Introduction
The rapid rise of artificial intelligence (AI) is driving change in how some businesses operate, offering new, quick and innovative ways of working. Tools such as ChatGPT are now widely used and whilst the benefits continue to gain much attention, so do the debates on its limitations.
At Contracts Advance, we have seen an increase in the use of AI, particularly ChatGPT, in the bidding process. Bidding can be a time-consuming and expensive process, so it’s easy to see why some businesses use it. However, many also ask if it can be solely relied upon in bidding or if human input is always going to be needed to be successful. We decided to put this to the test by asking ChatGPT to generate a response to a tender and then asking our senior bid reviewer to evaluate the response.
What we found
When asked ‘Detail what your Social Value offer will be in relation to this equipment supply contract and how it will be of benefit to the Trust and its patients’ ChatGPT generated a response in seconds, answering the question in a well-structured manner. However, when our senior bid reviewer analysed the response, he was quickly able to identify key areas that were lacking in content, context and evidence. This may seem an obvious result as, after all, ChatGPT does not know individual businesses, but when it comes to bidding content, context and evidence are key to achieving a high score.
Content
As said above a key takeaway from reviewing the response was that the content was vague. Content and clarity are hugely important to a tender response as a commissioner can only give points for what they see in front of them. As AI relies on the knowledge that is input, it’s easy to end up with generic responses lacking in depth and clarity. As you can see from the ChatGPT response below, this can raise more questions than answers.
Context
Although AI can create grammatically correct content, we’ve found that it lacks the skill, data and knowledge to write deep, compelling responses with context and a strategic focus. Ambiguous language throughout a bid response can leave commissioners asking questions and language such as ‘we aim to’ or statements such as ‘we’re better than most of our competitors’ need context to ensure credibility. Our ChatGPT response below raised questions due to the language used with no context.
Evidence
Evidence is arguably one of the most important parts of a bid. As AI relies on algorithms and data it is almost impossible for it to input the required evidence, examples and case studies in the bid response. You can see in the ChatGPT response below that a vague statement was generated referring to the benefit of a partnership, with no evidence.
Additional limitations identified
Along with the key elements discussed above, it’s also important to note other limitations so that AI can be leveraged effectively.
Lack of Authenticity
As mentioned AI relies on algorithms and data to generate content, and because of this, data can often be re-produced again and again resulting in responses that lack authenticity and uniqueness. If multiple companies used the same AI for a bid, commissioners could end up with a pile of generic responses, making it impossible for companies to stand out. We at Contracts Advance have seen this already and we’re only in the early days of businesses using AI in bidding.
Lack of Human Touch
AI responses without human intervention can be shallow and almost present an overview rather than the detailed response required in bidding. AI is not able to provide a response with the knowledge, language, evidence and strategic depth that is required to score the highest points in a bid. And although good as an aide in bidding, AI cannot replace human ingenuity.
Benefits of using AI
When reviewing our ChatGPT-generated response, the limitations of solely using it did outweigh the benefits. However, there are some very clear and tangible benefits of using AI in bidding worth highlighting.
Summarising and condensing complex questions/information
AI is extremely good at condensing information into manageable and easy-to-digest content. Therefore, it’s a brilliant tool to support the bid process in the first stages.
Frameworking and planning
AI is great for supporting writers with the planning element of writing a response and can be a great starting point. By providing AI with the specification and response requirements, it can produce an outline of what the response should include and can provide a structure to follow.
Content and idea generation
AI can support writers in the development of ideas, win themes, social value initiatives etc. By using AI to help brainstorm, you can invigorate responses by providing enhanced content. However, it’s crucial that these ideas are enhanced by human creativity and knowledge, to enhance with detail, evidence and strategy.
Research and competitor analysis
Similar to summarising and condensing information, AI is a great tool for sector/competitor research and analysis. Using AI to condense information can be cost and time effective with the information gathered enhancing the bid.
Conclusion – The difference between using and leveraging AI
In conclusion, there is no denying that AI can support the bid process and aid businesses in improving responses. However, it’s important to use AI to enhance responses rather than completely rely on it to write them. Our test using ChatGPT provided a response that would have achieved a low score due to ambiguity, questions having to be asked and lack of evidence. If the response had been used as a guide, enhanced with human creativity and ingenuity, the result would have been a stronger bid response.
One final point – What AI has to say about AI
After this test, we asked AI to tell us why AI is unreliable and shouldn’t be used for tender applications… this is what it had to say:
AI is unreliable and should not be used for tender applications due to its inability to fully comprehend the nuanced and complex requirements often outlined in such documents. Tender applications demand a deep understanding of specific industry standards, legal regulations, and the intricate details of the proposed projects, which AI systems frequently misinterpret or overlook. Additionally, AI-generated responses can lack the strategic insight and tailored precision that human expertise provides, leading to submissions that are generic and potentially non-compliant. The risk of errors, omissions, and a lack of accountability further undermines the credibility and competitiveness of AI-driven tender applications, making them an unreliable choice for such critical tasks.