Making AI work for digital government

The conversation around artificial intelligence (AI) in the public sector has rapidly shifted from futuristic speculation to practical implementation, writes Stella Power, Managing Director of Annertech.
For government decision-makers, the challenge is no longer deciding if AI should be adopted, but determining how to deploy it securely, responsibly, and effectively to improve public services.
At Annertech, we believe AI should not replace human judgment. Instead, it should act as an exceptionally efficient assistant, taking care of administrative burdens so public sector teams can focus on what matters most: serving citizens.
Driving innovation
Annertech has been a Drupal Certified Platinum Partner since 2024, actively helping to develop robust, open-source tools and ensure that Drupal retains its reputation as the premium open source CMS for government digital platforms.
Unlike restrictive, costly proprietary systems, this collaborative open-source approach allows public bodies to pool resources, in line with the Government’s Build To Share strategy. Improvements made by one department can feed back into the wider community, ensuring public sector platforms remain secure, sustainable, and cost-effective.
The foundations for scalable public sector AI are already being built by the global Drupal AI initiative.
AI in action
To understand the tangible benefits of this technology, we can look at a recent digital-first platform Annertech developed for a prominent public sector upskilling initiative. The project aimed to future-proof industry training by connecting educational providers, tradespeople, and small businesses with curated resources.

While the platform was highly successful, managing thousands of course listings, images, and varying data sources manually was an administrative bottleneck. Here is how we used AI to transform their operations:
1. Streamlining messy source data
Data feeds from external training providers are notoriously inconsistent. For example, National Framework of Qualifications (NFQ) levels might arrive formatted variously as 9, Level 9 or NFQ Level 9 (Masters). Similarly, delivery modes are often a chaotic mix of part time or online classes.
Instead of importing this fragmented data as it arrives, which would break structured search filters, our AI tools step in to interpret and normalise the variations. It automatically maps them to the correct, unified taxonomy terms, keeping the user experience seamless and the data functional.
2. Course-to-career matching
Linking hundreds of training courses to specific career profiles and skills taxonomies is traditionally a laborious, error-prone process.
We implemented AI-powered role matching for all course imports. The system reads incoming course information and automatically suggests which professional competencies and career paths they relate to. Administrators merely curate and confirm the suggestions, keeping the platform accurate and scalable as content grows.
3. Automated accessibility and smart tagging
The platform manages an extensive library of imagery across course listings and editorial content. Manually writing accurate alt text for web accessibility is incredibly time-consuming, meaning it is often inconsistently applied.
Annertech integrated an AI-powered alt text generator into the website’s content management system. When an editor uploads an image, the system automatically analyses the visual content and generates a draft description using Claude AI.
The editor simply reviews and approves it. We applied the same intelligence to automate image tagging, completely eliminating manual data entry and preventing assets from getting lost in the digital abyss.
4. Automated course aggregation
Rather than requiring staff to manually browse external websites for relevant new educational offerings, we developed intelligent scrapers. These gather new data automatically and assign a relevance score based on keyword matches. Administrators move from searching to curating, significantly improving long-term platform sustainability.
Open source sharing
In line with the Irish Government’s Build to Share strategy, which champions the creation of reusable digital infrastructure to eliminate redundancy and save public funds, the intelligent architecture we built for this project is already being deployed again to power a similar platform for another public sector initiative.
It is a live demonstration of how open-source innovation scales: a technical solution built for one department immediately drives down costs, accelerates delivery times, and streamlines administration for another.
Trust and responsibility
Deploying AI in government requires a higher level of scrutiny than in the private sector. Public trust is your most valuable currency, and keeping it requires adherence to strict governance principles:
- Human oversight: AI is the co-pilot; humans hold the steering wheel. Every piece of AI-generated content or classification must go through a human approval workflow before publication.
- Data privacy and GDPR: You must know exactly where data goes. Public bodies need absolute certainty regarding data processing locations, local residency rules, and robust contractual protections with LLM providers to ensure personal citizen data is never compromised.
- Managing bias: Because AI models are trained on historical data, they risk perpetuating existing inequalities. Regular auditing and inclusive testing are essential to maintain public equity.
Navigating your AI journey
Adopting AI does not mean completely overhauling your digital infrastructure overnight. The most successful implementations start small.
We recommend identifying your organisation’s biggest operational pain points; whether that is an overwhelmed contact centre, inaccessible PDFs, or administrative back-end bottlenecks. From there, find an enthusiastic pilot team within your organisation to test the waters. Once one department demonstrates a clear win, wider organisational change becomes much easier to manage.
The digital landscape is moving quickly, but you do not have to navigate it alone. Partnering with an agency that is embracing the open-source AI frontier responsibly ensures your organisation moves from firefighting administrative burdens to delivering proactive, human-centric public services.
T: 01 524 0312
E: hello@annertech.com
W: www.annertech.com





