Name: Using AI to Master Oral Presentations
Uploaded: 2026-05-07T19:31:04.332Z
Duration: 1 h 1 min 16 s
Description: Using AI to Master Oral Presentations

Transcript for "Using AI to Master Oral Presentations": Hello, and welcome, everyone, to today's Deltek's thought leadership webinar. I'm Nora Bashore, product director of AI proposal here at Deltek. Before we get started, there's just a few quick reminders. For your the best webinar experience, we recommend you use Google Chrome or Firefox. If you have a question, please type it into the q and a box anytime during the presentation. We will address as many questions as we can after the presentation. And if we don't get your question live, your contact information will be shared with Red Team Consulting so they can get you the answer that you need. Resources, including today's slide deck, can be downloaded in the doc section, which should be in the upper right corner of your screen. And in case you need to leave earlier or wanna listen again, you'll receive a link to an on demand recording of today's webinar via email within about twenty four hours after the session ends. Today, Deltek is pleased to partner with Red Team Consulting to bring you actionable intelligence and strategies you can use right away. Today's discussion topic is very timely, and it's called using AI to master oral presentations. Our expert presenters are Blake Harvey and Neil Levine. Blake and Neil, the floor is yours. Excellent. Thank you so much, Nora, and, really appreciate Deltek giving us this opportunity to speak with everyone today and, putting together these great webinars and a great webinar series moving forward. Believe this is first of a few we have slotted this year. So very excited to tackle this topic and and share our thoughts with everyone on how to use AI to master orals presentations. This is an area that Neil and I are very passionate about, and we'll we'll talk through more today. And, we just really think there's underutilized, areas of orals presentations with how you could use AI to prepare for them and and and rehearse with your team. And so we'll cover a variety of topics today. But if you go to the next slide, you can see our faces, who's speaking with you today. So I'm Blake Harvey. I'm the president at Red Team. I've been with the company for close to eight years now. And, at Red Team, we work with over 200 companies a year, and we're well connected to over a thousand companies. And we're constantly speaking with industry leaders and executives on how they're thinking about growth, what what is impacting them in terms of the acquisition landscape, how they're using AI in their growth life cycle. And so, that's some of the perspective we're bringing as part of this presentation, just our lens at the intersection of GovCon and government in terms of, how we view the market from that lens. So, Neil, I'll pass it over to you to introduce yourself as well. Great. I'm Neil Levine. Thank you very much for inviting me here to speak today. I've been, in this field for about, thirty or so years, a little bit more than that. And, I've been at a Red Team working with Red Team for about three and a half years. And, been doing orals for quite a bit, very interested in in AI. I've been the president of two organizations and four been the, director of proposals at four. So, I moved around quite a bit. But, really have some actionable information for you today and really excited to get, going, with you now. So, you know, with that said, let's dive right into some material. What you're gonna walk away today with is three primary things. First, a process for building oral content with with AI. We'll show you how to use evaluation criteria, that's built around building blocks to build, slide content, speaker notes, and, how to do messaging faster. And, we're gonna give you the ability to spend your time on practice instead of slides. The second area we want you to walk away with is, the ability to to generate AI powered evaluation evaluator, q q and a simulation. And we'll show you ways to generate realistic kind of nightmare type questions and how to record practice sessions and use AI as a tireless evaluator that helps your team get sharper every single day. And the third takeaway is, sample prompts and practical methods you can use on your next orals competition and, an honest conversation about where AI helps and where it might hurt you, if you're not careful. So from there, let's dive in, and let's start with, No. Real quick, Neil. If. I could just and one thing I, forgot to mention in my opening is that, please, if you think of any questions as we go throughout this, please, don't hesitate to use the q and a. We'll be monitoring that throughout our presentation, so if any questions pop up and they're relevant to the slide we're talking to, we'll throw those in live, and then we'll we'll have q and a time at the end. So, yeah, just encourage everyone, to to use that q and a feature. So sorry, Neil. Go ahead. no no problem. So I I like to tell stories. And if if you've been part of a down select proposal lately, this is gonna sound really familiar to you. I was on a team recently, on a a major federal opportunity, and FAANUS one was a past performance evaluation. And we spent weeks on it. You're probably familiar with the experience, pulling the right