AI (Artificial Intelligence) appears to be taking over the world of writing. Using a bot, like Chat GPT, is a new holy grail for those who struggle with writer’s block; those whose English teachers prompted a lifelong aversion to words; dyslexic people; and those who are simply too crazy-busy to craft a well-written document.
By Nick Cowan, General Manager Plan A, and Caroline Boot, Director Clever Buying.
Writing beautifully worded documents has never been easier if you believe the hype. Since great bid writing has always been an important key to winning complex contracts, we wanted to know if this could finally take that annoying and time-consuming task away.
So, we asked: can a ‘bot’ write tender responses that take no prisoners on deadlines and quality, and save time, hassle and costs?
We talked to several companies who used Chat GPT to write a tender response, and here’s what they told us.
“What came out was relevant to the question and beautifully written. It captured the key general answers to the questions in the RFT; things like our statement on Quality Assurance, and Environmental Management. Most of the responses were actually pretty relevant to the project!
“But, it stopped there. Basically, the responses could apply to any company. Stuff that’s always won us top scores in Non-Price Attributes wasn’t there.
“Like the clever thinking that our team put into tailoring content that tapped into our company’s unique strengths to address the unsaid hot points for evaluation. And, the quotes from happy contract managers that we have tucked away. And, the ability to weave in subtle references to the skeletons in our competitors’ closets that we knew were bugging the evaluators.”
“The scores were not exceptional. Middle of the pack, which is okay if you can put in a cheap price and win that way. As long as you don’t lose your shirt in the process.”
The tenderers could make use of the material crafted by Chat GPT for some sections of their tender, where project specific intelligence isn’t needed. But, it shouldn’t be a surprise that a Chat bot can’t generate a bespoke methodology, that taps into the discussions and observations that were made on the site visit.
Chat GPT has no way of knowing information that you, or your client, gleaned through brainstorming areas that didn’t go well in the last contract. Which is the kind of tightly focused intel that gets winning scores in tender responses, which smart client agencies thirst for.
We asked some evaluation team members if they had recognised tenders that might have been written with the help of AI. We got a grin and a quick response from one procurement professional.
She said that they spotted AI “from a mile away”, when two tenders came in from different suppliers that had huge similarities in the way they were written.
“It was clear that the tenderers had put little real thought into how the question applied to their project,” she says.
“Chat GPT might be able to write beautifully worded responses, but it won’t be able to manage the project! The result? Both those tender responses scored in the major reservations band.”
That said, Chat GPT could have its uses. Like:
- If there’s 60 percent or more on the price, the non-price attributes are a waste of time anyway. Chat GPT might get you a quick, and hopefully conforming, result.
- If you’ve never written a tender response before, Chat GPT will give you a few clues. But you’ll need to check everything, super carefully!
- If don’t really care about winning the job, but want to put in a bid to let the agency know you’re there – go for it! Just make sure that the Chat bot doesn’t harm your image, rather than impressing your potential clients. Irrelevant, generic responses have a habit of undermining goodwill.
In the short term, Chat GPT will certainly put pressure on tender writers who do little more than cut and paste the content from the last bid.
Almost every tender scoring schedule has its antennae out for generic mushy content that could apply to any tenderer on any contract – that inevitably goes to the bottom of the pile.
Just as a top-notch RFT or RFP includes scored Non-Price Attributes and asks contract-specific questions, winning tender responses demand tailor-made ingenious responses.
Winning requires the input of your keenest minds to craft a bespoke, clever and tailored response. It’ll be the one that seeks and delivers on the client’s hidden agendas, and subtly reminds evaluators about the skeletons in your competitors’ closets. Winning responses tap into that elusive in-depth knowledge of what evaluators look for and how, exactly, they score it.
Maybe one day, AI will be able to tap into the unconscious, the site-visit discussions, and the reputational intelligence that the best proposal writers use to generate content that scores at the top of the ranks.
AI might work if agencies rate a beautifully worded generic response, as high as one that shows the tenderer can complete the technical aspects of the job. But we don’t see that coming any time soon.
Put simply, if you want to win, there are no magic shortcuts. High scores are awarded to those who put serious effort and real (not artificial) intelligence into:
- Understanding their clients’ real and hidden needs;
- accessing knowledge outside the public arena to understand their fears and uncertainties;
- bringing the strengths of your company’s thinking and planning to directly address the questions asked, in a project-specific context.
So, go ahead, use the ‘bots’ to replace or update the generic mush that forms your first attempt at a tender’s library. Maybe you don’t need a typist anymore!
But then, pull out the stops using the expertise and insight that taps into the personal experience, the hidden messages, the anecdotes that prove your company deep understanding of your client’s issues on this contract … will deliver.
Parting words from Jeremy Sole- a final column