Using AI in the bid process

AI promises to be an incredible asset to bid teams. Its ability to curate copy at pace is breathtaking and, in skilled hands, will be a game-changer for bid teams.

Early adopters with the resources to implement AI stand to gain an advantage over those who hesitate or can’t keep pace. Some will develop an in-house solution and try to balance the control it gives with the development cost and time, while others will choose an ‘off-the-shelf’ application with varying degrees of price, functionality, and business support.

One of the most talked-about uses for AI is how it can help bid teams ‘get to first draft’ in a fraction of the normal time. But, like me, you might be thinking, “What else can it do?”

‘Know your customer’ is a mantra as old as sales

One example is the extent to which AI could conduct buyer research. Finding out their procurement history in your market, buying trends, supplier selection, spend data, and opportunity pipelines is just the start. Is there other relevant information out there such as Freedom of Information releases, annual reports, strategies, public records, press releases and social media that, when combined, builds a picture of the buyer organisation and what they care about? The data is already out there, but it’s the speed at which it can now be compiled that’s exciting. Publicly available supplier codes-of-conduct and policies could provide insights into what a buyer looks for from a supplier in more general terms such as behaviours, cyber security, and ESG.

With AI compiling this information into an accessible report with recommendations, the deal team is better equipped to show empathy and understanding that gives a buyer confidence their solution really will solve their immediate and wider strategic problems.

To bid or not to bid

A company’s decision to bid has traditionally been based on several factors ranging from gutfeel to forensic analysis using complex decision matrices. Often, the decision can boil down to a company’s need to hit its sales targets, making it difficult to no-bid because the gap it leaves in the sales funnel can be hard to replace in the short-term.

AI could help inform a company’s decision to bid by assessing several factors simultaneously. These could include wide-ranging market intel, PESTLE analysis, customer analysis, a detailed assessment of the invitation to tender, and the solution required to comply with the buyer’s specification of requirements. This could result in a calculated probability of winning when considering the various risk and reward factors.

However, using AI could be a double-edged sword when qualifying deals, as there could be a temptation to bid more in the belief that using AI to generate proposals faster with fewer resources will mean it requires less work and is therefore worth a punt, something that evaluators won’t thank you for!

Bidder beware

One aspect of the Procurement Act 2023 bidders will want to be aware of is the consequences of being labelled an ‘excluded’ or ‘excludable’ supplier. The Act provides detailed guidance on this, but for the purposes of this article it can be simplified to say that deciding to bid for a contract where the probability there being serious performance issues is high could give rise to extremely severe consequences, including termination of the contract in question and all other public sector contracts the supplier holds. If you decide to bid, don’t just focus on winning – focus on delivering, too.

Exploring the data room

One of the most tedious jobs of any major deal is sifting through the dumping ground of documents compiled by both the buyer and incumbent supplier over the term of the current contract. Yet it’s vitally important as the buyer expects suppliers to take responsibility for doing their due diligence when submitting an offer capable of acceptance. Data rooms (they were actual secure confidential rooms back in the day) are online libraries with sometimes thousands of pages in hundreds of files ranging in relevance and importance. Some of the most important data are the list of staff eligible to be transferred under TUPE, details of existing contracts, and current contract performance reports, which help a bidder more fully understand what they’re taking on.

This is one area where AI could excel if we can find solutions that protect the sensitivity and privacy of the data being processed. One significant benefit could be to analyse the data room contents and cross-check it against the invitation to tender, clarification questions, and the contract to identify gaps, contradictions, errors, risks and opportunities. Done well, this could have significant time saving benefits and help de-risk contract awards for all parties.

Whatever you do, don’t load customer documents into open models such as ChatGPT.

Asking the right questions

Clarification questions can often be badly written, which is ironic given the due care and attention applied to every other aspect of the bidding process. Questions can also be unnecessary and expose a lack of understanding – if you speak to procurement professionals, one of their biggest frustrations is when a bidder has not read all the documents. Worse, the questions don’t always explain why they’re being asked and the consequences of not receiving an informative answer. It therefore shouldn’t come as a surprise when answers are vague or caveated (“it’s up to bidders to decide”).

A niche but beneficial use of AI could go beyond simply writing the clarification question to determining what to ask in the first place and why. Having rapidly analysed the full set of tender documents and the data room, AI could recommend a priority list of questions ranging from essential to nice-to-have. It could also apply insights from customer and competitor analysis to ask strategic thought-provoking questions that will help enhance your brand throughout the process.

Navigating governance

At an old company I worked at, we used to only half-joke that most of the time was spent on bid governance gates to keep the senior leadership team happy and secure their approval to submit a bid, and once the final gate was passed, we could sit down to write the bid.

If there was only one thing that I could use AI for, it would be to prepare the ‘governance packs’ required to progress the opportunity through the various gates. An opportunity summary, win probability, risk and opportunity register, and financial profile could be developed using the available data.

Bid to project handover

When the time comes to handover from the bid to contract delivery, AI could compile the required information for the delivery team using the information already gathered throughout the bid stage. This could include a Project Initiation Document, risk register, method statements, transition plans and other controlled project documents. Apart from speed, the benefit of AI doing this would be the continuity of information from the bid stage, rather than starting from scratch.

In summary

There are several areas where AI can help bidding companies do more than ‘get to first draft quickly’. AI is an emerging technology in a fluid market and therefore its use cases are also still emerging as people begin to imagine the art of the possible.

It’s application across multiple data sources can help deal teams quickly assess risks and opportunities, but care must be taken to sense check and validate the output before acting.

I believe we’ll soon see a flip from AI performing traditional tasks faster to AI driving changes to how competitive tendering can evolve to benefit buyers and bidders alike. This is made more likely with the introduction of the new Competitive Flexible Procedure.

There will be pitfalls along the way, but perhaps those that fail fast will ultimately reap the best rewards in the long-term.