narratives, lining up the metrics, and getting the references nailed down. Right? And it was hard work, and and we did it well. And, then after we did all that hard work, we kinda waited a bit. Right? We didn't know when the down select notification was coming, and we didn't know if we were gonna make the cut. So the team, eased off a bit. Right? Everyone went back to their day jobs. The proposal efforts slowed way down. Then we got some good news. The email came and we made the down select, so we had about two weeks to put together an oral's presentation. And you can probably guess what happened then. Most of those two weeks went into building the deck, slide content, graphics, refining the message. You know, it was real work, but it ate up the calendar. And by the time we got to rehearsals, we had two days. Maybe it was three. And, q and a practice, we squeezed in one session. Just one. And, the presentation went well. You know, it was it was okay. And the slides definitely looked clean, and the q and a was just a little bit wobbly. And, you know, it was good. It was acceptable, and we lost. Right? So here's the thing. It doesn't have to be this way. And, hopefully, today, we're gonna give you ways to do a lot better. Blake, how about you tell us a little bit about the the the market you see out in the field? Yeah. Yeah. Sure. So when we're talking with companies, you know, as I mentioned, we're well connected with many companies in this industry and talking with leaders on how they're using AI and how they're, approaching the growth life cycle. And I would say most companies at this stage, they have a tool implemented. They've they've moved forward with an AI tool for their growth, but they've implemented it to varying degrees, I I would say. And so, most of the companies I speak with, you know, use one of those tools, but maybe they're using it to respond to RFIs or they're using it to pull in past performance information so it's quicker to gather that. But it's not too heavily integrated in the growth life cycle and not too much being used in preparing for orals, which is why we thought this topic was relevant today. But you can't really blame companies for not being fully embracing or, fully implementing AI because if you think about it, I mean, ChatGPT was rolled out in late twenty twenty two. And it you know, at that time, it wasn't very good and there were still concerns around, well, if I give my data to the model, like, will they then use it and could other people access my data. Right? So it wasn't really until 2024 where these models were being able to be rolled out at the enterprise. And so here we are in, it's almost halfway through 2026, which is crazy to believe. But, you know, if you look at other technology advancements in the past, the cycle is much, much slower. And so, you know, companies are still getting these tools, embracing these tools, figuring out how to implement it into their workflows. And but one one thing I wanna mention about the tools is that, you know, our position at Red Team is is not so much specific to individual tools. You know, there's many good tools out there for, AI and and government contracting, but our our thought is more on, the process and how can these tools be used most effectively to increase your p win? To avoid the scenario, Neil, you know, Neil just walked through a, you know, not a great scenario. But if that company used AI to prepare the slides, to allow more time for practice, there might have been a different end result. And so the strategy is how do you implement these AI tools thoughtfully and in a, in a in a approach that will actually increase your p win, save time, and allow everyone to focus on the higher value tasks. So that's that's really what we're seeing in the market now and hopefully what we share today will, allow everyone to to use these AI tools a little bit more, moving forward. Thanks, Blake. So the problem we we wanna, you know, put in put in front of you is is that, most teams spend about 80% of their prep time building slides and maybe about 20% on rehearsal. Excuse me. I'm getting over a cold here. Percent on rehearsal and q and a practice. And by the time they get to the work, the work that actually wins, which is, you know, fine tuning the message, practicing under pressure, preparing for tough evaluator questions. They're really out of time, and they're exhausted. And they walk into the room hoping for the best, instead of knowing that they're ready. So what AI makes possible, and I'm sure, most of us in the room know this, is compressing that content building phase. And the goal is not to automate orals, although there are a fair share of people think that you can push the easy button. It's to get to the high value human work sooner, in my opinion. If you can build your slide content in days instead of weeks, you now have weeks of rehearsal and q and a practice. That changes the outcome. And we're gonna show you specific methods on how to do that. I do wanna mention because, Blake and I get asked about this all the time. We have a proven AI enabled repeatable proposal process that can produce a submission ready complex proposal fast, And it's working very well and we've adopted the methods that we use in that for our orals, preparation process. And so anyone who's interested in that process can, talk to