How AI can make the difference between winning and losing construction tenders

While bidding on tenders remains the default route to winning new work in the construction sector, it has for too long been a largely manual, time-consuming and error-prone workflow. Now the rise of powerful large language models is encouraging companies to experiment with AI to streamline processes, inform decisions and improve win rates.

With tenders becoming increasingly complex, the push to embrace AI couldn’t be more timely, and the government has advised public sector procurement teams to expect an increase in AI-supported bids. But, in the rush to adopt new technology, some firms are learning the hard way that AI can as easily create problems as solve them. AI generated “hallucinations” and misplaced trust in AI-generated content have already led to negative consequences, including lost contracts and reputational damage. So, before exploring where AI can help, it is worth understanding where it can go wrong.

Where AI can go wrong

A Polish construction company was recently reported to have lost a €3.7m government contract because its proposal contained information suspected to have been made up by AI. It initially won the tender for maintenance work on southern Poland’s road network, but the decision was overturned when a rival bidder spotted references to non-existent legal rulings in its documents.

Mistakes like this are not restricted to construction. A major UK accounting firm was also reportedly caught out when a report it prepared for the Australian government was found to contain errors likely to have been caused by AI hallucinations. The document – a review of welfare compliance systems – allegedly contained references to non-existent sources and experts.

The takeaway from these examples is that relying blindly on AI to create content – including its ability to accurately gather external information – can backfire. But used carefully, and fed with accurate internal data, AI can reduce much of the friction from bid management.

AI-driven qualification

Early in the bid cycle, for example, AI can be used in tender review to help teams prioritise opportunities.

Traditionally, reviewing requests for proposals and tender documents manually has soaked up many hours. And with so much information to go through, it is easy for bid teams to miss essential details, meaning they end up wasting valuable time on tenders they may not have a chance of winning.

AI agents can review tenders running to several hundred pages within seconds, summarising key requirements and identifying red flags such as compliance gaps, legal hurdles and other obstacles. Armed with the right internal data, agents can also compare tender requirements with past performance on similar projects, assessing how the company stacks up against likely competitors and how strongly it can score on key decision-making criteria.

AI-driven qualification like this can help firms make bid/no-bid decisions faster and with greater accuracy, minimising time wasted on unwinnable bids. Because AI tools learn from each submission, the value they deliver in processes like this increases over time.

Taking the pain out of bid co-ordination 

One of the biggest complaints about responding to tenders is co-ordinating input from across the business. Senior bid managers report sometimes spending half their week chasing subcontractors or specialist subject matter experts for input on specific parts of tenders. As well as requesting detailed background on all contractors, complex tenders can typically include hundreds of criteria that require input from experts in areas such as design and engineering, health and safety, cost planning, finance, supply chain, legal compliance and sustainability.

By centralising all tender information and communications within a single environment, AI-driven platforms can free senior bid workers from the administrative grind of assigning tasks, co-ordinating activity and chasing experts for input. AI agents can automatically map every tender requirement and use internal data to generate a draft response for each, tagging the appropriate expert or subcontractor to review. Each contributor can update their section with instant access to the original data source – whether that is a previous proposal, an internal document or an email thread. The system will send automatic reminders to ensure tasks are completed on time. In this way, AI can turn a chaotic mix of emails and Teams messages into a structured workflow. The result is fewer delays and last-minute panics – and more time to develop informed bids that comprehensively respond to the tender.

Learning from every bid

Many bid teams fail to properly analyse and learn from past successes and failures. Which criteria did we score highly on in our bids over the year? Where do we consistently underperform? Why do we always lose when up against certain competitors? 

While public authorities are required by law to deliver post-bid feedback – and many private firms will provide it if requested – too often nothing is ever done with these insights. Important learnings are left buried in emails, Teams chats or online meeting notes. AI can help by aggregating external and internal feedback and turning it into meaningful data. With AI-driven analytics, bid teams can home in on valuable insights – such as where they consistently underperform or which criteria cost them most points – providing the data to improve future win rates.

Becoming AI-ready

To get the most out of AI, companies need to feed it with relevant information. In bid management, this should not stop at uploading past proposals and tender documents. The broader and richer the range of relevant material it can draw on, the greater the value it can deliver – from product and service overviews to documentation on internal processes, pricing models and supplier data. One of the keys to becoming AI-ready is ensuring that information across the organisation is well organised, clearly structured and easily accessible to AI technologies.

As tenders become more demanding, AI is set to become a powerful enabler across the entire bid cycle, from qualification and co-ordination to post-bid analysis. The key is knowing how to use it. Use it blindly and it can produce errors or hallucinations that can derail bids. But feed it with relevant data and combine it with the experience and judgement of human bid professionals, and it can automate processes, enable faster, more accurate decisions – and improve bid performance.

Matthijs Huiskamp is CEO and founder of Altura