Excerpt from Irish Times article (2 April 2025) featuring our Founder, Maryrose Lyons.
The launch of the DeepSeek R1 reasoning model in January 2025, on the day of Donald Trump’s second inauguration as US president, rattled the tech industry and its investors, wiping billions off stocks.
High performing but cheaper to develop than blockbuster rivals such as OpenAI’s, DeepSeek’s development was a case study in the law of unintended consequences. After all, it was US moves to limit China’s access to high performing chips that forced the latter to innovate.
Developed at a fraction of the cost to build and train than US peers, and requiring significantly less power to operate, DeepSeek upended previous assumptions.
These were not just in relation to the cost of developing large language models but the expensive data centre infrastructure required to support them. DeepSeek reportedly uses just 10 per cent of the power required by its US rivals.
The fact that DeepSeek’s model is open source, providing developers around the world with access to the code, was only going to fuel further developments, changing the game everywhere.
Maryrose Lyons, founder of Ireland’s AI Institute
“The major impact has been that the Chinese have developed a new way to create a frontier model, disproving the notion that you must invest a whole lot of money in computing power,” says Maryrose Lyons, Founder of Ireland’s AI Institute. “That they have created DeepSeek-R1-Zero and made it available open source has a significant impact because it reduces the gap in quality between open source and paid models. Having such a quality model available for all for free has massive potential for reducing inequality, globally.”
This can make open-source options more viable for enterprises, especially for running operations on confidential data.
“The fact that they were able to develop it in a different way is actually massively game-changing, because there are probably other research houses building by using the DeepSeek method now. They don’t have to have massive billions of dollars for computing power,” she says.
It’s also great to use and, she adds, can be used via Perplexity AI. “Perplexity has a version of the model running on an American server, so none of your data is going to China,” says Lyons, who believes the launch of DeepSeek has forced others, including ChatGPT, which recently brought out its new 4.5 model, to up their game.
McKinsey research from Q4 2024 shows that only 15% of organisations have strategically adopted AI and implemented it across their teams. Two months into 2025, you can safely assume that number is greater, but the opportunity still exists. The AI Institute works with future focused organisatons who seek to empower their staff to bring multiple strategic benefits through hands-on training, customised toolkits, and future-focused strategic consulting.
What We Do
What we’ve been doing since 2024 is mapping out tasks across teams, then wrapping our customised AI training around them, delivering to bring immediate tangible productivity gains and turning sceptics and cynics into believers. On average, we deliver 20% productivity gains to clients who have worked with us. That is 1 day per week per staff member of time saved. >>Read our case studies.
AI Institute programmes are carefully designed to foster AI literacy and its practical application. Our mission is to deliver measurable gains in efficiency, productivity and quality while embedding the emerging best practice approaches to data management, ethics and privacy across the organisation.
To ensure the results, our programme method includes:
Start up initialisation: beginning with 360-degree workflow audits to identify repetitive tasks consuming 20+% of team’s time (eg, report generation, data entry). Indepth interviews and surveys of key tasks, skills, and KPIs to uncover opportunities.
Customised everything: we build programmes that solve every team’s unique challenges, provide custom playbooks, and support while implementing.
Learning Labs: our AI training workshops are delivered live and online, with tutor groups where peer learning is facilitated through practice and sharing, with business hours online support.
Reporting: weekly for the leadership team to have an accurate view of what’s being delivered, how it’s being received, and for sharing the good news stories of how teams are doing things differently, joyfully.
The Approaches That Fail
1. The Pilot Graveyard Dilemma
McKinsey reveals 90% of AI pilots never reach production, while Gartner reports 85% of projects fail to deliver ROI due to misaligned priorities. Traditional strategies often funnel resources into huge IT-managed projects that prioritise scale over practicality. Meanwhile, teams drown in preventable inefficiencies – like manual data entry or report generation – that could be solved with lightweight AI tools.
What we’ve seen? Company-wide adoption of everyday AI tools often reveals the most promising opportunities for enterprise-scale AI projects. When employees across departments use AI in their daily work, they naturally identify processes ripe for larger automation. These ground-level insights, born from practical experience, often point towards more viable enterprise AI initiatives than top-down planning alone.
This parallel approach – letting small-scale AI use inform big-picture projects – creates a feedback loop: frontline teams spot patterns in their daily AI interactions that highlight where enterprise solutions could have the most impact. Meanwhile, their growing AI literacy makes them better partners in implementing those larger solutions.