us about that. But we wanna stay very focused on the orals preparation process in this proposal. So, with that with that said, we wanna get specific now on how AI helps with with orals, and we wanna move into the details. There excuse me. There are three areas where AI can really help. The first area is in message clarity and slide content. AI can generate slide content that is anchored to the evaluation criteria. And I really wanna make sure that that that's emphasized, evaluation criteria specific, not generic content that could be generated by anyone that has an AI tool. Each slide gets one clear claim with mapped proof behind it. Your key discriminators get placed at high visibility moments. You're opening your transitions, your close, and he gets speaker note frameworks that give presenters the substance they need without scripting them word for word. The key here is that AI accelerates the content generation. You still make the strategic decisions about what claims and what to emphasize. The AI assembles the content from those decisions, and you'll see what I mean in just in just a second. The second area is where it gets really interesting, and this is newer territory. You can record a practice session, full oral rehearsal, and transcribe it. Then you feed that transcript into your AI tool along with the evaluation criteria criteria and your scoring guide. What happens is the AI acts as part of your evaluation board. It reviews your session as if it were the government evaluator, and it can identify where your messaging was unclear, where you make claims without backing them up with evidence, where your timing was off, where transitions between presenters were rough. And here's the thing that makes this powerful. It won't get tired like humans do. Each presenter on your team could do an hour a day of practice with AI asking them questions and evaluating their answers. That's a volume of rehearsal with feedback that is simply not possible with human reviewers alone. Your human reviewers get fatigued by the third run through. The AI just doesn't do that. You can also practice q and a verbally. Transcribe your answers and have the AI evaluate whether you actually answered the question, which frequently people don't, and whether you provided proof and whether you you stopped when you should have. So that's the that's the second area. The third area is whether you're ready for q and a, and you can simulate that really effectively. And this is where AI adds maybe the most obvious value. You can generate 20 or more nightmare questions from your actual proposal content. Not generic questions like tell me your management approach or something equally obvious, the questions that target your specific proof gaps, your staffing inconsistencies, and your transition assumptions. Transition meaning, you know, like, if you're, transitioning from another subcontractor. You practice answering them using a structured cadence. We'll show you in just a few minutes, and the AI never gets tired of asking you hard questions. It doesn't soften its feedback. Because it's your third rehearsal of the day. So those are three three of many areas that AI can really make the difference. And here's what changes when you bring AI into oral, prep. The blank page goes away. AI doesn't decide what to claim necessarily, and it doesn't replace your judgment. Humans still make every strategy call and every messaging decision, but the work of turning a decision into a draft slide, a draft answer, or a draft framework happens in minutes, minutes instead of days. And your team doesn't sit down to an empty deck and a ticking clock. They sit down to a draft, really like a red team quality draft they can already react to. And that changes where the team's hours go. Think back to the team in my story, the one that spent two weeks building the deck and squeezed in one q and a session. With AI handling the production work, those hours move. They move to the work that actually matters, sharpening the message, iterating on the slides, etcetera. Right? And this is the part I want you to take with you. The work that orals that wins the orals is the refinement, and, you know, I don't wanna belabor this too much, but this is where the win happens. So, maybe, Blake, you can talk to us about, FEDSIMSA. Yeah. For sure. And so before Neil's gonna dive into the process in a little more detail after these slides. But, before we get there, I wanted to spend a few minutes on FedSim because FedSim style orals, which is now GSA assisted acquisition services or AAS, runs those procurements. But FedSim is its own unique animal in terms of how those oral presentations go and what the orals slide deck needs to look like. And this isn't a FEDSIM training, but just to make sure we're all on the same page, just wanna level set on what FEDSIM is, that it's that delivery model run by GSA AAS. Most of the opportunities are through, the Department of War, but there's also civilian customers that do FEDSIM as well. Orals are mandatory. Every FEDSIM deal has orals presentations, and your technical response is often your actual oral presentation. And so a lot of the orals decks for FEDSIM are is very dense in nature and has a lot of more content on it than your typical oral slides would have. Your slides are also locked in at the time of the submission, which