More precise context hints
Describe needs, situations, and contexts based on how demand actually appears and develops, rather than relying only on taxonomies built from assumptions.
DECIVUM TRACE · CONVERSATIONAL DECISION INTELLIGENCE
TRACE reveals how the demand you want to capture with ChatGPT Ads actually takes shape.
It reconstructs how customers and prospects actually use ChatGPT and other generative AI platforms to express a need, explore it, compare alternatives, and make a choice. It connects those behaviors to observable journey signals and turns them into actionable guidance for context hints, messaging, landing pages, and tests.
What it enables
From real demand to concrete campaign decisions.
TRACE strengthens the decisions media teams and agencies already need to make: which conversational contexts to prioritize, what message to use, what experience to provide after the click, and which hypotheses to test.
More relevant messaging
Use decision criteria, doubts, alternatives, and reasons to believe that emerge along the journey to build messages that better match what the person is trying to decide.
Better aligned landing pages
Align the destination experience with the questions, proof points, and alternatives that brought the person to that stage of the journey.
More informative tests
Set explicit hypotheses around context, messaging, and expected outcomes so you can understand what worked, what did not, and what to test next.
TRACE does not replace media planning or media buying. It helps decide which decision contexts are worth prioritizing and what to test.
The blind spot
A campaign can be well executed and still start from a weak assumption about demand.
Intent is not static in conversational environments. A person may start with a problem, explore it, discover alternatives they had not considered, and change their decision criteria as the conversation unfolds. A list of keywords or hypothetical prompts captures only part of that process.
The question evolves
What the person wants to understand at the beginning may be very different from what ultimately determines the choice.
The alternatives change
Brands, products, and solutions can enter or leave consideration as the conversation progresses, including alternatives the customer had not previously considered.
The criteria become more specific
Price, reliability, service, features, risk, and trust can take on different weights as the need becomes clearer.
Outcomes leave signals
Visits, leads, conversions, sales, and other available signals help show which journeys are associated with concrete outcomes and which hypotheses deserve further testing.
The TRACE framework
Five dimensions to reconstruct the decision context.
TRACE follows the journey from the situation that triggers the search to the observable evidence that connects the decision to concrete actions and outcomes.
Trigger
The event, need, or tension that sets the search in motion.
Request
What the person is trying to understand and how the request takes shape and evolves.
Alternatives
The brands, products, services, or solutions that enter consideration.
Criteria
The criteria the person uses to compare, rule out options, and reach a choice.
Evidence
The observable signals that help show what happens next in the journey and with what outcomes: visits, leads, conversions, sales, and other measurable results.
Output
TRACE turns Decision Intelligence into a brief that campaign planners and measurement teams can use.
The output is not another dashboard. It is a structured foundation for defining stronger campaign hypotheses and knowing what to observe once the test goes live.
Conversational Demand Map
Triggers, requests, alternatives, criteria, and decision patterns that describe how demand takes shape and develops.
Priority Decision Contexts
The contexts in which a product or service may be relevant, prioritized according to the available evidence and their usefulness for campaign planning.
Planning Intelligence Brief
Guidance for context hints, messaging, landing pages, test hypotheses, KPIs, and signals to monitor so campaign results produce insights that can be applied to future tests.
TRACE does not replace media teams, agencies, buying platforms, or reporting tools.
It works upstream: helping teams understand which decision contexts to prioritize, which hypotheses to test, and which signals to use when interpreting results.
The same DECIVUM engine
TRACE changes the use case, not the DECIVUM engine.
TRACE uses the same data sources and methodological discipline as DECIVUM: behavioral research, client data, analysis of generative AI environments, clearly defined evidence levels, and, when the scope supports it, a connection to economic value.
How DECIVUM worksSearch in ItalyHow people search, evaluate, and choose, studied since 2004.
Behavioral researchSurveys and, when scope and sample size allow, panels that directly observe selected interaction sequences.
Client dataAnalytics, CRM, leads, conversions, sales, and other proprietary signals that help connect the journey to observable outcomes.
Generative AI environmentsAnalysis of questions, answers, alternatives, sources, and criteria that emerge across decision journeys.
Evidence levelsA clear separation between observed, reported, generated by models, inferred, validated, and estimated evidence so different kinds of information are not treated as equivalent.
When prioritization matters
Not every decision context deserves the same budget.
When the available data supports it, TRACE connects behavioral patterns and observable outcomes to the economics of the defined scope. This helps separate what is interesting from what is material enough to justify a test, an investment, or a higher priority.
Initial conversation
Are you considering investing in ChatGPT Ads?
The first step is to understand what demand you want to capture and how much you actually know about the decision journey behind it. We can start with your market, the data already available, and the decisions your media team needs to make.
If TRACE can improve the quality of planning, we define a scope. If the available evidence does not support a sufficiently robust analysis, it is better to know before building the test.
Marco Loguercio will reply personally.