Good research starts on both sides of the negotiation

Before a customer meeting, it is easy to collect plenty of material: a company website, an annual report, job advertisements and a few news items. That does not yet produce a reliable picture of the negotiation. What matters is which information changes your next question and which assumptions you still need to test. Your own side deserves equal attention: what can you actually deliver, promise and justify commercially?

Clarify your own starting point

Begin with a short assessment. What problem does your offer solve? Which capabilities are available today? What conditions apply to implementation, resources and price? Record your objective, internal responsibilities and your alternative if no agreement is reached. The Harvard Program on Negotiation recommends considering your alternatives alongside both parties’ interests and limits.

This preparation prevents a familiar disconnect: the opening pitch promises a benefit that your organisation can deliver only under additional conditions. Keep those conditions visible. An appealing sentence then becomes a proposition that can be checked. Involve the people who will later be responsible for delivery and commitments. Their constraints are part of the offer, even when they do not appear in its headline.

Turn information into testable statements

Now consider the target customer. Public company sources can show which markets, products or investments the business describes. You may infer that your offer could be relevant. A specific budget, purchasing intention or priority of your contact usually requires further confirmation.

Consider a fictional example. A machinery manufacturer announces an additional site. The announcement supports the claim that an expansion is planned. The idea that this will create training needs is an interpretation. Whether that need exists, and who owns it, remains a question. A useful opening might be: “You have announced the expansion. What requirements does that create for onboarding new teams?” This leaves room for the customer to correct your interpretation.

Start with four fields

Keep your research in a compact structure:

• Statement: What does the source actually say?
• Evidence: Where, when and in what context was it published?
• Interpretation: What could it mean for the conversation?
• Open question: What would confirm or challenge that interpretation?

Open the source yourself and inspect the relevant passage. AI-generated citations particularly need this check; NIST explicitly recommends verifying sources and citations in its generative AI risk profile. Three copies of the same announcement are not three independent perspectives. If sources conflict, retain both versions and their dates in your preparation.

Translate research into better conversations

AI can help organise material and draft questions. Ask it to look for contrary evidence and missing information as well. A useful instruction is: “Separate supported statements, plausible interpretations and open questions. Do not infer a person’s motivation from a company profile.” Then check the output against the original sources.

Stop preparing when your next questions are clear and additional research is unlikely to change an important decision. You need a reliable starting point for an open conversation, without knowing every answer in advance. After the meeting, replace assumptions with confirmed statements or discard them. Research becomes a learning process that can also help with smaller, regularly recurring deals.