means as you're rehearsing, you won't have time to update the content. And then also there's a q and a session as well where there's thirty minutes of government caucus and then there's, all the questions are delivered at once. And the reason why the government, likes using AAS or FEDSIM model is given that a very high rate in their ability to win protests. So it's almost bulletproof in that regard. Now how could you use AI you in preparation for FedSim? You know, there's still the a lot of the, topics that Neil just went through still apply for FedSim, but it it shifts a bit. And so the fact that your technical volume is your actual orals presentation, your decks need to be more dense in nature, and that's something you would need to prompt your AI tool to let it know, hey. These slides need to cover all these requirements. We need to address it in writing in addition to speaking in the presentation, and we might wanna give it one of your prior FEDSIM decks if you've done a FEDSIM deal before. So it has a reference of how much content should actually be on there, what it should look and feel like. And so prompt adding those details to your prompts in your AI tools, will help, prepare for the actual, when you when you iterate through to prepare the deck for the presentation. Your presenters need to be key personnel on the contract, for FedSim. And so that means, you know, executives are allowed, but sometimes it disrupts the synergy of the group. And so you're only allowed your key personnel. And if someone backs out, then a different key personnel could cover for that person. But you have to keep that in mind as you're rehearsing that, you know, you can't swap out with anyone else at the company. It needs to be one of those key personnel. And then also the thirty minute caucus, the government will deliver all their questions at once. And so you can use your AI tools in the rehearsals to help anticipate what questions might be asked. And I would also add that if you know who might be in the evaluation board, you can tell that to your AI tool, build a persona of that individual. So if that that person has spoken at different events and you know what really makes them tick or things that are their pain points, or maybe you have a capture deck and you build a profile on this customer and you know they're gonna be on the evaluation board, put that into your present into your AI tool as well so that as you're running through the rehearsals, the evaluators will review that. Your AI evaluators will review that from that lens, and it'll really help to with you getting better feedback when you're going through the rehearsals. There is very strict scoring in FEDSIM as well. You really have to speak to what's on the slides. One thing that's unique to FEDSIM that, probably. is also similar to the rest of the government as well, but, AAS really spells this out, that they read your slides before you present, and then they'll also reference your slides after you present as part of the evaluation process. So when you're presenting to them, they're actually sitting there listening, actively listening just to what you're saying. They they already read the slides, so they don't need to read the slides as you're as you're speaking. So that means it's even more important for your team to get the rehearsal down, you know, make sure the synergy is aligned among the group. And that's where, you know, the AI tools can help, and we'll cover this in more detail later. But when you're rehearsing, it's always good to record the rehearsal and feed in that transcript and into your AI tools and then get the feedback from the government or just get the feedback from in terms of presentation. Did you make, you know, did you say, too much like I just did? Right? It'll pick up those filler words and the natural transition. Did you transition from one person to the next well? So it'll help there, but what it won't help with is the actual team interaction. How do you interact with your other colleagues? What is your body language like? Are you making good eye contact? Those are things that the AI tools aren't gonna be able to pick up on, so it's still very important that you have an expert orals coach or a person in the room who's helping to pick up pick up on those nuances. So we wanna be upfront on where AI can help in orals prep preparation versus where there are still some limitations. And then also the ninety minute hard stop, that does impact, quite a bit in terms of your rehearsals, and you gotta really be within that time frame. If if you're at ninety minutes, the government will just cut you off in Fed Sim. And so you have to make sure you're within that ninety minute cutoff. And so when you're rehearsing, you can make sure your timing, your your cadence is right. And if there are issues with timing, that's where you could use the AI tools to say, hey. Where should we trim out we we ran hundred two minutes on this rehearsal. Where can we trim out twelve minutes? Where what sections should we pull that from? And then it'll give you individual recommendations to each presenter, to each slide, and then each presenter can then rehearse individually with that feedback, and then you can come back together as a group and rehearse again. So those are just some FedSim specific nuances that I wanted to cover just because FedSim