이 에피소드에 관해
In this episode of Data Chat, Iain Curtis sits down withJordan Connachie, MD of Kori Construction, a specialist main contractor in the later living and care home sector turning over around 120 million pounds and running about 15 projects at once. Two years ago Jordan took a punt and hired a data engineer before he could prove the return. This is the story of what happened next.
Jordan walks through Kori's move from Sage 50 and spreadsheets to a single, governed data platform that now answers plain language questions in one click. He is refreshingly blunt about why he started with the boring layer underneath the data rather than chasing AI, why he bans the word AI in his own product meetings, and how a clean data set turned into faster, better decisions on snagging, safety, workforce and margin.
If you lead a construction business and you are wonderingwhere to start with data and AI, this conversation is for you.
Key Topics Discussed:
- Why Jordan hired a data engineer, not an analyst, and why the difference matters
- The two phase journey: four years of point solutions, then stitching the data together
- “This is not an AI strategy, it is a data strategy with AI sprinkled on top”
- Building Lumius: a clean, governed data set with an LLM layered on top
- Standardisation and running construction like a manufacturing environment
- Data backed decisions: snagging trends and the move from three point to five point PPE
- Workforce and programme tracking from man hours by trade
- Bringing the team on the journey: transparency, adoption and the culture piece
- Buy versus build, and why Kori had to build until now
- Commercial discipline: protecting margin and growing only on the right work
- Cost versus value, the growth mindset, and the haves and the have nots
- What is next: predictive alerts and the outsourced brain
