The fastest place to look for a new client is often not a prospect database. It is the set of people who already know your work, have worked beside you, hired you before, introduced you to someone, or trust a person who can make an introduction.
To find clients in your existing network, map the relationships you already have, define one specific offer and buyer, score business fit separately from relationship strength, look for a credible reason to talk now, and make a small contextual ask. Do not message everyone. A short list of ten well-supported opportunities is more useful than a thousand names with no relationship path.
This guide gives you a product-agnostic process you can run with a spreadsheet. It also explains where a CRM or AI system can help and where your judgment still matters.
Last verified: August 18, 2026. The outreach-law references and linked product boundaries were checked against the official sources listed below. This is practical business guidance, not legal advice.
Why your network is different from a cold list
A cold list starts with demographic or firmographic fit: job title, industry, company size, technology, location, hiring, funding, or buying intent. Those signals can identify plausible buyers, but they do not create trust.
A warm network starts with a relationship path. The person may be a former client, colleague, collaborator, classmate, investor, community peer, vendor, referral partner, or a second-degree connection with a credible introducer. The relationship does not guarantee a sale. It changes the opening: you can begin with relevant shared context instead of inventing familiarity.
The strongest warm opportunities usually have all three of these conditions:
- Fit: the person or company plausibly needs what you sell.
- Relationship: there is real trust, shared history, or a credible introduction path.
- Timing: something changed that makes a conversation useful now.
Many network searches fail because they optimize only one dimension. A close friend with no need is not a qualified lead. A perfect buyer you met once in a crowded room is not meaningfully warm. A past client can be a strong relationship but a poor opportunity if there is no current reason to talk.
Build a relationship-source map
Do not begin by scrolling through one app and trusting whatever comes to mind. Build a source map so quiet but valuable relationships are not hidden by the people you interact with most often.
| Source | What it can reveal | What to record |
|---|---|---|
| Former clients and buyers | People who already know the quality of your work | Work completed, outcome, last contact, current role |
| Former colleagues | People who have seen how you operate | Team, shared projects, trust level, where they moved |
| Referral partners | People who serve the same buyers without competing directly | What they refer, what you can refer back, recent exchanges |
| Professional communities | Peers who know your expertise through repeated participation | Group, shared topic, meaningful interaction, permission to contact |
| Email and calendar history | Real conversations that may be absent from a contact list | Last substantive exchange, meetings, promised follow-up |
| LinkedIn connections | Career changes and first-degree professional relationships | Current role, company, connection date, shared history |
| CRM and client records | Commercial history and stalled opportunities | Past proposals, deal result, objections, renewal or re-engagement date |
| Personal notes | Context that software rarely knows | Relationship nuance, preferred channel, sensitivities, promises made |
You do not need to connect every source to a new tool. Start with data you are entitled to use. LinkedIn lets members export their own connection data, with some email addresses omitted according to each connection’s privacy choice. Google also lets account holders download selected account data, although an export can take time and includes privacy responsibilities once the file leaves Google.
The important distinction is authorization: a first-degree connection, a message history, or a contact in your address book is not permission to ignore the person’s expectations or applicable law. It is evidence that a relationship may exist. You still need to decide whether a business message is appropriate.
Define one offer and one buyer before scoring
“Anyone who needs consulting” is not a target. It is an instruction to produce noise.
A usable network-search brief has four parts:
- Offer: the concrete problem you solve and the commercial shape of the work.
- Buyer: the role or decision-maker who can recognize and sponsor that work.
- Situation: the events that make the problem urgent or visible.
- Exclusions: people or companies you should not approach.
For example:
Find marketing leaders at growing B2B software companies who may need a positioning and website reset after a new funding round, leadership change, or category shift. Exclude direct competitors, current employers, companies with an active agency of record, and anyone I dismissed previously.
That brief is narrow enough to judge. It also makes it easier to explain why someone qualifies instead of hiding behind a mysterious score.
Score fit and relationship strength separately
Do not compress every signal into one number too early. A two-axis score preserves the difference between “right buyer” and “credible path.”
Fit score
Use a simple 0–5 rubric:
| Score | Fit evidence |
|---|---|
| 0 | No plausible need or explicitly excluded |
| 1 | Industry or role is adjacent, but the offer is unlikely to apply |
| 2 | Some need is possible; too little evidence to prioritize |
| 3 | Role, company, and problem are aligned |
| 4 | Strong alignment plus a visible business trigger |
| 5 | Exact buyer, exact problem, and a current reason to act |
Fit evidence can include a relevant role, business model, team size, recent initiative, new responsibility, product launch, hiring pattern, public request, past proposal, or known operational problem. Avoid treating funding or headcount growth as proof that someone wants your service. They are prompts for research, not conclusions.
