Use Case: Aggregate Viewing / Reading / Listening or Buying Data to Enable Genuinely Personalised Recommendations
Comprehensive recommendations based on a complete data-set and detailed, curated preferences.
Overview
People will regularly see recommendations popping up, for example on a streaming service, or a retailer, for something they have already watched, read, listened to or bought.
That happens because the services in question don’t have a complete view of the person's history. And therefore any recommendation made is only based on what they can see has been done on their platform; not what the same person has watched/ read/ listened to or bought elsewhere.
DataPal recommendation engine enables people to bring together their activities on multiple platforms; rate them and then uses AI to augment that history. That rich, deep augmented history can then be made available to service providers who can use it to better understand their user base, and to make highly relevant recommendations.
Context
Recommendations engines have become business as usual in online commerce and entertainment over the last 25 years. They just sit in the background; none of us is surprised or amazed by them as we may have been in their early days. In turn, none of us get too surprised when, amongst the flow, recommendations show up for things that we have already watched, read, listened to or bought.
Challenge
By now it is pretty obvious to most users of web platforms that recommendations based on activity or ratings on a platform can only relate to a subset of their overall watching/ reading / listening/ buying behaviour. So recommendations in the current paradigm are constrained in their inputs. But they are also constrained on the output side. Clearly platforms will only recommend products and services that they themselves sell, and not ‘whole of market’ which would be preferable from the individual/ buyer side.
Solution
DataPal Recommendations Engine addresses these challenges by providing people with the tools to gather their data from multiple sources. Then curate the combined data-set, and use that to identify recommendation characteristics. And finally apply those characteristics to a market level data-set of possible product/ service options. And all done without sharing significant volumes of data with other parties. Recommendation options are pulled to the individual and their data rather than the existing activity and related personal data being shared with the market.
How It Works (Flow)
Data Ingestion: Individuals import or connect up their various feeds of watching, reading, listening or buying data with their DataPal account.
Data Cleaning & Parsing: DataPal automatically standardises formats, matches related activities, and fills in missing context.
AI-Based Augmentation: External data sources and machine learning enrich transaction records with product, merchant, and contextual metadata.
Human-in-the-Loop Review: Users can review and enhance records — adding personal notes and ratings.
Granular Permissioning: The user may choose to share all, some, or even just one activity record with third parties — time-bound, purpose-specific, and revocable at any time.
Ongoing Learning: Once an activity has been augmented, DataPal auto-recognises and enriches future similar records.
Actors
Individual / Data Owner: Ingests, enriches, and manages their transaction data and permissions.
Activity provider: The sources of original watching, reading, listening or buying record for ingest.
Organisation / Service Provider: Requests specific, purpose-limited access to verified transaction data (e.g., for lending, loyalty, or service fulfilment).
DataPal Platform: Provides ingestion, cleaning, AI augmentation, and permission management under the IEEE 7012 MyTerms model.
AI Agent / Developer: Builds value-added services — notifications, reminders, or price-comparison assistants — using DataPal’s APIs.
Benefits
We can outline the benfits to each stakeholder both in the immediate short-term and the longer-term:
Individuals Gain clear, usable activity records; control exactly what’s shared; unlock new value from their own data. Longer term individuals build an automated, personalised data vault that continuously learns and enriches itself over time.
Activity record providers have the ability to improve the experience of their users. Could also establish themselves as users of the aggregate data-sets of activities.
Organisations / Service Providers Access accurate, consented, and purpose-specific activity data on which to run recommendations and better service personalisation.
Developers / Ecosystem Access a composable, consented activity data platform for building innovative service and agents. Longer term grow new business models based on privacy-first, customer-controlled data collaboration.
Society / Regulators Demonstrates practical privacy innovation and ethical data sharing. Longer term this encourages transparent, compliant data ecosystems aligned with public trust.
Outcomes
Significant improvements in completeness and accuracy of genuinely personalised recommendations across a wide range of categories.
Higher conversion rates from inbound recommendations through completeness and accuracy of the data driving them
Increased user trust and engagement, measured by active consent renewals and sharing frequency.