Writing Smarter: The Role of Artificial Intelligence in Winning Proposals

Ashley (Kayes) Floro, CPP APMP • February 23, 2026

The use of generative artificial intelligence (AI) tools is becoming almost commonplace in our daily lives, much in the same way that we use cell phones and the internet without a second thought. And why not—it’s so easy to enter a prompt into ChatGPT and use the response to start an email, thank you letter, or social media post. So, it’s not surprising that the explosion of generative AI tools like ChatGPT, Claude, and others has sparked major interest—and debate—among those in the proposal community. 


AI promises so many benefits to our proposal process: it can save time and increase productivity, it can help brainstorm win themes and messaging, it can improve readability and clarity, and it can serve as a knowledge assistant. This can help reduce burnout on proposal teams since AI is saving the team time doing the “grunt work” so that proposal team members can focus on higher-value tasks. 


However, AI use does not come without potential risks, including confidentiality and data security risks. Inputting sensitive content—like solution content, past performance content, pricing, client names, etc.—into public AI tools can breach Non-Disclosure Agreements (NDAs) or contracts, expose proprietary or competitive information to third parties, or violate privacy regulations. Knowing the potential benefits, as well as the risks, you and your teams may be asking how you can responsibly use AI to draft sections, develop win themes, or tailor boilerplate content. And how can you do so without introducing AI generated errors? Let’s dig into it!


How You Can Responsibly Use AI to Draft Sections, Develop Win Themes, or Tailor Boilerplate Content

If used responsibly, you can use AI prompts to suggest creative ways to frame your team’s differentiators and benefits, help align your solution with the customer’s priorities and hot buttons, and provide alternative ways to express key messages to keep language compelling. You can also use AI to help rewrite dense, jargon-heavy technical text into clear, persuasive, customer-focused content. Here are some ways you can minimize risk while using AI to support proposal development:


  • Set a policy and train your team 
  • Never input confidential or proprietary information into public AI tools
  • Always treat AI output as a draft, never a final deliverable


Set a policy and train your team

Before you allow teams to start using AI to support proposal development, you should create a written guideline or policy. The policy should define when and how team members may use AI, which tools are approved, and how to mitigate risks while using AI. The policy should define clear boundaries for AI use, considering where it adds value and where it doesn’t. Remember, AI is good for things like:


  • Drafting boilerplate content, like company overviews or standard capabilities statements (after it is fed with source materials)
  • Generating ideas for win themes based on key differentiators and customer hot buttons you provide (note that capture still has to provide this information!)
  • Helping tailor boilerplate to meet solicitation requirements
  • Checking readability and suggesting stylistic improvements
  • Summarizing content


AI is not good for:


  • Making compliance decisions: it may miss mandatory requirements
  • Factual accuracy: it can invent or misstate facts
  • Interpreting ambiguous solicitation language: human judgement is still needed for this!


Once you set the guidelines, train your team! Make sure everyone knows and understands the guidance. Be sure to highlight the data privacy risks and concerns. For many companies, entering proprietary data into a public AI tool may be grounds for termination. 


Never input confidential or proprietary information into public AI tools

Although we have already touched on this point, it really deserves some additional attention. But the bottom line is that you should never paste client-sensitive content, proprietary solution information, or internal pricing into tools like ChatGPT unless your company has a private, secure instance. Many public AI tools store user inputs and use them to further train their models. That means your sensitive information could remain on their servers indefinitely—and possibly reappear in responses to other users. Even with private, secure instances, some companies may be concerned with data breaches or cyber-attacks. If this is the case, consider developing a policy to redact or anonymize sensitive names and figures before asking AI to help tailor content, even in your private instance of the tool.


Always treat AI output as a draft, never a final deliverable

While I was at the APMP Bid and Proposal Conference in Nashville earlier this year, I heard a story about a team that was thrown out of competition because they used AI to write their proposal and then didn’t tailor it. The customer told them that another team submitted the exact same response. I am not sure I was able to hold back the level of shock that a team would submit content without adjusting it—but after hearing that story, this really needs to be said. For so many reasons, you should always review, revise, and tailor the content you receive from your AI tool. The best advice I have heard is to treat AI like a junior writer or assistant—its suggestions still need review, fact-checking, and editing by your experienced proposal team. Just like we have always done as part of our proposal best practices, have a human team member review all your content for compliance, accuracy, your specific proposal style guide, and tone—especially the content developed initially by AI.


How Can You Avoid Introducing Errors or “Hallucinations” When Using AI?

In the context of generative AI tools, a hallucination happens when the AI generates output that is factually incorrect or fabricated, but it still presents the information confidently, as though it were true. Examples of common AI hallucinations include:


  • Inventing a certification your company doesn’t hold
  • Citing a law, regulation, or standard that doesn’t exist
  • Referencing past performance examples or customer names that aren’t real
  • Providing made up statistics or figures 


Hallucinations happen because AI models don’t actually know or understand facts—they predict likely sequences of words based on patterns they’ve seen during training. When they can’t find the answer in the data they were trained on or in what you provided in your prompt, they sometimes simply generate something that sounds plausible. 


This is why, especially when teams are using AI, it is critical for proposal managers and Subject Matter Experts (SMEs) to actively stay involved throughout the process. Here are some steps you can follow to help avoid hallucinations and errors when using AI:


Always start with a compliance matrix

Get back to the basics: build your compliance matrix first and track every requirement explicitly. You can have AI generate a first cut, but then you need to go back and add in everything that the tool may have missed. Next, just as best practice has always told us to do, have a human peer review and validate the matrix. Then use the matrix as the source of truth, and verify that every section written, whether by AI or by a human, maps back to the correct requirement. Remember, AI can help you phrase responses, but it doesn’t reliably recognize all mandatory instructions, page limits, or formatting rules.


Use AI as a support tool, not a decision-maker

Remember that AI can draft, suggest, and rephrase, but it doesn’t understand the legal or contractual weight of a solicitation. We’ve mentioned this already, but always have a SME and/or compliance lead review every section of the response. Don’t throw out your best practice review cycles!


Fact-check everything

Again—another “old school” best practice that has become ever more important in the age of AI. Because AI is prone to hallucinations, it may confidently invent product specifications, certifications, client names, or achievements. To catch these errors, first make sure you provide the AI with accurate source material. Then as part of your review process, require the reviewers to check all the information, including data points, dates, names of agencies or organizations, and references to laws, standards, or regulations.


Feed AI verified content: don’t let it guess

Don’t ask AI open-ended questions, such as “What are the key benefits of our solution?” unless you also give it source content to work with. Instead, first give the tool your actual product or solution specifications, differentiators, and past performance examples and ask it to organize, rephrase, or summarize those into proposal language. Remember, only do this if you are using a paid/private version of the tool. 


Leverage secure, organization-approved AI tools

If you are going to introduce AI tools into your proposal process, it’s best to use private or enterprise-grade AI systems that can be fine-tuned on your approved boilerplate and style guides. Public tools are trained on general data and will be less aligned with your standards—and they also pose serious confidentiality and data security risks if not used carefully. 


Final Thoughts

As generative AI tools continue to evolve and become more embedded in our workflows, proposal teams have an opportunity to harness them responsibly and effectively to produce efficiencies in our winning proposal processes. By understanding both the benefits and the risks, and by establishing clear policies, training, and review processes, you can use AI to enhance productivity without compromising compliance, accuracy, or confidentiality. Ultimately, AI should serve as a supportive assistant, not a substitute for human judgment, expertise, and quality control. Remember:


  • Always fact-check AI output against trusted sources
  • Provide the AI tool with accurate, complete input material (don’t let it guess)
  • Have SMEs review content for accuracy
  • Don’t let AI generate sections from scratch without oversight


With the right balance, AI can help your team work smarter, reduce burnout, and deliver stronger, more competitive proposals!

