AnyIdea AI and AI Prio Score

AI technology & foresight

AnyIdea's AI layer combines Synthesis and generative AI. The module Foresight is primarily based on our specially developed synthesis AI. This is the result of many years of research with academic partners and FFG Austria and is being continuously optimized. It processes enormous amounts of data (input) to provide well-founded insights and decision-making bases (output).

Trend Agents: Context & Focus

Trend agents are leading the way for synthesis AI. Through your inputs, the AI analyses, links and weights relevant data with pinpoint accuracy. In this way, important signals are not only identified, but they are also transferred to prioritized trend clusters based on identified patterns.

Details of data and data categories considered can be found here.

Data-based trend identification

Identification is consistent “Bottom-up”: The AI analyses the entirety of all signals and evaluates them with a KI-prio score. This methodical evaluation makes the relevance and potential impact of a trend completely comprehensible and always keeps it up to date.

Learn more: What trends does AnyIdea work with?

The AI Prio Score

This score provides an objective assessment of signal relevance. It is based on the distribution, frequency and development of topics over time. This gives you precise information about the probability of occurrence and the Time to Impact of technologies and trends.

The bottom-up calculation (methodology)

In addition to synonyms, our models also capture complex semantic similarities. They are supplemented by proprietary algorithms that weight data categories and results:

• Extraction: Documents and articles are analysed and core information extracted.
• vectorization: Using deep learning, this information is transferred to a multidimensional vector space and clustered there.
• Weighting: The thematic relevance is determined using cosine similarity, timeliness and signal strength.
• categorization: Data sources are specifically weighted. For example, scientific publications are classified as more future-oriented, while news data is incorporated differently into the time calculation due to their high volatility.

‍