AI, Machine Learning, construction and bots

Construction is waking up to the opportunities posed by artificial intelligence and machine learning to mine rich data and deliver powerful business insights and predictions.

Artificial intelligence (AI), machine learning and data analytics have been mentioned sporadically in Extranet Evolution since 2013, when I looked at software tools used by India’s Transbit Technologies to automate some construction document control processes. There was then a gap of about three years before the next EE mention of AI, in September 2016 (France’s Finalcad was using AI for object recognition) and one more in 2017 (Unearth using AI for site surveys). The frequency of mentions has since accelerated, with six EE articles touching on AI in 2018. EE covered:

  • ShapeDo’s visual comparison technology (April 2018)
  • Salesforce’s growing technology stack (June 2018)
  • Concord Project Technologies’ T-Con advanced work packaging platform (July 2018)
  • Kreo’s efforts to revolutionise BIM with AI (August 2018)
  • Bricsys’s incorporation of AI into its design and collaboration tools (October 2018), and
  • HoloBuilder adding “SiteAI”, to automate  progress control (December 2018).

UK government policy on AI and construction

In the UK, AI is set to become increasingly important and prominent in the UK architecture, engineering and construction market.

AI sector dealThe tone for this development was set in February 2015 when the UK government’s Digital Built Britain post-Level 2 BIM strategy was first unveiled (post), talking extensively about open data and big data. And over the past four years, successive government and industry reports have increasingly sought to move the industry debate beyond its (over-)obsession with BIM and embrace wider digital thinking. In November 2017, “AI and Big Data” was one of four Grand Challenge areas in the Government’s Industrial Strategy, with AI and both construction highlighted as subjects of forthcoming sector deals. The AI sector deal was published in April 2018 (link) and the Construction Sector Deal followed in July 2018 (link). In the meantime, the National Infrastructure Commission published its Data for the Public Good report in November 2017 (link) setting out a roadmap towards a ‘national digital twin’ for the UK’s national infrastructure enabled by secure sharing of high quality, standardised data (incidentally, Neil Thompson wrote a fascinating article [£] about connecting data for Building magazine recently).

Joined-up thinking across Whitehall departments and agencies, and across industry and academia, is evident in recent announcements about these areas. In November 2018, for example, the Cambridge Centre for Digital Built Britain published the Gemini Principles intended to underpin the UK’s digital twin strategy (link). And there have been two moves this month:

  • nPlan - logoOn 5 February 2019, UK Research and Innovation announced £18m funding through the Industrial Strategy Challenge Fund for research, development and innovation projects to support the construction industry. The use of AI to predict and plan construction projects and avoid possible delays was one idea to win funding. The project, by nPlan, Kier and the University of Cambridge, will use AI and algorithms to better predict, plan and schedule construction projects (nPlan featured in this Building magazine article [£] last month). Another project will look at integrating voice-activated AI and augmented reality in the assembly of components to speed up construction and increase productivity.
  • On 21 February, the government announced it would investing in growing the UK’s skills in the AI sector, creating a nationwide programme of industry-funded AI Masters courses coupled with work-based placements, and funding research at 16 AI centres for doctoral training.

Botmore looking for partners and pilot projects

BotmoreAgainst this backdrop, both established technology businesses and startups are targetting AI opportunities in UK construction, and looking to increase R&D work in these fields.

In addition to those already mentioned, Extranet Evolution recently spoke to Aydin Ozcekic, CEO of Botmore Technology.

Bringing together AI and machine learning experts at Istanbul Technical University with counterparts in the UK, Ozcekic said Botmore is developing a ‘digital assistant’ capable of analysing data from existing applications (he mentioned Oracle’s Primavera, Autodesk’s BIM 360, and Tekla BIMsight as examples) and then helping engineers make informed predictions (Ozcekic has written about deploying AI in construction, and about data sources in construction, and spoke at Digital Construction Week in London in October 2018).

ConBot is a prototype application helping teams combine conventional project data, sensor-derived data (the ‘internet of things’) and interactive chatbot technologies, using natural language processing techniques. Ozcekic is interested in talking to potential technology partners, and also to construction businesses that might want to pilot test the technologies on construction projects.

Update (17 April 2019)Interesting US-oriented article on the rise of AI and other technologies in construction and real estate.

Permanent link to this article: http://extranetevolution.com/2019/02/ai-machine-learning-construction/