Start-up streamlines construction process with AI document handling

With Civils.ai, construction companies can save up to 20% in costs.

Artificial intelligence (AI) has found its place in the construction industry, mainly in document handling. Major construction projects often face significant delays due to the manual management of critical documents like PDFs and CAD files, leading to rework and human errors.

Recognising these inefficiencies, construction AI startup Civils.ai is leveraging AI to transform the industry, helping players save 10% to 20% in costs by efficiently automating complex documentation.

“What we’re trying to do in the long term is to eliminate all the rework that [is] due to manual activity on design,” Mirko Vairo, COO and co-founder of Civils.ai, told Singapore Business Review. “The impact of these reworks is huge in the industry. Sometimes one structure or infrastructure can go billions over the budget.”

When Vairo, an AI expert, met his co-founder Stevan Lukic — a former civil engineer and web developer — Lukic pointed out numerous inefficiencies in his engineering work. These were mainly related to the document sets they needed to analyse for the project, which most of the time are on hard formats.

“We said, okay, why don’t we try to use AI to automate some of these processes of extracting data, finding information, and eventually start designing the structure directly with the support of AI,” Vairo recalled.

With Civils.ai, project documents are collected from construction companies and analysed and sorted by AI to help engineers easily extract data. This eliminates the need to spend time on documents, allowing engineers to focus more on the design of the project instead.

AI in construction

Vairo pointed out that in the current AI construction scene, different solutions can be classified along two axes. One axis represents the main scope of the tool, including information storage systems and information analysis systems. The other axis represents the target users, which include technical and non-technical teams.

“The information storage systems are the most famous ones: Autodesk Cloud, Procore. These [are] big names within construction, providing services [for a long] time. [But] you have the new players incoming, trying to provide some vertical service for the industry on some specific topic,” he said.

“The other axis is the target users which include the non-technical teams, [These are] people that do not work on the design and the implementation of the calculations and everything. And on the other side, you have people that are more technical, like the engineers or the architects,” he added.

Vairo emphasised that it is important to know how well solutions address specific use cases. He noted that whilst many AI tools are being launched, the real challenge lies in their effectiveness with data and the depth of their engagement with particular use cases.

Looking ahead, Civils.ai remains confident in its position in the industry in the face of potential competition — since AI is increasingly being innovated and used in all sectors.

“If you want to be very vertical to some use case, you need to build new and different things that you integrate altogether in order to approach that use case in the right way and give back an outcome that is useful for the engineers,” the AI expert said. “Now, to get to that point is not easy. It takes months of learning, of trying stuff on that specific use case.”

Civils.ai is confident it is at an advantage, being the first mover in this specific segment. Vairo affirmed the wealth of knowledge they generated over the last two years. “Even in the coming years, months… [we] will generate a huge gap against the others,” he said.

In addition to pioneering this segment, Civils.ai has focused on search engine optimisation (SEO) from Day 1 to drive website traffic and attract potential customers. Through this strategy, Civils.ai has attracted 25,000 unique visitors per month, with a growth rate of 10%.

“We have a bunch of free tools that are useful for engineers on projects [and] people search for this tool for help, and they find our free solutions; and so people are attracted to come in on the website and start using our stuff. Then they start to be aware about all the other premium services we do, and they contact us on the other side,” Vairo said.

“We have started to position ourselves and our website on the AI side before the craziness of the last year when ChatGPT launched their solution, and everyone was searching about AI.”

In just two years, Civils.ai has captured the attention of some of the largest global players. In its base in Singapore, the start-up has shared its expertise with industry giants like Arup and AECOM. Globally, it has worked with renowned names including Mott Macdonald, Jacobs, and Kajima.

Vairo revealed they have two main target clients across the world: contractors and engineering consultants. “[These are] people [that support] whoever is running the project in designing or running other peripheral activities related to analysing and designing the project. On the other side, you have contractors [and] these are people that will get the job [done] and manage the different teams involved within the project,” he said.

Civils.ai offers its clients two options: they can either integrate its services directly into their existing processes using APIs or seamlessly leverage its intuitive front-end if they lack an internal solution. Companies can significantly cut costs by using Civils.ai’s platform, which reduces manpower needs and minimises rework caused by human errors.

“We target high-value contracts from 50k above and the way of making profit is pushing the customer to use more and more of the products and increase the value, because we are generating an impact on their productivity, but especially [for the] costs [of] the entire project,” Vairo said.

Funding and initiatives

Since its inception, the start-up has skillfully manoeuvred through the industry’s challenges, buoyed by $300k in funding from esteemed investors such as Antler and Iterative.

“When we entered the market, there were some events [that shocked] the world, like the wars, inflations and all the other things, and that completely changed the approach of VCs on the market in terms of how they judge investments,” Vairo recalled.

But to keep venturing into this path, Civils.ai needs to secure additional funding to meet the new requirements for early-stage start-ups. “We are currently raising a seed round that targets $1.5m, speaking with a few investors and hope to close it soon,” he said.

“So, the funding at the moment is not huge; I would say we are close to $100k. That’s why we are at the market raising. But with revenues increasing, the burn rate is going down month by month. At some point we could also be sustainable without investors, but obviously, that will slow down the pace for us quite a lot,” he said.

In the next few months, Civils.ai will work on its product roadmap to optimise the information extraction process for engineers with a focus on CAD files analysis which contain the architectural and structural information of the projects.

“[Engineers] need to find [the right one amongst] thousands of these sketches, [extract] the information inside, and then [use that information for] some reporting or analysis they need to run. So, the main target for this year would be reading [these] CAD file[s] very well; that is one of the main challenges the industry has. And then, automatically checking [that] information, verifying [it] to be compliant with local rules or specifications they get for the project. This will be the target for [us] by the end of 2024,” Vairo said.