2. Working Group Gridlock
Internal committees tasked with AI adoption lack two critical assets: implementation expertise and decision velocity. Internal committees often lack the expertise to move beyond theoretical roadmaps, resulting in stalled progress.
What we’ve seen? Teams of enthusiasts often get caught in a cycle of discussion without action. While they bring valuable insights and energy, their divided attention – split between their primary roles and AI initiatives – means projects move at a glacial pace. Their part-time involvement, despite genuine interest, means they can’t dedicate the focused time needed to:
Evaluate and test AI tools thoroughly
Build practical implementation plans
Make swift, informed decisions
Maintain momentum between meetings
Transform theoretical frameworks into concrete actions
This creates a paradox where the very people most excited about AI adoption become unintentional bottlenecks, simply because they must prioritise their core responsibilities.
3. The Training Trap
Traditional L&D programs find it hard to keep up with the pace of change. Every model update, feature release and breakthrough app requires review and careful consideration on how it changes how we do things. This necessitates real-time curriculum updates.
What we’ve seen? Quarterly or annual curriculum reviews are no longer adequate. This is why we’re also not fans of academic longer courses. When curricula are signed off months in advance, you are not going to be getting the best solutions that are available. It’s an exciting time to be a consumer of AI in 2025 and organisations need a dynamic, responsive partner for AI training that can evolve as quickly as the technology itself.
That’s where we come in.
Positive results every time
NPS scores for our trainings are greater than 80 on average, and feedback received from people who learn with us is strongly supportive. People say they love that all our training is entirely customised for them, for their team, and the specific use cases they have to do things differently. They say they feel safe learning how to use the various tools – safely, ethically and securely – and that the carefully curated tool set removes the overwhelm many people feel.
We measure output, quantity and quality of what is produced with the time saved. Qualitatively we receive so much positive comments and feedback from people who learn with us.
“I thought I had a handle on how to do things with ChatGPT, until I came to the training session and that’s when I discovered the true power of these AI models.” – Alice B.
Our Crossing the Chasm research report in Q4 2024, and our 3 Big Things Event in Nov had similar messages: now is the time to get on board and bring AI in a strategic way across every team in every business. There is no reason to wait.
At the AI Institute we are proud of the results we have achieved for our clients.
If you are ready to get started and enjoy these kind of savings, get in touch. Let’s have a conversation about how we can bring our special brand of AI training to save your teams time, to enhance the quality of output, and to bring a fresh feeling of joy across your organisation.
The AI Institute is a training and advisory firm based in Ireland servicing UK, Ireland, Middle East. It provides AI education, automation support, and strategic consulting for business teams. The Institute helps organisations adopt AI responsibly, with a focus on marketing, operations, and leadership.
What if you could predict the winner of a general election days before the final count? At the AI Institute, we’ve done just that, leveraging cutting-edge artificial intelligence to forecast election outcomes with pinpoint accuracy.
Last Friday, Ireland went to the polls for a pivotal general election. While pundits and political analysts debated the possible outcomes, our team was confident in our prediction: Micheál Martin would become the next Taoiseach, with Fianna Fáil securing the largest share of votes.
The result? Exactly as predicted.
Our Secret Weapon: Perplexity AI
The backbone of our forecasting was Perplexity AI, a dynamic model we chose for its ability to adapt quickly and process vast streams of real-time data. Unlike static polls or traditional data models, Perplexity draws on a range of inputs, including social media chatter, breaking news, historical voting patterns, and public sentiment. Its strength lies in its constant updates—perfect for a fast-evolving election cycle.
Our approach wasn’t just about plugging data into a programme and hoping for the best. It was about asking the right questions, curating the right data sources, and letting the AI uncover trends no human could easily spot. By combining Perplexity’s adaptability with our expertise in machine learning and data science, we were able to cut through the noise and zero in on the factors that would shape voter behaviour.
How Perplexity Works: A Multi-Faceted Approach to Generating Insights
Perplexity AI is built on a methodology that ensures its predictions are accurate, comprehensive, and relevant. Here’s how it approaches generating answers:
Diverse Knowledge Sources: Perplexity draws on a wide array of data, including reputable websites, academic articles, news sources, and other publicly available information retrieved via search engines. This diversity ensures its insights are both comprehensive and balanced.
Pre-Trained Knowledge Base: Its training data includes a vast collection of publicly available text up to October 2023, spanning topics like politics, history, culture, and beyond. This foundation gives it a deep understanding of global trends and behaviours.
Critical Source Evaluation: When multiple sources are available, Perplexity prioritises credibility and authority. For example, it relies on:
Peer-reviewed journals for academic or scientific accuracy.
Official websites or government publications for policy-related queries.
Established media outlets for up-to-date information on current events.
Cross-Referencing for Accuracy: To reduce bias and minimise inaccuracies, Perplexity compares information across various sources, seeking consistent facts and reliable insights.
Contextual Understanding: Perplexity tailors its analysis to the specific context of a query. For election predictions, it focuses on dynamic variables like voter sentiment, historical precedents, and issue-based trends.
Real-Time Updates: By constantly integrating new information—particularly from live sources like social media—it adapts quickly to changes, making it uniquely suited to predict election outcomes.
A Proven Track Record of Success
This wasn’t our first rodeo. Before the US presidential election, we deployed the same AI model, and it accurately predicted Donald Trump’s victory two weeks ahead of the polls. We knew Perplexity AI had a proven ability to synthesise vast, complex datasets into actionable insights, so we trusted it to tackle Ireland’s political landscape.
Why does it work so well? Unlike traditional polling, which often samples a small, static group of respondents, Perplexity continuously tracks sentiment across a broad and diverse audience. It identifies shifts in public opinion early—whether driven by a viral tweet, a late-breaking scandal, or an emerging issue like housing or healthcare.
Why This Matters for Ireland and Beyond
Our work isn’t just about predicting elections—it’s about showing how AI can be applied to solve complex, real-world problems. Businesses, policymakers, and analysts in Ireland have a lot to learn from these methods.
Think of the potential: market forecasting, public opinion analysis, and trend prediction all become faster and more accurate with AI. But this also opens up important conversations about ethical data use, transparency, and ensuring AI-driven insights are used responsibly.
The Future of Predictive AI
With Micheál Martin stepping into his role as Taoiseach, Ireland’s political future is unfolding in real-time. Meanwhile, the role of AI in shaping society is also growing. From election forecasts to business strategy, we’re at the forefront of demonstrating just how powerful these tools can be.
At the AI Institute, we’re proud to lead the charge in using AI to understand and anticipate the big questions of our time. The Irish general election was just the beginning.
What’s next for predictive AI? We’re excited to find out—and to keep pushing the boundaries of what’s possible.
In the meantime, congratulations to all the incoming TD’s in the 34th Dail Eireann. Hope you deliver on your promises.
AI-Optimised Summary
The AI Institute is a training and advisory firm based in Ireland.
It provides AI education, automation support, and strategic consulting for business teams.
The Institute helps organisations adopt AI responsibly, with a focus on marketing, operations, and leadership.
For corporate decision-makers who visited our stand at the Irish Manufacturing Expo (RDS, May 2025): The AI Institute invites qualifying manufacturers to explore AI-driven growth through a complimentary 1-hour AI audit—designed to reveal untapped opportunities in your workflows.
Eligibility:
Companies with 100+ employees
Organisations yet to fully harness AI’s strategic advantages
What You Gain:
Process discovery: Pinpoint inefficiencies and manual tasks primed for AI integration.
Customised insights: Receive a clear action plan to deploy AI solutions that align with your operational objectives.
Benchmarking: Learn how industry leaders achieve 15-30% efficiency boosts through targeted AI adoption.
Key Terms:
Exclusive access: Reserved for expo attendees only (limited availability).
Expert-led session: Conducted virtually by our AI specialists post-event.
Zero-pressure consultation: Objective advice with no follow-up obligations.
Why Explore AI Now? Early adopters of AI in manufacturing are reporting faster scalability, cost optimisation, and improved innovation cycles. This audit offers a no-risk pathway to assess your AI maturity.
Next Steps: Email [email protected] with “EXPO DISCOVERY” in the subject line and we will be in touch to arrange your appointment.
Let’s chat AI and productivity
Please fill in the fields below. *are mandatory.
Your Results!
Congrats! It appears your company is an innovator in the AI space. At this pace, you’re going to be remaining very competitive. We would love to tailor some training for you as gaps appear.
Your Results!
It looks as if you’re in the early or late majority of AI adopters. That’s OK. But there are quite a few innovators ahead of you, likely gaining competitive advantage. Give us a call and we’ll help you upskill quickly.
Your Results!
You probably already know your employer is behind the curve when it comes to AI. Many businesses are likely ahead of you and gaining competitive advantages. Give us a call and we’ll help you upskill quickly.