From Submission to Success: How Lessons Learned Drive Better Proposals

By Ashley (Kayes) Floro, CPP APMP • March 30, 2026
When was the last time your team truly examined why you won—or lost—a proposal? Every submission your team makes, win or lose, contains a roadmap for doing better next time. Yet many organizations treat each proposal as a standalone event, moving quickly from one bid to the next without pausing to reflect on what worked, what didn't, and why. This is a costly mistake. A structured lessons learned program, built into every stage of the business development lifecycle, is one of the most powerful tools a company can use to sharpen its competitive edge. Conducting Lessons Learned Conducting lessons learned after each proposal submission is a critical part of the business development lifecycle. It helps companies understand where they are excelling and where they need to improve. To ensure the experience is fresh in everyone's mind, each member of the proposal team should document their impressions — both positive and negative — within the first week after submission. Sample questions to consider include: Was the proposal development schedule reasonable and realistic? Why or why not? Were there any bottlenecks or major issues? If so, what were they, and how could they be mitigated in the future? Did the team work well together? If not, how could team dynamics have been improved? How effective was communication among the team? What went well? What could have been improved? Did any unexpected problems occur during proposal development? If so, how could they be mitigated going forward? Did the team stay within its B&P budget? If not, what could have been done differently? What worked best during the capture and proposal effort? What areas require improvement? A practical way to gather and analyze this feedback is to send a survey to each team member using an automated tool, which makes it easier to collate and compare responses. After Action Report Once the results are in, the Proposal Manager should review the feedback and prepare an After Action Report that details lessons learned and recommended next steps. This report should be shared with the full proposal team to ensure that insights are carried forward into future efforts. Lessons Learned Session Additionally, after contract award is announced, the team should conduct a formal Lessons Learned Session to document and discuss observations, findings, and conclusions — win or lose. By understanding where the team encountered roadblocks, and where the customer found gaps in the response, the team can address those issues and strengthen both the process and the final product on future efforts. Equally important: identify what the team is doing well and make sure those practices are preserved and repeated. Analyzing Trends and Updating Standard Operating Procedures (SOPs) Conducting lessons learned after each proposal is valuable, but the benefit compounds when you step back and look at the bigger picture. On an annual basis, review your After Action Reports and lessons learned debriefs as a body of work, and analyze them for recurring themes and patterns. As the year wraps up, whether you follow a corporate fiscal year or the calendar year, ask yourself: What challenges keep surfacing? Where does the team consistently perform well? Sharing these trends with your team creates a culture of transparency and accountability, and helps focus improvement efforts where they matter most. More importantly, translate those findings into action by updating your business development and proposal SOPs. If internal feedback shows the team is consistently scrambling during production, adjust your SOPs to launch the production process earlier. If customer debriefs repeatedly cite a lack of customer understanding, take a hard look at your capture process and strengthen your call plan execution. Continuously refining your processes in response to real data is one of the clearest paths to improved performance—and more wins. Final Thoughts Every organization in this industry wants to win more, and win rates are often cited as the headline measure of a business development organization's health. While they are a useful starting point, win rates alone don't tell the whole story. Too many variables influence any single outcome. What matters more is building the discipline to learn from every effort, regardless of the result. A consistent lessons learned program, paired with annual trend analysis and a willingness to update your processes, creates a feedback loop that makes your team sharper over time. The companies that win consistently aren't just the ones with the best writers or the biggest budgets, they're the ones that treat every proposal, win or lose, as an opportunity to get better.

