By Kevin McCaffrey, CEO/Founder Tr3Dent
Artificial Intelligence has rapidly moved from experimentation to implementation. Organizations across government, infrastructure, technology and the smart cities sector are increasingly deploying AI-powered tools to improve productivity, accelerate decision-making and unlock new efficiencies.
Yet as AI adoption grows, so does a significant challenge.
How can organisations trust the information AI produces?
Many users have experienced the frustration of feeding prompt after prompt into an AI platform, hoping it will eventually deliver the right answer. Hours can be spent refining instructions, only to receive responses that appear convincing but are incomplete, inaccurate or entirely incorrect.
This challenge is known as AI hallucination, where an AI generates information that sounds plausible but has little or no basis in fact.
While this may be frustrating in everyday use, the implications for organisations can be far more serious. In government, infrastructure, compliance and regulated environments, inaccurate information can create operational, reputational and governance risks.
So, what is working?
The answer increasingly lies in grounding AI.
Moving Beyond Prompt Engineering
Many organisations initially focus on improving prompts to achieve better results from AI systems. While prompt design remains important, leading organisations are discovering that the greatest improvements come from controlling what information the AI can access and use.
Grounding AI involves connecting it to trusted, verified sources of information and establishing clear operational boundaries around how it generates responses.
Rather than relying solely on publicly available information or historical training data, grounded AI systems reference approved sources of truth that have been specifically identified and validated by the organisation.
This approach significantly improves accuracy, consistency and trustworthiness while reducing the likelihood of hallucinations.
A Real-World Example
Tr3dent recently worked with a large international NGO undertaking a significant digital transformation program.
The organisation was leveraging Agentic AI to assist with the development and maintenance of complex ecosystem designs, stakeholder maps, value streams, SWOT analyses and strategic roadmaps.
However, there was a critical requirement.
Every output generated by the AI needed to align with globally recognised frameworks such as the United Nations Sustainable Development Goals (SDGs) while also maintaining compliance with emerging regulations including the EU AI Act.
While the use case appeared straightforward, it highlighted an important lesson: achieving meaningful value from Agentic AI requires far more than simply deploying the technology.
Success depends on establishing trusted foundations.
Creating a Verifiable Source of Truth
To address this challenge, the AI was grounded using a combination of internal repositories and externally validated reference sources.
Internal organizational knowledge was connected through secure repositories, while external sources included official SDG frameworks and relevant legislative requirements.
This ensured the AI could prioritise verified information when generating responses.
Instead of relying on potentially outdated information or drawing conclusions from unrestricted internet sources, the system was required to reference approved datasets and trusted frameworks.
The result was greater confidence in the outputs being generated, improved consistency and stronger alignment with organisational objectives and regulatory requirements.
Why This Matters for Smart Cities
As cities become increasingly data-driven, the importance of trustworthy AI will only grow.
From urban planning and infrastructure design to sustainability reporting, digital twins and community engagement, AI is becoming embedded within many aspects of city operations and decision-making.
However, the quality of outcomes will always depend on the quality and reliability of the information being used.
Grounding AI helps ensure that decisions are informed by trusted data, recognised standards and verifiable sources rather than assumptions or unreliable information.
For smart city projects, where decisions can influence public services, infrastructure investments and long-term strategic planning, this level of confidence is essential.
Key Takeaways
The experience highlights several important lessons for organisations implementing AI:
• Ground AI systems using trusted internal and external data sources.
• Establish clear governance frameworks and operational boundaries.
• Prioritise verifiable information over unrestricted public data.
• Align AI outputs with recognised standards, regulations and organisational objectives.
• Focus on building trust, transparency and accountability from the outset.
As AI adoption accelerates, organisations that invest in strong governance and trusted foundations will be best positioned to unlock its full value.
The future of AI is not simply about generating faster answers.
It is about generating answers that can be trusted.
SCC Marketplace Partner Spotlight
Tr3dent is a Smart Cities Council Marketplace partner helping organisations navigate digital transformation, AI implementation and governance challenges. Through practical experience deploying Agentic AI solutions in complex organisational environments, Tr3dent helps clients establish trusted foundations that improve accuracy, compliance and confidence in AI-driven decision-making.
To learn more, visit the Tr3dent Marketplaceprofile and connect with the team.
For more information on Smart Cities Council programs:
