The purpose of this paper
This insight paper focuses on Artificial Intelligence (AI) as an emerging technology and how it might impact organisations that produce bids and proposals. The paper examines the potential pros and cons of integrating AI into a bid function and the wider organisation.
How long has AI existed?
AI is not a new concept and has long been the basis of many sci-fi films and books. An early example of AI in the business world is the Logic Theorist developed in 1955 and used to mimic human problem-solving skills in relation to mathematical problems.
What is AI?
AI refers to the development of computer systems that can perform tasks that typically require human intelligence, such as visual perception, speech recognition, decision-making, and language translation.
AI involves designing algorithms and models that can learn and improve over time from data inputs, allowing them to make predictions, recognize patterns, and automate processes. This includes machine learning, deep learning, natural language processing, computer vision, robotics, and other related fields.
Believe it or not the explanation above has been generated by using ChatGPT and the full explanation ChatGPT offered included potential sectors which could be “revolutionised” by AI including healthcare, finance, transportation, and education. The explanation also identified a number of risks, including privacy and security which will be discussed later in this paper.
Findings of using AI for bidding
As a test case ChatGPT was used as part of stage 2 (Bid Strategy and Kick off) and stage 4 (Bid Response Production) of the Contracts Advance Bid Management Process. A very specific question was asked in relation to producing a Bid Strategy to win a healthcare service that is currently being procured, followed by a series of questions to elicit responses to questions in the ITT.
ChatGPT provided very useful information on the process of producing a bid strategy but struggled to detail an approach that would win. The key elements that were suggested included local research, value proposition, competitive analysis, bid team formation, pricing, partnerships, quality and outcomes, marketing and communications, and the submission process. Whilst information was given on what to do, no information was provided that could be considered useful as a bid strategy.
The responses generated from ChatGPT for the selected ITT questions provided a richer level of information. Responses that related to processes (e.g., complaints) were better than responses that related to more complex areas like service delivery model and workforce planning.
So, how useful is AI when bidding?
The conclusion from this test was that for process related responses, AI provided a good outline draft or an ideas framework that was a starting point to be developed by more experienced bid professionals who better understand the local context and how to develop content that explains:
- What the bidder will do
- How the bidder will do it
- Why the bidder is doing something
- When the bidder has done something before or when they will do it for the specific contract
Contracts Advance has developed its bid management process to maximise scores for written responses based on a winning strategy and this is founded in answering the question in the right order, using the right language, provide evidence and examples, explaining the benefits to customers and consumers and showcasing any added value. In the test undertaken, ChatGPT was NOT able to provide the same level of sophistication as human bid writers and managers.
Risks associated with AI
Whilst there are clearly benefits to using AI to generate ideas and save time in initial drafting, there are significant risks that need to be understood by organisations. Where organisations use AI as part of the bid team/process, the following risks need to be considered and mitigated:
Programming Bias – As AI applications are programmed by humans, there will be ethical and political bias naturally built into each application. This bias may provide information that is not appropriate for some responses, particularly in areas like substance misuse where views are often polarised.
Decision making – The level of decision-making that AI is allowed to undertake should be closely monitored, particularly in relation to scenarios where there could be cataphoric implications in delivery. Taking an extreme example of driverless cars, if there is no other outcome than a car colliding with many pedestrians, how does the AI decide which one it will hit? This is similar if AI were to be allowed to make decisions on pricing or complex delivery models with no true understanding of the real world or the impact of that decision on business performance or safety.
Sharing learning with competitors – As more people/organisations use and test the range of AI platforms, each AI application is learning and absorbing the questions/content and storing this . This content may then be available to other users/organisations as part of the data that the algorithms are interrogating. Without realising this, organisations could be helping competitors simply by testing an AI application, potentially impacting on security, privacy and protocols.
Considerations for the future
AI almost certainly is a valuable addition to a bid team if the constraints are understood and it is used appropriately. It would appear that the best use of AI at this stage of its development is as an idea generator and an initial draft writer for less complex/process-based responses.
In the future, as AI becomes more sophisticated and it becomes more intuitive in piecing together disparate components within a bid, it may be possible to generate a high-quality bid purely using AI. However, as discussed above, human oversight is strongly recommended to ensure that what is generated in virtual reality is safe and applicable to the real world.
As a business Contracts Advance will continually monitor the AI landscape and share its findings with our customer base to ensure we maximise AI’s value.