AI-generated cover and inline illustrations.
A business-to-business pitch, or B2B pitch, proposes a product or service to another business. Before asking for a meeting, the sender needs to show what a conversation could help that business decide.
Preparation starts with finding a real fit. The sender develops one small idea and shows enough evidence for the recipient to assess it.
AI can help organize this preparation. It cannot establish a business need from a few public pages, supply trust, or promise that a meeting will turn into a client.
Start with a useful task
Choosing a prospect because it is a large company or appears to have a healthy budget says little about what it needs.
Begin with a real question or task. Which business has a need that the sender is equipped to address? What would make the proposed work useful to both sides?
For a service business, that might mean making a specific customer step easier to understand. For a supplier, it could mean a small trial that helps a buyer assess whether a product fits. The potential partner is relevant because of the work, not merely their size or status.
Public professional material can help establish that fit: a company’s service pages, a published interview, a research summary, or a speaker’s recent talk. Focus on the work and the business’s stated priorities. Private details and personal vulnerabilities are not needed.
Separate an observation from a diagnosis
A public page can show what a business has chosen to explain online. It cannot, by itself, explain the business’s results.
An agency may observe that a salon’s English page does not explain how a first consultation works. It cannot conclude from that page alone that the salon is losing foreign customers. The salon might answer those questions well through another channel. It might not want more bookings. Its main constraint might be staff time rather than demand.
A useful research note keeps three things apart:
- What was directly observed, with the page and date checked.
- What that observation might mean, clearly treated as a possibility.
- What the business would need to confirm before agreeing to the work.
This simple separation changes the tone of a pitch. “Your booking process is costing you sales” becomes “I could not find a first-visit explanation on the English page. Is that something customers ask about?”
The second version identifies an observation and leaves room for the salon to explain its situation.
Make the proposal smaller than the ambition
Consider a hypothetical agency preparing a proposal a salon has asked to receive. The example is fictional and does not describe PIN client work.
The agency wants to help the salon serve English-speaking customers. That is too broad for a first request. It could involve language support, service design, content, booking tools, and staff capacity. A promise to solve everything would hide the decisions that still need to be made.
A narrower idea is easier to assess: one first-visit explainer that answers the salon’s most common consultation questions and points to its existing inquiry route.
Before making that proposal firm, the agency would need to confirm the salon’s actual questions, available language support, approval process, and capacity. It should not advertise support that staff cannot provide. It should not rebuild a booking flow before learning why the current one exists. Our guide to making an offer usable in Korea explores those delivery details.
The first message can show a short outline and one relevant work sample. It does not need to include an unpaid full strategy. If there is a fit, the next step can be a small paid trial project, often called a pilot, with an agreed scope.
Before a paid trial begins, both sides should agree on its scope. They should know what will be delivered, who supplies the facts, who approves the work, when it will be ready, and how they will judge it. The agreement should also say what is outside the project. Agree on the fee, included revisions, and how additional requests or approval delays will affect cost and timing.
Evidence should match the promise
A beautiful design sample shows design ability. It does not, by itself, show that the design produced sales.
A testimonial may show that one client was satisfied. It does not prove a result will repeat elsewhere. A strong portfolio can establish craft, but the sender still needs to explain which part of that work relates to the proposed task.
The best proof is relevant and modest. A sender can explain what they made, what their role was, and what can actually be verified. Client work should be shared only when permission allows. A self-initiated sample can be useful if it is clearly identified as a sample rather than a commission.
Someone without a large portfolio can still make a credible request. One thoughtful piece of work is more useful than a long list of implied connections. The missing experience should shape the size of the proposal, not be covered with bigger promises.
Check permission before sending a sales proposal
A useful idea is not permission to send advertising. A public business email address or a personal introduction does not, by itself, establish consent to receive a sales pitch.