and GSA AAS is is a bit different in terms of how they compete and how you deliver orals to them. Great. Okay. So I hate going to AI presentations where, it's all overview. So here comes the me. We really wanted to give you something where you had, method and not just, what I would call overview. So, you know, please ask questions if this isn't clear. We won't clear. We wanted to make this high enough level that it could be a general presentation, but meaty enough that, you were getting some content. So don't hesitate to ask questions if you don't understand what we're talking about. Our our method, which is similar to what we do in proposals, is to use a building block method. Very different than just writing a single prompt that generates a a proposal in a whole, you know, one fell swoop. So it's letting the AI act as kind of a configuration manager, maybe even a little bit like the proposal manager, the orals, manager, maybe even a little bit like an orals coach as you develop this, and it operates very fast. So this is you know, you work through this portion, what's seen on this slide in about a day, maybe half day to a day. And you build these building blocks, each of these rows as separate documents or separate elements or separate parts of a single document depending on how you wanna do it. And you look at the pieces each time they're built, and you can edit them and have input to them and interact through the chat interface or whatever your tool, permits. And the idea is that there's three levels here, all very much based on your evaluation and scoring and the instructions for what your orals are. And, you know, the three levels here are you're basically gonna answer all of the questions that you're being asked, particularly based from the evaluation criteria. You're gonna make sure that all of the claims you're making in that, a value valuation based scoring, there's proof behind it. And that after you've done those two things, you're gonna make sure that you've, differentiated yourself, in those two elements. And those are three of the main layers that you put in place before you do everything else that goes on your slides. And there's other things that go in the slides, but those are the top first three things that you do. Each of them, a a single pass that the AI does giving you an artifact. And scoring answers become headlines and key messages on the slides. Your proof becomes evidence in your speaker notes or maybe some of the content on your slides. And the differentiators become high visibility placement, in your slides. And you actually make a map of where you're gonna put them on which slides as as you go. And so each slide becomes anchored to an evaluation criteria. And so you you can tell that you're gonna score well just by reading the titles of each of your slides, and the AI has assisted you in doing this. And every claim you make is mapped to contract meant metric, a qualification, or a resource, so that you're not making empty claims. And the evaluator can answer why you, before you've done any of the other material that bogs down the presentation and gets you, then gets you distracted. So you have that piece done, you know, maybe in the first hour of of starting because you feed in all of your past information that you've worked in and generate these three layers and validate them maybe even in the first hour or so of working in this or the first half day. Now as you start building your slides, in concert with your AI tool, you iterate. I like to do it in an interview fashion where you say, you know, build me my first slide. Why did you choose these things? How did you choose these things? What's my central graphic? And then why did you pick that? How do you make this better? And then I do the next slide and the next slide and the next slide. You can do this very quickly and get your draft set of slides in a day, maybe two days, depending on the complexity of what you're dealing with. And now you have your first set of drafts, and you can package that together and then start iterating on both with your people and with the tool saying, what are the weaknesses of this? How does this score against the criteria? And continue to refine and iterate as as you go. Before I show you what a sample prompt might be to get you here, you know, if, you have a question go ahead. One, one question that popped up is, related to the evidence in the speaker notes. And this person was asking, what are your thoughts on using speaker notes versus using speaker scripts versus using nothing. So it's I guess it's three different lines. Either notes, maybe this is bullet points, things like that, or actual scripts, or just try you know, no notes. Just off the cuff trying to present without anything to reference to. So at this first part, I'm not at the speaker note part. I'm just trying to get slide content. But the next the next step after that when when I'm getting to notes, my, my recommendation is you go to what I call a speaker framework, which is a little bit more detailed than an outline. I, highly recommend, unless you're working with the most nervous of writers, never to generate a script and, encourage your your, your, speakers never to generate a script. Because once you generate that script, it's very hard to get the writers to not try to memorize the script or to, deviate from