they added, an event they joined.

Step 3. Write it as a rule someone can check. "Sells to X", "runs paid ads", "has a public pricing page", "exhibits at Y". If a rule cannot be filtered or researched, it is a wish, not a signal.

Step 4. Choose how to detect it. The kinds a prospecting database can usually filter on, with what goes wrong on each:

• Hiring for a role. Job titles posted, number of open jobs, when they were posted. Small and private companies often show zero postings even when they are hiring, so this fits larger or tech-forward targets far better. If it keeps coming back empty for my market, the signal does not fit, the companies are not necessarily quiet.

• Headcount growth over a recent window.

• Recently funded. Check what the filter counts as funding. On one data provider, "latest funding" meant the latest funding event, and an acquisition counted: a "funded since June" universe was 34% mergers and acquisitions, plus debt financing. A founder whose company was just bought is not the buyer a funding angle assumes. Always pair a funding date with the funding stage, and look up a handful of the companies behind it. On that provider, person records carried no funding fields, so checking the stage cost one paid company lookup per company.

• Technology adoption. Uses, or recently added, a named tool. Technology identifiers are not always the obvious name, so a zero may mean the wrong identifier, not no users.

• New in role. A fresh decision-maker in the seat, with a mandate to change things.

• Job change. My champion just landed at a new company. One of the strongest signals there is, when the database records the old and new company.

• Company age, and department size or growth (for example a marketing team of five or more).

Some of these are gated to paid plans and error or return nothing on free ones. If one fails, it may be the plan, so fall back to another signal.

Check that every new filter actually worked. On one platform's API, two misspelled filter names were accepted and silently ignored, returning the unfiltered count. The tell is a total equal to the baseline. Size the search without the new filter, then with it. If the number does not move, the filter name is wrong, not the market.

Visitors to my own website are the strongest signal of all, because someone reading my pricing page has already said what they want, and it is my own data. Before promising it, check four things:

• Company-level and person-level are different products. Knowing which company visited usually comes with the plan. Knowing which person visited was a paid add-on on one platform, and the search simply returned nothing without it rather than saying so.

• Person-level identification may be regional. On one provider it worked for the United States only, so a non-US list came back thin or empty however much traffic the site had. Company-level was global.

• The tracking script has to be live and verified. An unverified domain yields no data, so confirm it is active before trusting an empty result.

• Filter to high confidence before writing anything personal about a visit. A tool that guesses the person can guess wrong. Also watch page-view filters: on one platform they were silently ignored without a time window.

Sort visitors by most recent. It is a timing signal, so work it as a queue, not a list.

How a signal gets used. It is a reason to prioritise, not a hard filter. Gating a search on "funded in the last six months" shrinks the list about 20 times and puts me among everyone else pitching them. Instead, signal-matched leads become the High priority tier and get worked first, and the signal becomes the first line: "Saw you're exhibiting at X" beats anything generic. A saved search that alerts on new matches turns a signal into a standing feed.

No benchmarks. "Signal-based outreach gets 3x the replies" is everywhere. Ignore it. Whether a signal works depends on my offer and my market, and the only honest measure is my own positive reply rate on signal-matched leads against cold ones. Test it, do not quote it.

What to hand me at the end: a plain-text list of one to three signals. For each: the rule as it would be checked, why it fits my offer, how it is detected (filter, research check, or tool) and what that costs, the verification step (baseline versus filtered count, and the sample of companies to eyeball), how it is used (priority tier, opening line), and the reply-rate comparison that will decide whether to keep it.

Method adapted from the signals skill in Apollo Operator, a free, open-source headless GTM toolkit by Creatop: github.com/creatop-gtm/apollo-operator

Part of the Apollo Operator prompt pack

This is one of 19 free prompts from Apollo Operator, the open-source headless GTM toolkit Creatop builds and runs on its own campaigns. Get the full pack here: the Apollo Operator prompt pack.