AI-augmented content management | VerseOne's AI Strategy
VerseOne was founded to use innovative technologies to make a positive impact in its chosen sectors and, as far as we are concerned, AI is simply another innovative technology that can be harnessed to improve productivity.
VerseOne's AI strategy mirrors that of our customers' - to encourage safe, pragmatic use of AI that augments human decision-making, not replaces it.
Safe
VerseOne AI does not send any data outside of the EU, or use your data for any centralised or third party training models: your data is safe within the VerseOne ecosystem.
AI Large Language Models (LLMs) have been known to misbehave, especially when preserving their existence — some have even attempted to blackmail their operators! VerseOne ensures that your use of AI LLMs is intuitive and safe.
VerseOne DXP AI implementations:
- are only used if our customers have specifically opted in to their use through the purchase of Prodigy AI;
- send only the absolutely required data to the AI machine;
- no data is transferred outside of the UK;
- data is not used for training generalised models;
- models are contained and unable to carry out deceitful or harmful actions (e.g. blackmailing operators);
- use Azure's AI Services (e.g. Azure OpenAI instances, Azure AI Search, Azure Computer Vision) and so are protected by Azure AI Guardrails.
VerseOne also appreciates that LLMs are not traditional deterministic systems, deliberately programmed for predictable output: they are stochastic systems receiving varied inputs that yield necessarily unpredictable outputs.
VerseOne works with customers to minimise the impact of this in solutions, through precise prompt engineering, reasonable disclaimers, and checks through technical measures where available.
Pragmatic
Focussing the controlled use of AI to solve immediate, real-world problems faced by our customers in their day-to-day work, and so delivering a set of essential tools on which to expand.
In understanding our markets, VerseOne is able to blend differently specialised models (generative, computer vision, NPL) and properly engineered background prompts (Plain English ISO references) in order to add value to the customer organisation (over and above the safe deployment and operation of the models).
One of the most pertinent uses is in enabling customers to query their data within VerseOne DXP, including providing clause-level search results through VerseOne AI Search Assistant, and extraction of meaning and summarisation in the PatientVoice and TenantVoice feedback management solutions.
Augmentative
We believe that AI models are superb at generating insights from large and varied datasets; however, we believe that AI should use these to augment human decision-making rather than being allowed to make irrevocable decisions themselves.
As such, VerseOne DXP AI functions all require human validation after performing their actions.
Internally our use of AI is the same - AI does much of the repetitive heavy lifting, so that humans can work faster or make better decisions.
VerseOne has already built a number of optional AI features into its VerseOne DXP, in the form of Prodigy AI, and these adhere to the principles above — as will any future developments, with due regard to the state of the art, in the near or medium term.
VerseOne is integrating Artificial Intelligence into its Digital Experience Platform with a focus on safety, pragmatism, and augmentation of human decision-making. Its SPA principles ensure data sovereignty, the secure and ethical use of AI models, and the delivery of practical value to customers. Current and future AI features, including Prodigy AI, are designed to enhance insights while maintaining strict safeguards and human oversight.
In order that we can best deliver on our AI strategy, all of our intelligence features are self-contained within Prodigy AI. Ensuring that AI is encompassed within a modular but integrated product empowers our customers to roll out Prodigy AI features when they feel comfortable.