the script. If you start with an outline, they'll learn that there's two or three different ways to, go through that, that that framework well. And that's what you want is the flexibility on their feet to have multiple ways to talk through it so the material gets ingrained in their minds. So I like to, in concert with the with the AI and human intervention and, you know, discussing it, really having a conversation with the AI, generate a framework. So. really kind of a a middle hybrid of the three options I was given. Yeah. Yeah. And I would add, I like the the outline you mentioned, Neil. Normally when I'm presenting, I'll have bullet points that I that I can reference And also anchor words because if you get lost and maybe you're you lose your train of thought and you need to collect yourself very quickly, an anchor word that's just maybe it's, I don't know, it's thrown out an example, basketball. And you remind remind yourself, oh, yeah. I needed to talk about dribbling down the court and, doing a layup. That was my next, like, you know, next minute or so of of the talking point. So that those anchor words can help you just recollect yourself and remember where you need to walk through. And sometimes that can be on the slides, you know, if it's an in person, presentation, sometimes your slides are your anchor. But, I would just add that in there. But, yeah, I totally agree with everything Neil said as well. But what I wanna make is a as a really important point on this slide, which is easy to miss, is you have these these documents, these little building blocks, and these decisions that you've made, which the the AI can keep in either its memory or in these documents. So not all of the information for each of the slides is here. So if you change your strategy as you go through a little bit, the AI is gonna be able to adjust. And instead, you're really talking about your strategy upfront instead of looking at a particular graphic and saying this line's too thick or this word's misspelled, which is what most groups do for a predominant amount of the time, talking about things that aren't really gonna impact the win. What you wanna do particularly at the beginning about talking about things that are in going to impact your score and have major impact. And then as you get towards the end, that's when you talk about formatics and how the graphic looks and does the graphic tell the story. The AI actually forces you, can force you to do things the right way. And I think that's the biggest thing that the AI can do for you, not auto gen a slide for you. That actually is just doing it the bad way faster. So I put here sample prompts. These prompts aren't necessarily like, oh, copy down these prompts. These prompts are just kind of starting points for you. On the left, what you see is, you know, the framework of a a prompt where you attach your scoring answers that I just talked about, your proof inventory, your differentiator statements, and you tell the I the AI, for each evaluation criteria, generate a slide headline that mirrors the criterion language. Two to three supporting points with specific proof map to each, a separate note framework like we just talked about, and the presenter will, put it you know, that framework, they'll put their own words into it, and a transition line to the next slide. You know? And, you know, that's a good starting point for the beginning because you can kind of look at that and work on that and get that in place. And then in the next iteration, you can start putting more into the slide, the graphics, and, the other things that need to be there for our full fledged orals. On the right, this is kind of your q and a simulation prompt. You attach your technical approach, your staffing plan, your proof inventory, including the gaps and the evaluation criteria, and you tell the AI, act as the government evaluator who's read our proposal that's looking for weaknesses and generate 20 hard questions that target inconsistencies, challenge weak claims, and probe assumptions, and ask for specific examples of past failures. Right? And so, hopefully, this kind of, you know, framework kind of shows you where you wanna go. You don't wanna just say, you know, give me fifty, forty, slides based on these instructions. It that's just too that's just too high level. Right? And so the next part is rehearsal. Maybe, Blake, you can talk a little bit about that. Yeah. Yeah. I'll jump in here, Neil. So, really, you know, I I covered this a little bit on the FEDSIM slide, but, really, this is where we see AI as, as very impactful for preparing for your orals presentation. And there's a few tips or techniques that we think, you know, we would recommend companies implement as they're going through the process of preparing, and for their actual oral presentation. So the first thing is to record your practice session. So this is better if it's a virtual presentation. First of all, at Red Team, we always tell our clients you should rehearse how you're gonna present. And so if you're gonna present in person to the government, you should rehearse in person. If you're gonna present virtually to the government, you should rehearse virtually. And so it it's this is works better if it's a virtual presentation because when you record it, the transcripts that it outputs actually says who said what at