A signal's journey: from publication to decision

What is a signal, and what is not one?

Not all information is a signal. A press release repeated in fourteen outlets is a single signal; a competitor’s job advert can be one if it reveals a new line of business. A signal is defined by its relevance to the organisation’s information needs, not by its volume or reach.

The most valuable signals are often also the hardest to see. Ansoff called them weak signals: early, incomplete indications of a change which, if detected in time, leave room to prepare before the information is complete^[1].

The seven stages of the journey

1. Capture. The declared sources are reviewed continuously: news and websites, patents, regulation and official gazettes, tenders, competitors’ websites and specialist sources.
2. Filtering. Each watch’s criteria let through what fits the defined needs.
3. Classification. Copies of the same publication are grouped, and each one is tagged by what it is about: technology, entity, type of event, place.
4. Review. A person with judgement validates what is relevant and discards the noise.
5. Retention. What is validated is saved with its tags, its comments and the link to its source, and becomes part of the history.
6. Distribution. It reaches each recipient with the detail they need, for example through newsletters.
7. Decision. The team interprets, debates and acts. It is the only stage that no tool can carry out for you.

The sequence matches the one set out in UNE 166006:2018 for the operation of the system: source control, search, processing, value-adding, distribution and storage^[2]. ISO 56006:2021 also points out that strategic intelligence is cross-cutting and not confined to one department: within an organisation, many journeys coexist, one for each area, and sharing the same base is what makes it possible to cross-reference them^[3].

Where each stage fails, and how it shows

The capture failure is the most serious, because it leaves no trace: nothing that comes afterwards will correct it, and nobody misses what they never saw.

Example: the journey of a regulatory signal

Illustrative example. Every day the BOE publishes a summary, also available as open data (XML and JSON) for automated capture^[4]. One day a draft royal decree appears that affects an additive your company uses in one of its products.

1. The provision is captured because the official gazette is one of the sources of the regulatory watch.
2. It comes in because it contains the name of the additive and a term from the scope of application, both of which are in the filter.
3. It is classified as a regulatory change, and the reviews published afterwards by several sector outlets are grouped with it instead of arriving as separate entries^[5].
4. The analyst confirms that it affects a product line, and adds a comment: ‘review adaptation deadlines with Quality’.
5. The signal is kept with its source. Months later, when the final version is published, the history makes it possible to compare both texts.
6. The head of Quality receives it in the newsletter and takes it to the product committee, which decides to bring forward tests with an alternative additive.

How Innguma follows this journey

At Innguma, each stage has its place. Factory, the space where Innguma processes and structures signals, brings together capture, filtering, AI classification and validation; the AI groups duplicates and classifies each signal by category, technology, entity, type of event, place and tone. The Library keeps what has been validated, and newsletters are sent from it, with summaries generated by AI. Each signal links to the publication it comes from, so the journey can be checked from start to finish.

Innguma offers connectivity through the Model Context Protocol (MCP), currently compatible with ChatGPT, Claude, Microsoft Copilot and Gemini. If your organisation enables it, from the assistant you can run the watch cycle: organise and add sources, modify filters, review the inbox, publish and query what has been published. What is automated, and what goes through human review first, is decided by your team. The AI processes. Your team decides. Knowledge accumulates.