Decision Console

Decision Console is the screen for reviewing the current state of a selected deal through Bayesian judgment and ontology evidence. It does not merely show the P(Win) number. It explains why the judgment appeared, what is blocking progress, what evidence must be confirmed next, and how an adopted recommendation connects to actual sales activity.

Decision Console includes AI analysis. AI analysis is not a separate product or separate screen. It is a Decision Console function that strengthens judgment explanation based on server-prepared evidence and the current deal context.

Location: Sidebar → Ontology → Decision Console

Why This Screen Is Needed

Sales leaders do not only ask, "What is the probability?" The more important questions are: "Why does the system judge this way?", "What is blocking the deal?", and "What must we verify before the next meeting?"

Typical pipeline probability often depends on stage convention or the salesperson's optimism. EXAWin+ reads activity records, selected signals, Bayesian updates, silence, momentum, threshold, and decision impedance together.

Decision Console gathers these signals into one operating screen so the team can discuss a deal by evidence rather than intuition.

Basic Flow

Use the project list on the left to select an accessible deal. Only deals within the user's company and permission scope can be selected. When a deal is selected, the judgment area on the right refreshes for that deal.

For the selected deal, the system prepares recent Bayesian updates, silence information, accumulated silence penalty, momentum P(Win), and ontology judgment context. Based on that evidence, Decision Console shows the current state, bottlenecks, recommendations, similar context, and follow-up tracking.

Users first review P(Win) and judgment status, then read bottlenecks and next evidence candidates. If needed, they request AI analysis to read a natural-language explanation of current deal evidence. If a recommendation is valid, the user adopts it and records the actual follow-up activity in Activity War Room.

Core Judgment On The Screen

Decision Console shows the current Bayesian state of the selected deal: P(Win), distance from threshold, confidence, latest update, silence, and momentum.

P(Win) is the posterior probability calculated from recorded activities and selected signals. Threshold is the decision reference for considering movement to the next stage. Decision impedance interprets how blocked or passable the judgment is through the relationship among current probability, threshold, and k.

Silence and momentum must be read together. In long sales cycles, no news is not always neutral. If recent activity stops or customer response is delayed, judgment should reflect that time flow.

Tabs And Views

Progress constraints show why the deal is not moving easily. The important question is not only the probability itself, but which evidence is limiting the judgment.

Bottleneck shows candidate bottlenecks. A bottleneck does not simply mean "bad." It is a point that must be checked before the next action: contact, decision maker, stage, activity gap, negative signal, or missing evidence.

Next action shows candidates suggested by the system based on current evidence. This is not a command that users must execute. It is a judgment candidate the sales team reviews with the real context.

Similar deals provide comparison context. Seeing what happened to deals with a similar flow can support the current judgment.

Sales action tracking shows how follow-up activities connect after recommendation adoption. If a recommendation is adopted, related actions, preparation, contact attempts, and customer interactions can be recorded in Activity War Room and connected to the recommendation.

How AI Analysis Is Used

AI analysis in Decision Console reads the selected deal's evidence and provides natural-language interpretation. When the user requests AI analysis, the server prepares the deal context and evidence pack. This pack may include deal state, Bayesian judgment, activity context, and evidence lists.

AI output must pass server validation. A generated sentence does not automatically become system judgment. The server checks format and evidence connectivity, and only validated results can be treated as an interpretation draft.

Even if an AI request is temporarily incomplete, the server-confirmed Bayesian judgment and evidence remain intact. Decision Console's core judgment does not depend only on generated text. AI analysis is an explanation layer that helps people read the judgment inside permission, evidence, and validation boundaries.

Meaning Of Recommendation Adoption

Adopting a recommendation means: "We will treat this recommendation as a tracked sales action." Before adoption, follow-up activity is not connected as an evaluation target for that recommendation.

After adoption, users can record follow-up sales activity in Activity War Room. For example, additional questions sent to the customer, internal preparation, or actual customer contact can be connected to the adopted recommendation.

If the recommendation is not adopted, it is not treated as an executed follow-up target. This distinction lets the system learn which recommendations were actually acted on and whether judgment improved afterward.

Relationship With Activity War Room

Decision Console reviews judgment. Activity War Room records actual activity.

Use Activity War Room to record new meetings or customer contact results. After adopting a recommendation in Decision Console, move to Activity War Room to record the actual work.

This separation is intentional. The judgment screen does not arbitrarily change data. Real work outcomes are recorded in the standard activity screen, and those records flow back into Bayesian updates and ontology analysis.

What Not To Overtrust

Decision Console does not guarantee the future of a deal. P(Win) is a judgment based on recorded evidence and configured model settings. Customer internal issues, competitor movement, budget changes, and decision delays that are not recorded cannot be known by the system.

The screen is not a tool for declaring that a deal will close or fail. It is a tool for judging which direction current evidence supports, what must be confirmed, and which follow-up action should be prioritized.

AI analysis is also not the final decision maker. It helps explain the evidence, but customer messages, pricing conditions, contract strategy, and actual sales decisions remain the responsibility of the user and company policy.

Good Habits

Before opening Decision Console, check whether recent activities in Activity War Room are sufficiently recorded. If the data is stale or signals were selected inaccurately, judgment quality will fall.

When reading a recommendation, do not ask only whether the recommendation is right. Ask what customer response or internal evidence is needed to verify it. The value of Decision Console is to make the next validation question clear.

When using AI analysis, compare the explanation with actual screen evidence. If needed, inspect relationships in Ontology Inspector or use Ontology AI to explain the same evidence in another way.

Read Next

To inspect the evidence relationship structure, read Ontology Inspector.

To use natural-language evidence analysis with Decision Console, read Ontology AI.

To record activities and follow-up actions, read Activity War Room.