what time. And if you rehearse in person, everyone kinda gets jumbled into the one device that's recording while you're in person. And so, the feedback might not be as specific to individuals. And there's some ways to work around that, but generally speaking, that's kind of how it works. But recording your session allows you to then have a transcript of exactly what was what was said, the cadence of how it was said, and then you can feed that into your AI tool. So that's the second step is feeding your transcript into the AI tool. And I'll normally download a JSON format because it's easier for the machines to read, And so I'll I'll feed that into the into the tool. And you'd also it would be good to set up a a rag as part of this process. And if you're not familiar with what a rag is, it's a retrieval augmented generation, and it's a technique that combines two things. It's basically the retrieval piece is it's retrieving information from your own data, your own dataset. And I saw Brooke asked a really great question, about if you're submitting slides at the same time of your proposal, do you use the capture plan to inform the scoring answers? And yes. Absolutely. And so that's what a a rag that's that's one of the pieces I would recommend to feed into your rag would be your capture plan. So your rag for preparing for orals is all information related to the opportunity. So that's the RFP. That's your actual oral slide deck, the rehearsal transcript that we just talked about, the evaluator profiles, so who you think is gonna be on the evaluation board, what's their role, their background, you know, feed in in maybe post from their LinkedIn or if they commented on things, their personality, their preferences. Right? All that type of information, maybe agency account plans. As Brooke alluded to, your capture plan, which includes more information on your win strategy, your win themes, your discriminators, your ghosting criteria, all that kind of stuff. So that's all in your rag. And then when you ask your AI to evaluate you, which is that third step, that's where it will look at your transcript and evaluate your actual presentation against all of that data in the rag. And that's where it's really powerful because it's not only looking at what you're saying against your slides, but it's also looking at what you're saying against some of the ghosting criteria you should have said or your win themes. And it can act as the evaluation board. What I would recommend doing is writing a prompt that says, hey. The evaluation board consists of these five people. Here's their personalities, and actually get feedback from each evaluator. And you can have them cross collaborate and argue with each other. You know, you could simulate a whole evaluation if you wanted to get real into it, with your AI. And these would all be AI agents or or personas, that would be acting as your evaluation board. But the the AI evaluation board would also understand the RFP and the evaluation criteria and what they're really looking for as part of the evaluation. And so that can be really powerful because that can tell you if, hey. Did the evaluators actually get your win themes? Did they did they hear it enough that it actually resonated with them? Did they get the message that you're trying to deliver? Right? And if not, how could you restructure your deck or how could you restructure the way you deliver your presentation so it clicks in their mind and they get it. And so, you know, the and the AI the AI evaluators can really be a good proxy for how the government evaluators will actually hear you and think about how they're gonna evaluate your presentation. So after you run through those three steps, you can then refine that with your team. So you take in all the feedback, you hand that off to your team, maybe you coach each person person individually, maybe each person then runs through this individually on their own sections, and then you can come back for group rehearsals and run through this whole process again. And, again, it can be really powerful. The one other tip I would add is that if you wanted to, what you could do is you could have someone else kind of, keying to the to the transcript as you're running through your presentation, keying up what slide you're on. So as you transition to the next slide, maybe someone else announces the slide verbally, maybe in a separate room or something so it's not distracting the presenters. But that just allows the slide to be, the the slide you're on to be noted in the transcript. So then the AI as it's evaluating and it's looking at your slides at the same time, it'll know what slide you're speaking to, which could, result in better feedback. So just just other things to to think about as you go through the rehearsals. Great. And so I'm gonna speed up a little bit just, for time. You know? And the AI can just definitely make sure you use the correct cadence when you're answering questions, and this is true of any, any oral presentation. You wanna make sure that the that there's the the MC that, frames the question, routes it to the right SME, delivers proof, and then stops. So that's true of any, any orals, and the the AI can definitely make sure that you're using the right, the right pattern. And, you know, we talked a little bit about graphics. You need to be a little bit