Why real-time data alone does not improve executive decisions

Decision speed depends on decision conditions

Confluent’s Quick Thinking 2.0 research reports that 92% of leaders say senior business decisions have become faster over the past three years. It also reports that 75% have regretted acting too quickly, while 71% have regretted waiting too long and missing an opportunity.

Confluent sells data-streaming infrastructure, so it has a commercial interest in arguments for real-time data. Its findings should be read with that incentive in mind.

The figures point to a practical management problem. Different decisions require different speeds. A cyber incident can demand rapid action, while a high-stakes investment can justify more deliberation. The useful question for executives is what consumes their limited attention while they decide.

Some of that attention can go to an earlier question: whether the evidence in front of them is usable.

Before the business decision comes a decision about the evidence

Consider a senior meeting where two teams arrive with different figures for the same business question. Before executives can discuss pricing, risk, investment or operations, they must establish which figure is current, complete and relevant. Someone asks for another report or contacts another team. The decision waits, or proceeds with the discrepancy unresolved.

Quick Thinking 2.0 reports several related findings.

These figures show that surveyed leaders report these constraints at the same time. They do not establish that difficulty accessing data or stale information causes greater reliance on instinct.

For an executive, a discrepancy between two figures creates another task: establishing which number applies before using it to weigh revenue, customer impact, regulatory exposure or risk. Resolving the discrepancy consumes time that could be spent examining the business choice itself.

The meeting points to a design question for information systems: how many questions about the evidence remain unresolved when it reaches the decision maker?

Freshness and trust are separate properties

Making information arrive faster addresses part of that question. Accuracy, relevance and appropriate governance require their own controls.

A dashboard may update every second while two teams still disagree about what a metric represents. Update frequency has improved, but the disagreement in the meeting remains.

Quick Thinking 2.0 reports that 91% of leaders would feel more confident in their decisions if they had access to real-time data. This measures reported confidence. Decision quality requires separate evidence.

For executives assessing real-time systems, freshness and decision readiness are separate properties. Freshness concerns how current the information is. Decision readiness also requires enough context to understand what the information represents and how it should be used.

That distinction suggests a clearer objective. Relevant information should reach the point of decision with its meaning and governance clear enough for the executive to evaluate it.

Move data uncertainty upstream

Data quality and governance controls can sit closer to the point where data is created. This operating design addresses questions about quality, meaning and permitted use before they reach an executive meeting or another downstream application.

The test is concrete. If two teams use the same metric, they should be able to establish what it measures, when it was updated and which rules govern its use before presenting it for a decision.

Governance therefore has a direct role in the decision process. It defines how information is managed and used. In a system designed to deliver data quickly, those rules also help determine whether an executive can interpret the information when it arrives.

Real-time data streaming, meaning the continuous movement of data between systems as events occur, can move current information across teams, systems and applications. Confluent has a commercial stake in this approach because it sells data-streaming infrastructure. The quality and meaning of the information being moved depend on separate controls.

The investment test is stricter than update frequency. If a system delivers fresher data while executives still have to determine which figure is valid, the meeting still begins with the same unresolved evidence problem.

AI makes input quality more consequential

Quick Thinking 2.0 reports that 62% of executives use AI to make the majority of their decisions. It also reports that 70% second-guess their own judgement when it conflicts with AI recommendations.

As survey findings, those numbers show why executives need to examine the information behind an AI recommendation. A recommendation about fraud risk, pricing, maintenance or a supply-chain response can depend on the data supplied to the system.

Executives therefore need to examine data provenance, meaning where the data came from and how it reached the system, as well as its freshness, relevance and governance. The Confluent findings show that, among the executives surveyed, AI already has a reported role in many decisions and can conflict with executives’ own judgement. The correctness of an AI recommendation requires separate evidence.

That makes traceability a practical requirement. When an AI recommendation reaches the decision room, leaders need enough information about its inputs to judge whether the recommendation deserves weight in that specific decision.

Key highlights

• Match decision speed to the decision: Faster is not always better. Executives should calibrate decision speed to urgency, risk and the time needed to establish whether the evidence is usable.
• Resolve evidence questions before the meeting: Conflicting, inaccessible or outdated data consumes executive attention before the business choice can be assessed. Move questions about validity, relevance and meaning upstream where possible.
• Treat freshness and trust separately: Real-time updates do not guarantee accurate, relevant or well-governed information. Assess whether data arrives with enough context and governance to support the decision.
• Move data uncertainty upstream: Apply quality and governance controls closer to where data is created. Shared metrics should have clear definitions, update information and rules for use before they reach decision makers.
• Make AI inputs traceable: As AI recommendations influence more decisions, executives need visibility into data provenance, freshness, relevance and governance. Traceability helps leaders determine how much weight an AI recommendation deserves.