What to Bring to a Cortex Consultation for a Customer-Specific Intent Plan

October 27, 2026•4 min read

The most productive Cortex consultation begins with your business problem, not a request for the largest possible list. Bring one priority offer, a definition of a suitable customer, examples of real outcomes, and the decision you want intent data to improve.

Those inputs make it possible to discuss a source plan, feedback process, and evaluation that reflect your niche. They also help distinguish a useful customer-specific model from a generic audience purchase.

Bring the offer you want to grow

Describe what you sell, who can use it, where you can serve them, and what makes the purchase worthwhile for your business. Include meaningful boundaries such as minimum project scope or supported applications.

Avoid presenting every product as equally important at the start. A focused first use case is easier to evaluate and gives the learning process a clearer target.

If your strategy recently changed, explain that. Historical customers may remain valuable examples, but they may not all represent the audience you want next.

Bring contrasting customer examples

Prepare anonymized examples of a strong customer, a poor-fit inquiry, a qualified opportunity lost for timing, and a sale lost because of an operational limitation. Explain why each belongs in its category.

These contrasts reveal more than a list of favorite customers alone. They help separate market fit from sales execution and capacity.

Include suitable customers who came through referrals or other channels if they can be used under the approved workflow. Preserve their origin so learning does not become mistaken campaign attribution.

Bring a proposed source map

List the websites, topics, and buying signals you believe are relevant, along with why they matter. Include adjacent interests you want to distinguish from your actual offer.

Treat this as a discussion document. Confirm which sources and levels of detail are available rather than assuming the service can observe every requested behavior.

Ask how feedback would change the interpretation over time. The goal is not merely to select sources once, but to create a process that can learn which signals fit your customers.

Bring your sales workflow

Explain who receives opportunities, how quickly they can review them, what the first action is, and where outcomes are recorded. Include realistic capacity constraints.

A customer-specific model needs reliable feedback. If the team cannot consistently identify qualification or loss reasons, the consultation should address that process before promising sophisticated learning.

Confirm supported integrations and handoffs for the actual systems involved. A manual pilot can be appropriate, but its effort should be understood.

Bring the measurement question

Choose what you want to learn. Examples include whether a source produces more serviceable inquiries, whether a combined signal improves a review queue, or whether feedback improves ranking at a fixed sales capacity.

Define the outcome, comparison, and buying-cycle horizon. Distinguish qualification from purchase, and attribution from incremental value. These choices determine which data is needed.

If you want an original benchmark, bring anonymized event timestamps, score snapshots, model versions, qualification rules, outcome dates, and reporting permissions as relevant. A benchmark requires inspectable evidence, not a desired headline.

Ask what the first 24 hours can establish

The supplied Cortex brief says scoring can begin about 24 hours after list subscription. Ask what the initial score represents, which inputs are available, and what setup checks should occur before scaling activity.

Do not confuse the start of scoring with proof of improved accuracy or completed revenue results. Those require reliable feedback and an appropriate observation period.

Set expectations for the first week around process quality and for later reviews around mature business outcomes.

Clarify the proprietary part

Ask how the model becomes specific to your niche, offer, geography, and sales cycle. Clarify access, retention, export, and permitted reuse of customer records and learning.

Customer-specific interpretation does not imply that every underlying contact is exclusive. Identity matching does not create consent, and a technical export does not establish eligibility for every advertising platform.

If clean-room analysis is relevant, define the permitted question, participants, data, and outputs. The controls depend on the actual implementation and agreements.

Leave with a concrete first decision

A useful consultation should identify a focused use case, the inputs required, the owner of feedback, and a realistic evaluation approach. It should also make limitations visible so your team knows what remains to be verified.

Book your Cortex Marketing AI consultation. Bring the business context that only you can provide. The purpose of two-way intent data is to turn that context and your real outcomes into a more useful understanding of the buyers you want to reach.

Back to Blog