Ernie - An Internal AI Assistant for ERNACT

At a Glance

Client: ERNACT
Sector: International public-sector network
Scope: Internal AI assistant for all staff
Key capabilities: Document Q&A with exact source references, structured report generation
Delivery: End-to-end software build and deployment, plus staff training

Who ERNACT Are

ERNACT (European Regions Network for the Application of Communications Technology) is an international digital transformation network established in 1991. Managed by Donegal County Council and Derry City & Strabane District Council, it works with member regions across Europe and Atlantic Canada.

Background

Dr Margaret Quinn, who was leading the initiative, described it as both an opportunity for operational improvement and a practical AI adoption exercise:

“Artificial Intelligence has massive potential to change the way businesses work and project management is no exception. AI can streamline processes, support greater efficiency and help us use our time more productively. So we are working with Liam to explore how we can take advantage of AI tools to improve our project management and project development and, at the same time, learn about AI adoption challenges.”

Dr Margaret Quinn, Programme Manager @ ERNACT

The Problem

ERNACT’s teams had built up hundreds of internal documents over years of cross-border project work - lengthy project reports and planning documents, many running to hundreds of pages.

When someone needed a specific detail - for example, how a particular challenge was handled across a previous set of projects - the only option was to manually search through and cross-reference multiple long documents. It was slow, and there was always a risk of missing something.

Staff also needed to regularly produce internal reports in standard company formats. The templates existed, but assembling each report still meant manually pulling information from various sources and formatting it by hand.

The Objective

Reduce the time staff spent searching for information across internal documents, and make report production faster and more consistent using the formats ERNACT already had in place.

What I Delivered

I designed and delivered Ernie, an internal AI assistant available to all staff across the organisation. It addressed two core workflows:

Document search and Q&A: Staff could ask questions in plain language across ERNACT’s full library of internal documentation and get answers drawn directly from those documents. Every answer included exact references back to the source material - so nothing had to be taken on trust.
Structured report generation: Staff could generate internal reports pre-formatted to ERNACT’s existing templates, pulling in relevant content automatically rather than assembling everything from scratch.

How I Approached It

I led the project end-to-end: from initial discovery and scoping, through solution design and build, to deployment within ERNACT’s own infrastructure, followed by hands-on training so staff could use the tool confidently from day one.

Under the hood, the system uses a retrieval pipeline that searches across ERNACT’s document library and returns relevant passages to generate answers.

A significant part of the technical work involved handling the realities of ERNACT’s document library: large, complex PDFs that needed to be processed reliably so that nothing was lost or misrepresented in the system’s responses.

Security and Compliance

GDPR considerations informed design and rollout decisions throughout. The system was hosted on ERNACT’s own AWS infrastructure, and personally identifiable information was handled especially carefully.

Outcome

Ernie is now part of operations at ERNACT. Teams can surface information across years of project documentation in seconds, verify answers against the original sources, and produce consistent reports with significantly less manual effort.

Work that previously required reading through and cross-checking multiple lengthy documents can now be completed much faster with clear source references.

You can read more about the project here.

“Liam designed & deployed our internal AI chatbot and provided training to staff on how to use it effectively. He was professional and straightforward, and provided valuable advice on data security. The project was a huge success, and the chatbot is now part of our operations. I’d happily recommend Liam to any organisation seeking help with AI.”

Caitriona Strain, General Manager @ ERNACT