Relationship score
Use a second 0–5 rubric:
| Score | Relationship evidence |
|---|---|
| 0 | No relationship or safe path to contact |
| 1 | Weak recognition: met once, connected without a conversation, or stale group overlap |
| 2 | Real but light interaction: a useful exchange, event conversation, or distant shared context |
| 3 | Repeated interaction, mutual collaborator, or credible introducer |
| 4 | Worked together successfully, meaningful referral history, or trusted peer relationship |
| 5 | Strong current trust and a natural reason to discuss the work |
Relationship strength should reflect actual evidence, not the social platform’s label. A first-degree connection can still be cold. A second-degree contact can be warm if a close mutual collaborator is genuinely willing to introduce you.
Timing signal
Use timing as a note, not a magic multiplier. Useful events include:
- a role change that creates a new mandate;
- a past client’s move to a new company;
- a public launch, rebrand, hiring plan, expansion, or strategic change;
- a contract, renewal, project, or budget cycle you already know;
- a direct request for expertise;
- an introduction offered by someone you trust.
The best shortlist is normally people with fit of at least 3, relationship of at least 3, and one credible reason to talk. Change those thresholds for your business, but do not let high fit erase a nonexistent relationship.
Research only what you need for a respectful first touch
Once you have a shortlist, verify the current facts. Do not manufacture intimacy by collecting trivia.
For each person, answer five questions:
- What do I know because of our actual relationship?
- What has changed since we last spoke?
- Why might my offer matter in their current role or company?
- What channel is genuinely available and appropriate?
- What is the smallest useful next step?
Separate facts from inferences in your notes. “Now leads brand at Northstar” may be a fact from a current company profile. “Probably planning a rebrand” is an inference unless the company has said so publicly. Your message should not present the inference as inside knowledge.
Choose the right warm path
There are three common paths. Prefer the one that requires the least invented familiarity.
Direct shared-history outreach
Use this when you have worked together, had a substantive conversation, or have another clear relationship.
Congratulations on the new role. I was thinking about the launch work we did together at your last company. If positioning the new portfolio is on your list this quarter, I would be glad to compare notes. No pitch deck needed; even twenty minutes would tell us whether it is worth discussing.
Why it works: it names real history, makes a relevant connection to the present, and asks for a conversation rather than assuming a project.
Warm introduction
Use this when a trusted mutual contact knows both sides well enough to make an honest introduction.
You know both of us, so I wanted your read before reaching out. Their team appears to be hiring around product marketing, and that overlaps with the positioning work I do. Do you think a conversation would be useful? If not, no introduction needed.
Why it works: it gives the introducer permission to say no. A forced introduction transfers social cost to the person helping you.
Reconnection before business
Use this when the relationship is real but stale and the opportunity is uncertain.
It has been a while since the studio project. I saw that you moved into an operating role and would enjoy hearing what you are building. I have also shifted my work toward launch strategy for service teams, but this is mainly a catch-up. Would a short call in the next few weeks be useful?
Why it works: it does not disguise a pitch as friendship. If business becomes relevant, it can emerge from the conversation.
Make the ask proportional to the relationship
The weaker the relationship, the smaller the ask should be.
| Relationship | Reasonable first ask | Usually too much |
|---|---|---|
| Past client with a known new need | Brief call about the specific problem | Sending a full unsolicited proposal |
| Trusted former colleague | Advice, referral, or a direct conversation | Asking them to distribute your pitch broadly |
| Credible mutual introduction | Permission for an introduction | Treating the mutual contact’s name as endorsement |
| Light event or community contact | A relevant question or resource | Claiming you “know each other well” |
| Stale connection with no current signal | Genuine reconnection | Immediate calendar link and hard sell |
Your goal is not to prove you can personalize a message. It is to make the next step useful to the recipient.
Track a warm-lead pipeline
Warm-network work becomes inconsistent when it lives entirely in memory. Use a spreadsheet or CRM with these fields:
- person, company, and current role;
- source and relationship path;
- offer or target this person was evaluated against;
- fit evidence and fit score;
- relationship evidence and relationship score;
- timing signal and verification date;
- suggested channel;
- next action and owner;
- status: review, approved lead, opportunity, won, lost, or do not resurface;
- last contact and result;
- permission, preference, or compliance notes where applicable.
Keep “pending review” separate from an active lead. That protects your pipeline from AI suggestions, imported contacts, or rough ideas that a person has not yet judged.
Also preserve dismissals. If someone is a poor fit or should not be approached, the system should not recommend them again merely because your target wording changed.
Where AI helps and where it should stop
AI can make a network review faster by:
- normalizing records from multiple sources;
- researching a person’s current professional context;
- explaining fit and relationship evidence;
- ranking a review queue;
- summarizing relevant history;
- preparing an outreach strategy or first draft;
- learning durable preferences from approvals and dismissals;
- keeping follow-up tied to the CRM and pipeline.
AI should not decide that a relationship exists when the evidence does not support one. It should not turn a contact source into permission, hide uncertainty, or send a sensitive message simply because it can draft one.
When evaluating an AI system, ask to see the evidence behind a recommendation. Then test whether you can correct the record, dismiss a suggestion permanently, restrict data sources, and require review before an external send.
How Every handles warm-lead discovery
Every’s AI lead-generation workflow is built for a specific version of this problem: finding possible clients inside relationship data a business chooses to use, then carrying approved opportunities into an AI CRM, proposals, invoices, and payments.