Six Strategies to Tackle Tight Page Limitations

By Ashley (Kayes) Floro, CPP APMP • March 25, 2026
Tight page limitations are continuing to be a challenge as contracting officers streamline their acquisition processes. When faced with tight page restrictions, we often find ourselves struggling with trimming five pages of material into two pages of allocated space. However, sometimes the content we are working with is so long because it is simply overly wordy. In this article, I present six tricks for eliminating waste. 1. Use Active Voice With active voice, the subject of the sentence comes first and performs the action in the sentence. Active voice is more straightforward and concise than passive voice. It typically results in shorter, sharper sentences. So not only does it take up less real estate, it flows better and is easier to understand. Passive: It was decided by the Program Manager to streamline the program. Active, Strong Verb: The Program Manager streamlined the program. 2. Eliminate Redundancies Remove redundancies that take up extra space and don’t add value. I present some examples below.
icons demonstrating how to write clearly

Write Clearly: Why “More” Isn’t Always Better

By Ashley (Kayes) Floro, CPP APMP • March 23, 2026
In the world of proposal development, there’s a persistent misconception that longer writing signals deeper thinking. Teams sometimes feel pressure to fill pages, add more qualifiers, or expand explanations in hopes that additional words will make their message more persuasive. However, the opposite is often true. Clear writing is powerful because it makes it easy for the reader to understand, evaluate, and remember your message. The goal should be clarity, not volume. The most effective writers know that concise, direct language carries more impact than dense paragraphs and complicated phrasing. In this article, we present seven practical tips to help you write more clearly and effectively. 1. Break Up Long Sentences and Paragraphs Long sentences are one of the most common causes of unclear writing. When a sentence stretches beyond 25–30 words, it is easy for readers to lose track of the main point. Instead of packing multiple ideas into a single sentence, break them into shorter, focused statements. Each sentence should communicate one main idea. Example Less clear: Our team will implement a comprehensive data management framework designed to enhance reporting capabilities while also improving accessibility for users across multiple departments. Clearer: Our team will implement a comprehensive data management framework. This approach improves reporting and makes data more accessible across departments. Shorter sentences reduce cognitive load and help readers absorb information quickly. Similarly, large blocks of text can overwhelm readers. Each paragraph should focus on a single idea or topic. If a paragraph begins to cover multiple points, consider splitting it. Shorter paragraphs make it easier for readers to scan and process information. 2. Avoid Nominalizations Nominalizations occur when verbs are turned into nouns, often ending in -tion, -ment, or -ance. While they are sometimes necessary, they can make writing more abstract and wordier. Whenever possible, convert nominalizations back into strong verbs. Example Wordy: The implementation of the solution will result in the improvement of operational efficiency. Clearer: Implementing the solution will improve operational efficiency. Strong verbs make writing more direct and easier to understand. 3. Choose Strong, Specific Verbs Weak verbs like make, do, provide, conduct, or perform typically require additional words to explain what is happening. Strong verbs communicate action more clearly and concisely. Example Weak: Our team will conduct an analysis of system performance. Stronger: Our team will analyze system performance. Replacing weak verb phrases with precise verbs makes writing sharper and more confident. 4. Remove Unnecessary Words Many phrases in proposal writing add length without adding meaning. Words like very, really, quite, and in order to clutter your sentences. Look for opportunities to tighten phrasing. Examples In order to → To Due to the fact that → Because At this point in time → Now The goal isn’t to eliminate detail, it’s to eliminate filler. 5. Use Active Voice When Possible Active voice makes it clear who is responsible for an action and typically produces shorter sentences. Passive voice can be useful in certain situations, but overuse can make writing vague and indirect. Example Passive: The report will be completed by the team next week. Active: The team will complete the report next week. Active voice improves clarity and accountability. 6. Use Lists When Appropriate When presenting multiple related items—steps, benefits, features, or requirements—lists can improve readability. Lists allow readers to quickly understand key points without digging through dense paragraphs. They also highlight structure and make complex information easier to follow. Final Thoughts When readers can quickly understand your message, they are far more likely to absorb your ideas and act on them. Remember: strong writing isn’t measured by how many words you use. It’s measured by how clearly those words communicate your message.