In Korea, Article 50 of the Information and Communications Network Act generally requires express prior consent for commercial advertising sent electronically, with limited statutory exceptions. Before sending, check the applicable rules for the recipient and channel, including sender identification, advertising labels and refusal methods. A polite opt-out or a request for permission is not a way around the prior-consent rule if the message itself advertises a service.
KISA’s detailed anti-spam guide explains that information specifically requested to assess a contract, such as an estimate, can fall outside advertising. The salon example below assumes the recipient has asked for that specific proposal.
Stay within that request. It does not cover a mailing list or unrelated offers. The sample shows proposal wording rather than a complete email. Before using it, check which sender details, advertising labels and refusal methods apply. If the basis for sending is unclear, check the guide or get qualified advice.
A genuine editorial interview request is not automatically a sales advertisement. Assess the actual purpose and content rather than treating every creator invitation as advertising or disguising a sales pitch as an interview.
Make the requested proposal easy to assess
A requested proposal can be short without being vague. It needs the relevant observation, a modest idea, one credible reason to consider the sender, and a clear next step.
For the hypothetical salon, a requested proposal could say: “Thank you for asking for an outline for an English first-visit guide. I suggest three short sections: what customers should share before a visit, what happens during the consultation, and how the service, upkeep, price and timing are confirmed. The guide would use facts your team approves and point to your existing inquiry route. Which details are missing or need changing before we discuss the work and fee?”
If the fictional agency has a relevant sample it can share, it could add: “I have linked a self-initiated first-visit guide to show how I organize the information. It is a sample, rather than client work or evidence of sales.” The sender would replace this with an accurate description of their own work.
The sections would need to match the salon’s actual services, customer questions and available language support. Any linked work sample should be genuine and shared with permission.
An interested reply can lead to a conversation about the business goal, the constraints, and a paid pilot. Our guide to useful consultation questions explains how that conversation can change the recommendation. An unambiguous no closes the request. Silence is not permission to keep pushing through more personal channels.
An introduction is not blanket permission
If someone offers an introduction, ask them to check whether the recipient wants the specific proposal. Give the contact a short note explaining who is asking, why the work fits and what the next step involves. They should not have to guarantee the work or share private contact details.
Use AI to organize the preparation
AI is most useful here as a research and editing assistant. It can organize supplied material, compare a proposal with a public business priority, identify unanswered questions, and shorten a crowded message.
The human sender still needs to open the sources, check their meaning, and decide whether the request is appropriate. A generated summary is not a verified fact. A plausible connection is not a real relationship.
A practical instruction is:
“Using only the material below, separate verified observations from assumptions. Include the source for each observation. Identify what still needs confirmation. Suggest one small proposal that fits our actual capabilities. Do not invent business problems, personal interests, past results, relationships, or likely revenue. If the recipient has asked for or consented to this proposal, draft a short message within that scope that is easy to decline. Do not infer consent from a public email address, an introduction, or silence.”
Only material suitable for that AI service should be included. Private customer records, confidential proposals, and personal contact lists are not needed to test this approach.
Measure progress without confusing it with success
The first useful outcome may be an answer that corrects the original idea. A business might say its English information is already sufficient, but that customers struggle to choose a service. That is a reason to revise the proposal, not a failed attempt to force the original sale.
Replies, qualified conversations, paid pilots, and repeat work are different outcomes. Track them separately. A meeting is not revenue, and a paid pilot is not proof of a profitable long-term service.
For the hypothetical salon, staff could first check whether the proposed explainer answers the right questions. After release, they could review the inquiries it supports, using only information they are entitled to access. Any comparison should account for other changes, such as promotions or seasonal demand. A rise in inquiries would not automatically prove that one piece of content caused it.
The method worth keeping
Before the next proposal, write down one observation and its source, one point the business needs to confirm, and one small task your team can deliver. Check that the recipient has requested or consented to the proposal.
Then choose a work sample that supports that specific task. If the evidence changes the idea, revise the proposal before asking for a meeting.
