Navigating the Impact of AI on Federal Proposal Development

I suppose you’ve heard about Artificial Intelligence (AI) in our business; all the cool kids are talking about it. There has been a flurry of presentations about AI at APMP and other events. And the proliferation of AI tools on the market is growing faster than the mold in my son’s bathroom (teenage boys; sheesh). Most people in our business are at least curious about the impact of AI on capture and proposals, and some are concerned about the impact of such tools on our jobs.

What follows is my opinion on the subject. Even at Red Team’s intergalactic headquarters, there are differing opinions on the value and future of AI in our business. So take this content as one source of input to help you think through the subject.

The Bright Side of AI

Let’s start on the positive side. AI tools are better than internet search engines in that they can produce content that is a synthesis of a large volume of available, relevant information. They have the potential to produce first draft content quickly and efficiently. The secret to getting good output from these tools seems to be to refine the prompt you feed the tool, and to continue to revise and refine the prompt until you get output that meets your needs.

The Limitations of Publicly Available AI Tools

An AI tool is only as reliable as the data it is trained on or has access to. Publicly available tools such as ChatGPT build responses to prompts based on public information the tool has scraped from the internet. That means they can’t create a response that is customized to a client’s proprietary solution or to the government customer’s needs. You can feed such a tool specific information on a client’s solution and past performance as part of your prompt, but you risk having that proprietary information ingested into the tool and made available to all other users of the tool.

The good news for our business is that there are some custom tools out there that are unique to a client, so that the tool ingests the company’s past performance information, prior proposals, and other internal information and the system works offline so that there is no risk of proprietary information spilling onto the internet. This seems like a useful advancement in AI tool technology.

One other note about AI tools – Tris Carpenter, Red Team’s General Manager of Strategic Growth, did a search and found over 30 AI proposal response management platforms. It’s obvious this is an area of growth, but what seems likely is there will be a contraction as tools mature, with some tools disappearing and others being absorbed by larger platforms.

Keys to Success with AI Tools in Federal Proposal Development

The following ideas seem to be key for companies to be successful using an AI tool in our market:

• Select a tool that will not expose company data to the internet.
• Provide customer information, along with solution data, win themes, and other capture information as part of the prompt to the tool.
• Refine the prompt you submit to the tool to get better results.
• Don’t expect the tool to create final content; at best, a tool can create first draft pieces.
• Be prepared to refine and, in some cases, drastically rewrite the output from the tool, to make sure the content is tailored to the requirements of the opportunity.

The Government’s Stance on AI

A note on where government agencies are with AI — many agencies are investigating ways to use AI in their acquisition processes, from solicitation development to proposal evaluation. But most agencies who have expressed an opinion about AI tool use by contractors don’t want those contractors to use AI to write their proposals. In fact, at the recent industry day for SEWP VI, NASA announced that bidders have to confirm in their proposals that they did not use AI tools in proposal development.

Final Thoughts

In conclusion, the rise of AI in federal proposal development is a fascinating and transformative development. It offers the promise of greater efficiency and capability in our work, but it also presents challenges related to data security and content customization. This is just the beginning of the conversation on the use of AI in our industry. Tools are maturing rapidly, and new tools are popping up every day. While AI shows some significant potential to create efficiencies, we might keep in mind this idea: Just because we can, doesn’t mean we should (at least until the ground settles a bit under our feet).

Written by Jeff Leitner. Connect with Jeff on LinkedIn!