Case 04 · xPRM · custom development
xPRM — in-depth contact analytics and business relationship management
Search across contacts and channels, automatic tags, enrichment of contact details and audience reports.
xPRM (https://xprm.io/) is a web platform for in-depth analytics of Telegram and Instagram channels and their subscribers, built for advertisers, agencies and channel owners.
It brings together data on channels, ad placements and results to make decisions on buying and selling advertising easier.
The system provides a single workspace where a team can search for and evaluate channels, plan campaigns, track subscriber dynamics and key performance indicators. Through integrations with bots and analytics services, xPRM automatically collects and updates statistics, flags the quality of channels and builds analytical reports.
Going forward, xPRM can scale beyond Telegram and Instagram to other social platforms while remaining a single hub for managing advertising assets.
How AI is used
- Search across the contact directory with free-form queries (hybrid search combining vector search with LLM-based pre- and post-processing of queries and results)
- Tagging the contact directory (professions, industry, role, etc.) with an LLM
- Summarising all the information in a contact card with an LLM
- Enriching directory information with agentic search, including web search
- Classifying bots among subscribers
- Understanding user queries
- The system uses AI to better understand the meaning of a query: it extracts names, phone numbers, roles and professions and synonyms, and uses them to find people more precisely.
- Semantic search
- In one mode, AI builds a vector representation of the text and uses it to find people who match by meaning, not just by exact word matches.
- Normalising Telegram data
- AI helps bring Telegram profile data (first name, last name, bio, etc.) into a clean, structured form.
- Automatic tagging and summaries of users
- For people's cards and profiles, AI assigns tags (professional, free-form, company, etc.), builds summaries and merges sources.
- Filtering interests from external sources
- When raw interests arrive, AI applies rules and filtering to remove unwanted categories and topics.
- Generating analytical text about channels
- AI produces a human-readable description of the audience and a final text analysis based on channel data.
- Working with an external interests API
- The system calls an external service to get interests, and parts of the pipeline then process and filter this data with AI logic.
- Tracking and controlling AI usage
- For every AI call, tokens, cost, time and errors are counted and a report is built to show where and how much AI capacity is spent.
Project interface
Fortune AI
Have a similar task?
In a free consultation, we'll discuss your process and possible uses of AI.