Every can work with Google relationship context you authorize, your own LinkedIn connection export, CSV imports, and contacts you intentionally save. It evaluates business fit and relationship strength independently. Suggested people land in a pending review queue; they do not become active pipeline work until you approve them.
Once approved, Every can prepare the outreach strategy and a context-aware draft using an available channel. Important external sends require confirmation or an explicit instruction. The relationship history then stays connected to pipeline work, proposal drafting, invoicing, and payment status.
Every is not a bulk contact seller, a background LinkedIn scraper, or a replacement for your judgment about consent and context. If you need a large net-new cold list, use software designed and governed for that job.
Outreach law still applies to warm contacts
“They are in my network” is not a legal category.
In the United States, the FTC says the CAN-SPAM Act applies to commercial email, including individual business-to-business messages, and requires accurate sender information, non-deceptive subjects, a valid postal address, and a working opt-out process. Read the FTC’s CAN-SPAM compliance guide before treating a former client or connection as exempt.
In Canada, CASL generally requires consent for commercial electronic messages. Government guidance explains that implied consent based on an existing business relationship applies only within defined time periods and with other requirements. Review the Government of Canada’s guidance on getting consent.
Other countries and channels have different rules. If your process crosses jurisdictions, uses purchased data, or runs at scale, get qualified legal advice and configure your system accordingly.
Compliance is the floor. Relationship quality is the higher standard. A technically lawful message can still damage trust if it uses private context unexpectedly or turns a dormant friendship into an automated sales sequence.
Common failure modes
Treating every connection as warm
A platform connection proves a graph edge, not trust. Require evidence of a conversation, shared work, repeated interaction, or a credible introducer.
Scoring only fit
This produces a cold list wearing a warm label. Keep relationship strength visible and independent.
Hiding a pitch inside a fake catch-up
If the business reason is already clear, state it. If you genuinely want to reconnect, do not force a commercial agenda into the same message.
Using private context as a personalization trick
Just because data is accessible does not mean it will feel appropriate in a message. Use context the recipient would reasonably expect you to know.
Asking for too much
Do not turn one introduction into a full proposal, or one reply into an automated sequence. Earn the next step.
Measuring activity instead of progress
Drafts, sends, and response rates are intermediate metrics. Track qualified conversations, opportunities, proposals, wins, and relationship damage such as opt-outs or negative replies.
Forgetting the people you rejected
A system that repeatedly resurfaces the same poor-fit contact wastes time and erodes trust. Preserve dismissals and the reason behind them.
A 90-minute network review
You can run the entire method without software:
- Spend 15 minutes defining one offer, buyer, situation, and exclusion list.
- Spend 25 minutes collecting 30–50 names from former clients, colleagues, partners, email, calendar, communities, and LinkedIn.
- Spend 20 minutes scoring fit and relationship separately from 0–5.
- Spend 15 minutes verifying current roles and timing for the top ten.
- Spend 10 minutes choosing the appropriate path and writing five small asks.
- Spend 5 minutes creating next actions and a date to review outcomes.
The output should be a short, explained queue. If you end with hundreds of names, your target is too broad or your review threshold is too low.
Methodology and sources
This guide combines a transparent two-axis qualification method with Every’s current product behavior. It does not rely on claimed universal conversion rates or manufactured benchmarks. Product claims were checked against Every’s current public feature, privacy, and pricing pages. Data-export and outreach-law claims were checked against the linked official sources on August 18, 2026.
Primary references:
- LinkedIn Help: Export connections from LinkedIn
- Google Account Help: How to download your Google data
- US Federal Trade Commission: CAN-SPAM Act compliance guide
- Government of Canada: Getting consent to send email under CASL
- Every Privacy Policy
- Every Pricing
FAQ
How do I find clients from my existing network?
Map your former clients, colleagues, referral partners, professional communities, email and calendar history, LinkedIn connections, and CRM records. Define one offer and buyer, then score each candidate separately for business fit and relationship strength. Research current timing only for the strongest candidates and make the smallest useful ask.
What counts as a warm lead?
A warm lead has a credible relationship path, not merely a matching job title or a social-media connection. Examples include a past client, a trusted former colleague, a repeated collaborator, or someone a mutual contact can genuinely introduce. The person must also fit what you sell.
How can I use LinkedIn connections without scraping?
LinkedIn provides an official account-data export for your own first-degree connections. Import only data you are authorized to use, respect missing fields and privacy choices, and do not treat connection status as consent to send commercial messages.
Should I ask my network for referrals or pitch them directly?
Choose based on fit and relationship. Ask a well-placed peer for a referral when they are not the buyer but know the market. Approach a possible buyer directly when the need and your shared context are both credible. In either case, make it easy to decline.
Can AI find warm leads automatically?
AI can research and prioritize candidates, but a person should review the evidence and decide who belongs in the pipeline. A useful system distinguishes fit from relationship strength, shows uncertainty, preserves dismissals, and requires appropriate approval before external outreach.