Sinead Shackley is the Senior Director of Engineering at Liberty IT, a leader in digital innovation employing over 800 employees across the island of Ireland. The company is part of the Fortune-ranked Liberty Mutual Insurance and one of the key drivers behind its global digital enablement journey.
What was the catalyst that convinced your company to seriously invest in generative AI technology? Was there a specific moment or project that proved its potential?
At Liberty IT, we regularly scan for emerging technologies as part of a process we used to call the Futurist Challenge and now call our Digital Progression Program. This used to be a quarterly process within the organisationto scan the landscape for new and emergent technologies todetermine their suitability for our world within insurance.Towards the end of 2022, we did a Futurist Challenge and at this event we identified GenAI as an interesting technology and created a white paper on it.
When ChatGPT went mainstream in early 2023we examined our white paper again and that started everything for us. Because we were already aware of the technology and had already considered how and where we would apply it within our insurance business, we were quickly able to stand up several foundational capabilities including our own in-house version of ChatGPT, our in-house AI platform which currently makes a portfolio of foundational and frontier models available to the use cases, we had already considered potential risks from the technology and so quickly stood up our Responsible AI approach. From there, we then built the rest of the ecosystem, including a north star GenAI architecture, identified use cases and identified and built out reuseable GenAI capabilities.
What measurable impacts have you seen in terms of efficiency, accuracy, or customer satisfaction since implementing these technologies?
We’ve lots of examples, too many to mention but to detail just a few:
Our own in house instance of ChatGPT, we have a significant number of users of this tool, with hundreds of thousands of hours saved, giving people time back to work on other tasks.
Our AI platform which supports a portfolio of frontier and foundational models. These models are in turn used by our teams in their use cases with projections in the many millions of dollars (across all parts of our business).
Our engineers use Github co-pilot to help with coding efficiency.
We also use GenAI in our tech modernisation journey, e.g. database conversions from Sybase to MS-SQLthat were converted seamlessly using GenAI.
We continue to experiment with various models and products. For example, we recently had hundreds of stored procedures that we needed to convert. We tried firstly using the OpenAI models and only had a low success rate. We sought advice from an external third party who offered to come in and do it for us. We thenexperimented again using Claude models and various prompts until we got 100% of the stored procedures converted cleanly and saved us significant external spend!
Have there been any unexpected applications or benefits of generative AI that surprised you?
Liberty has experimented with AI-powered knowledge assistants to help engineers and claims adjusters quickly surface information. The surprise has been how much tacit knowledge (often trapped in SharePoint, Confluence, or Slack threads) could be unlocked, reducing time spent hunting for answers.
What are the biggest technical hurdles you'veencountered in integrating generative AI into legacy insurance systems, and how have you overcome them?
Like I’m sure all companies, we encountered lots of hurdles, particularly in the beginning, when we literally knew nothing. I laugh when I think back to the times when we never had enough tokens and to get more, we had to complete a form and submit it to Microsoft!We also couldn’t get access to models etc, so not enough throughput to support our use cases, problems with resiliency. Layer on, net new things we hadn’teven heard of, content moderation and guardrails etc
Although I would say the largest hurdles aren’t tech related, they are people, process and data related.
It has taken a village of people within Libertyto overcome the hurdles, including colleagues within our privacy, compliance, data science, technology, security, infrastructure and procurement areas.
How has the introduction of generative AI changed the skill requirements for your engineering teams and what initiatives does Liberty currently have in place to address this? What new competencies are you prioritizing in hiring?
Generative AI (GenAI) is transforming how we work and serve customers. At Liberty, we’re equipping all employees with tools, practical training, and guidance to use AI responsibly to ensure that everyone can experiment with confidence, build skills, and help deliver faster, smarter outcomes together.
We have designed and developed a GenAI Learning Mission. This consists of several complementary workstreams for all members of our organisation whether in an engineering role, people leader role, product role etc. The mission includes:
GenAI Literacy
Tools & Ways of Working
Building AI into Apps
Evolving GenAI Skills
People Leader Journey (including modules on change management)
These are all supported with monthly office hours for everyone to come along and connect with SMEs.On top of this we have regular promptathons, unconferences and knowledge shares.
There's ongoing debate about AI replacing versus augmenting human workers. What's your perspective based on what you'veobserved in your organization?
AI tools are being positioned as augmenters of human capability rather than full replacements.
For example, engineers use AI to accelerate coding, testing, and documentation, but the decision-making, architectural thinking, and context-sensitive trade-offs remain firmly human.
Similarly, in customer operations, AI assists with triage, summarisation, and insights, while human employees handle nuanced judgment, empathy, and exceptions.
AI often reshapes roles rather than eliminating them.This creates a premium on adaptability and continuous learning—teams that can learn to leverage AI quickly tend to gain an edge.
There are undeniable efficiency gains: automation of document review, claims processing, or code analysis can cut turnaround times significantly.
However, I am seeing companies also recognise the human value—trust, empathy, ethics, and nuanced decision-making are essential in industries like insurance, where relationships and judgment play a big role.
The trend is toward a hybrid model: AI does the “grunt work,” humans do the “great work.”
What ethical guidelines have you established for AI use, and how do you enforce them across different teams and projects?
Our Responsible AI approval process is a multi-step governance process all Generative AI use-cases at Liberty must undergo.
A RAI Steering Group exists to ensure strategic direction. Separately, a cross-functional RAIC exists to advise on risk with experts from across Legal, Privacy, Security etc
To ensure we follow our “Integrity first” principle, we advocate for a “Responsible AI by Design” process. This means embedding governance in the product development cycle to facilitate an early understanding of the use cases' impact and risks. This streamlines development and deployment by reducing the need for costly corrections or alterations down the road. It also prevents rework as we know that our use of AI will be heavily regulated either by state and federal laws or by our insurance commissioners. Every proposed framework for AI regulations includes a requirement that we have a governance structure and that we document our assessment (and ultimate mitigation) of AI risks.
Sync NI's Autumn 2025 magazine explores how Northern Ireland is moving AI from pilot to production
This issue features exclusive insights from industry leaders on real-world AI implementation, responsible adoption, enterprise infrastructure, and how data-driven innovation is transforming businesses across the region.
Read the Big Data Special Edition free online →
Stay connected with NI's tech community:
Subscribe to our newsletter – Get the latest tech news, job opportunities, and events delivered to your inbox
Visit Sync NI – Your daily source for Northern Ireland technology news
