The Strategic Communications team had a reputation framework, but no reliable live measure of stakeholder opinion.
The team already had a mature fieldwork programme. It gave leadership an established framework and a lagging benchmark across important markets, stakeholder groups and international peers.
The problem sat with the wider monitoring capability. Its live media measure treated general sentiment around company mentions as evidence about the company's reputation. A hashtag, name reference or reply could bring a post into scope even when the opinion concerned something else. Keyword and Boolean rules also struggled as language, issues and entities changed.
The resulting score tended to be more positive. When the data was audited, much of it was noise. It could not reliably show what stakeholders thought about the company, who held the view or why it was changing.
The team needed to connect its existing reputation framework to wider stakeholder perceptions, live. The result had to answer three questions:
- How were stakeholder expectations shifting?
- How were competitors performing and repositioning?
- How was the company performing against those expectations?
TSC reduced a high-volume media universe to attributable stakeholder opinion.
Genie captures the agreed source universe before downstream analysis. In current operation, this includes more than 250,000 deduplicated social posts a month across the company and five international peers. The multilingual source mix spans six social platforms: X, Facebook, Instagram, TikTok, YouTube and LinkedIn. Online news and wider digital content enter the same data environment.
Core enrichment removes duplicate, spam and off-topic material and resolves the entities involved. A fleet of client-specific agents then reduces the universe in stages. It excludes content that is purely factual or connected to the company only by proximity. The remaining evidence expresses a meaningful opinion about the company. In this deployment, that high-quality signal usually accounts for about 15% of the deduplicated social universe. The proportion varies with the client and scope.
The analytical agents attribute each opinion to its source and label it against the client's reputation ontology. Targeted sentiment returns the direction of the opinion and an explanation of the classification.
Those labels feed a tailored balanced scorecard. Signals with greater measured visibility carry more weight within each reputation dimension. The dimensions then contribute equally to the overall score. This reduces the risk that one high-volume topic dominates the reputation picture. TSC analysts review the drivers, test anomalies and interpret differences across languages, markets and stakeholder groups. The client and TSC team refine the method and reporting together.
Scaled analysis of stakeholder perception and narrative exposure has turned into a live capability.
The team can now see which stakeholders hold a view, what shaped it and how it connects to the reputation framework. Each judgement remains linked to the source evidence and explanation behind it. This gives Strategic Communications a live measure it can interrogate between fieldwork cycles.
The same persistent dataset feeds a range of agentic deliverables. Recurring reports and event-driven questions draw on the same signal history, entity model and analytical framework. Each output adds context that improves the next.
The next expansion moves beyond stakeholder perception into worldviews. TSC is extending the models to identify the beliefs, priorities and assumptions through which stakeholders interpret the company and its external environment. This requires a larger and more varied data universe. It builds on the same persistent context rather than another separate monitoring setup.
The result is a growing reputation intelligence base. It can support more questions and more internal users without returning to a new collection and analysis exercise each time.
