‘Not being used to reduce labour’: Where we’re up to with AI on site

So, are we there yet?

To find out, GCR sat down with Kenny Ingram, VP of construction and engineering at software company IFS, to ask how AI is being used in the industry.

Like a construction site in the rain, the picture is muddy.

Where do you think the biggest untapped source of potential AI use is in construction?

We spoke to our advisory customers about 12 months ago to try and understand where they were on their journey and what they thought would be the most applicable use cases. This is always where the conversation starts and we got a lot of information that wasn’t very easy to make sense of. The general conclusion a year ago was that companies didn’t really have a clear vision of how they’re going to use AI.

I spoke recently to a massive construction company that’s trying to use AI agents for procurement but they were not able to automate the process, instead having to recruit procurement staff on the construction sites, so there was no business case there for using AI.

I have not seen any real evidence that AI is being used by companies to reduce labour. For example, a tier one UK contractor said it’s not looking to get rid of people and was looking to improve business processes, improve project performance and so forth.

A lot of construction companies include service management, asset management, maintenance, offsite modular manufacturing, rental equipment, predictive maintenance, all of which AI can help with: basically, the tasks that are more repetitive and volume based.

This is different to delivering 20 unique projects a year where it’s difficult to acquire data volume and conduct machine learning.

When I have previously attended roundtables with tier one and two construction companies in the UK, the number one thing people focused on was project forecasting.

It’s not necessarily about getting rid of people. It’s more about allowing people to do the job they’re meant to be doing. Let the system do the number crunching. Let it generate the information. Now you can spend the time actually working out what it means and making informed decisions.

And are there any sectors in particular you think are reluctant to adopt the technology?

All of them. I think the industry is risk averse and quite slow to adopt technology generally. There’s an overall resistance to weaning off Excel spreadsheets, most companies we speak to are usually a complete patchwork filled with different business systems.

It can be worse with bigger companies. I recently worked with one construction company and they were running huge road contracts all in Excel. It’s a £15bn business. They’ve got a finance system in place and all the back office, but not with the operational arm. It’s hard to believe they’re actually running a company that size and doing it in spreadsheets.

It’s not that Excel should be banned. It’s a good tool, but it shouldn’t drive all business processes because you end up with inconsistencies. People tell me they need help with data errors and spend weekends trying to get an invoice or an application out to the client.

Companies want solutions to problems. The bigger the company, the harder it is to make the change culturally because they’re very disjointed and politically very complicated. We end up selling more to the mid-market than we do to tier one firms because they are more receptive, you don’t have to go through layers of politics, which stifles change.

Bigger companies tend to be very siloed; it can seem they want to buy a system for every department.

Is AI creating a skills gap, or helping to skill employees?

Obviously there’s a skill shortage out there and we are asking how IFS can help with that problem.

We think AI will help companies plan better, cutting the amount of materials arriving on site late, or helping with difficulties like changing specifications.

We talk a lot about programmes, but should be driving execution processes. Procurement, labour, plant and site execution are detailed, short-term plans that generally aren’t synched with overarching plans.

I would argue that a lot of the inefficiency and lack of productivity you see is because of a reactive style of management that doesn’t plan properly. So I think what we can do is help firms have plans working in harmony, joined-up processes and less surprises. This lets employees create a plan they know is going to work, with labour and suppliers in tandem.

If you go to construction sites and ask them about inventory control, they don’t really have any. In general, they don’t have a clue what’s on site, so they buy stuff and somebody receives it, sticks a good received note (GRN) in their pocket and then after a week another person can’t find it.

An example of where the technology can help is when you have goods arriving on site, we can scan the GRN on a phone and the AI can read it and find the purchase order and do a receipt transaction.

Further Reading: