7 September 2023

When bidding for contracts, businesses should forget AI and focus on competition


It’s difficult to avoid fashionable speculation about how large language models like ChatGPT might be used by writers and businesses. Speculation is a word I chose with some care and with the one, hugely significant attribute ChatGPT can never offer – intelligence. It didn’t appear in my opening sentence because some super-sonic software tombola coughed it up first, out of hundreds of other options. It’s there for two important reasons. One, because I know from decades of experience that speculation is what businesses, NGOs and other major organisations really do week after week, when they compete with other businesses and NGOs for the multi-million pound government funded contracts that sustain many of them. And two, because I anticipate that my readers may spot and enjoy the connection using the word forges, between polysemy and irony. No software can do that. 

It’s an inconvenient but important truth, that for many businesses, particularly in the technology sector, the most important customer is so often the state. Business development is inherently a speculative activity. All public sector procurement happens within national and internationally agreed legal frameworks based on principles of free and open competition, which ideally deliver value for money for the taxpayer. These legal obligations and regulations mean that if you want to sell products or services at scale, then at some point your business will inevitably find itself in intense competition with well-resourced rivals. Failing to resource and reorganise themselves properly to meet this harsh reality is a key, evolutionary step on which many SMEs stumble, before wondering why they’re facing the receivers. 

This bidding, or tender process is the way government departments select the most suitable supplier of a service or product, because it lets them compare suppliers’ proposals against specific criteria. However, it is essentially a process almost entirely dependent on writing. There are rare occasions when the process permits regular contact between the client and a shortlist of suppliers, so that personal relationships and intra-personal skills like sales can play a part, but in most cases, what matters are the words someone in your business commits to the page. 

Whether you like it or not, turning business development ideas and discussions into profit, is usually all about English prose. Bidding is essentially about good writing. Which is of course, why speculation about AI and ChatGPT has become so common. It is however, a spectacular red herring.

Over two decades and for dozens of different business, NGO and research clients, I’ve worked on numerous multi-million pound bids as a writer and Lead Writer, both as a company employee and independent consultant. Those two differing roles matter because employees have different loyalties and pressures to external consultants, who are more free to observe and evaluate how teams of people work together in competition. Over those years I’ve learned a number significant lessons that any business development manager or sales director should find invaluable, when faced with pressure to use AI tools like ChatGPT. 

The first of these is that many bidders don’t evaluate the opportunity either thoroughly enough, or early enough. Large headline sums can be very alluring to sales teams, who don’t often stop to investigate and question their accuracy closely. This situation is often exacerbated by the late inclusion in to bid or not to bid discussions, of accountants, or in the case of technology, engineers, who often look on what an established business development team thinks is a realistic, even exciting opportunity, as just an unwelcome, additional chore. You will quickly discover where the power really lies in any business, if you make this mistake. 

The costs of bidding are high and start from the moment you dedicate someone to the opportunity, yet way more often than not I’ve seen businesses pull out of a bidding process at a very late, and costly stage, either because an individual senior accountant or engineer was never consulted, or decided they simply didn’t want to get involved out of pique, or disinterest. There are occasions when delay, or even God forbid, incompetence from the government procurement side means the costs of bidding just escalate, but experienced business development teams always factor in the inevitable delay; if not the incompetence. 

This points to a second significant lesson about all commercial bidding. It’s a team activity and although businesses love to deploy that word team at almost any opportunity, few grasp what’s really involved in putting the kind of team together who will come out winners in any best supplier competition. 

Team construction in the bidding world is typically a matter of role distribution. Someone, usually in sales or business development, decides what experience and skills are needed to compile a written response to the tender documents produced by the procuring organisation. They then assign individuals within the business to the team, based on those skills and on the content of the tender documents. If security is a key requirement of a technology service, then someone with security expertise will be asked to contribute, and so on. This is where things get really messy and where discussions about whether or not ChatGPT will help, are likely to be damagingly seductive and distracting. 

One of the most difficult challenges facing any bidder is making their written response coherent, when it may be lengthy, complex and authored by multiple employees. I was once asked to rewrite a 40,000 word document which my NGO client luckily realised (4 days before the submission deadline to the government agency that had procured it) was utterly incoherent because of multiple, international authorship. If there’s one thing large language models are specifically designed not to deliver, it’s a single, coherent voice. In most cases it’s inevitable that the first draft of a tender response will have to be written by many different individuals, with specific knowledge and expertise. AI has no facility either to recognise this, or to remedy it. It requires natural intelligence.  

However alluring, innovative or economical using something like ChatGPT might sound, it offers minimal value compared to easily the least well understood, but most effective bidding strategy. Wise businesses dedicate a team to any given competition, even to the point of locating them together in a single room, or removing them from other responsibilities. 

It’s one of the big dividing lines I’ve observed between businesses and NGOs, the latter typically thinking they can pull in expertise and crucially words in the form of bid text, from anyone who might have something to add, regardless of their location, current workload, writing skills or competitive appetite. 

If you work in sales or business development, and have found yourself seriously contemplating using something like ChatGPT to generate written responses to tenders, it really is a reeking red herring. Itself a simple combination of two adjectives that amply demonstrates why you shouldn’t do it. Invest that time and energy instead on identifying credible opportunities before matching them to colleagues who primarily relish competition, enjoy being part of a purposeful team but above all – who really like to win. 

Click here to subscribe to our daily briefing – the best pieces from CapX and across the web.

CapX depends on the generosity of its readers. If you value what we do, please consider making a donation.

Joe Nutt is an essayist who writes for a range of magazines and is the author of five books, mostly about poetry.

Columns are the author's own opinion and do not necessarily reflect the views of CapX.