careful with graphics. AI can certainly help. It's really good at brainstorming the right graphic. It's good at drafting a good graphic, and it's good at mock mocking up, certain graphics and critiquing your existing graphics. Right now, many of the tools fall a little bit short in coming up with, you know, production ready graphics. Sometimes it can hallucinate the details, and it it drifts a little bit when you give it brand standards. So, you know, who knows what's gonna happen in the next two or three weeks or two or three months or in the next year. But right now, you just have to be a little bit careful about what you do, with, you know, prime time graphics. So, we wanted to give you a little bit of where AI might hurt you in, graphics. Blake, did you wanna take this slide? Yeah. Yeah. So four ways the AI could could damage you or could be harmful. You know, the AI is not end all be all. Right? And there's obviously great benefits you can get from it, but some some areas where there is some caution. First thing is the fabricated specifics. So if, you know, if if sometimes AI will write something that's very compelling because it's telling you what you wanna hear. And then if that's not actually backed up or if that's not factual, sometimes it it can do that. It can put things in there. I actually was working on something the other day, and it wrote something that sounded great. And I asked this I asked it, this sounds fantastic. Can you just give me a source so I know this is accurate? And then it came back to me. Oh, actually, my like, basically, my bad. I I I that was that's not true. So it it corrected itself. Now if you're in a oral presentation and the government asks you where did you get this data and you don't know and it was made up, you're dead in the water. So you gotta make sure that there's no fabricated specifics in your presentation. The next thing is over scripted. You know, if you sound robotic, if you sound like you're reading off a script, that's not gonna be good. As I mentioned earlier, the FedSim specifically is looking for the team dynamic and how the presenters interact with each other. So if your team is cold and robotic or they just they didn't they didn't really mesh or it seems like the first time they met was right before they walked into the room, that's not gonna be good either. And so you wanna make sure you're avoiding that. Then there's a confidence trap. If you're if you do a few prompts and AI produces a great looking slide deck and it's polished and you might think with your team, oh, this is gonna be fantastic. Our slide deck looks so good. We're gonna be good to go here. That might harm you in terms of the slide deck might look good, but your if you don't practice and go through the rehearsals, your team might not be polished and it might not look good to the evaluators. And then the last thing is the judgment problem. You have to make sure you have your senior people in the room reviewing your presentation, looking for the team dynamics, the eye contact, that kind of stuff, but also reviewing your deck as well. Making sure that the orals deck is actually compelling and that it's, you know, hitting on the points you wanted to hit on. And so never underestimate. While the AI can definitely help accelerate some of these things and pull data more quickly, past performance references, things like that, you can't substitute for someone who's won multiple orals deals in the past and knows how these things are evaluated and knows what the government's looking for. You need that judgment in your process. So what I'd ask you to do, on your next orals competition is really three things. One, think in terms of some building blocks that you can manipulate. And when you iterate, the the AI uses those building blocks as input to making the the the drafts. If you think that way instead of just thinking about one large document, you will be amazed at how better your result, gets. I was amazed at how better the AI does both in making unique content and making improved content. Try it out on your next, oils competition. Use AI as the evaluator. It does a really good job of telling you here's the worst things in your deck, and here's a prioritized list of things working to work on. It helps you not, you know, work on commas and colors and things that aren't really going to help you win. Protect your rehearsal time. It's what makes a difference. The evaluators are people, and they want to see you comfortably working as as a team. Real quick, Neil. If you go back, someone asked a question. Sure. The term building blocks is unclear. Could you shed, a. little more light? I use building blocks as a word to say, like, many many queries. Like, when I said the three layers earlier on when I said, like, scoring questions and then differentiators and then proof points, I break those into artifacts that I can look at, and I call those building blocks so that that the, tool isn't constructing the final document all in one fell swoop. It's creating these building blocks or artifacts and, creating a document with those. So I can look at what it's using to then create the full document, and I can go in and edit them and think about them and add new ones in as I go through the process over maybe two weeks. So that when it iterates and creates the new document, it's iterating from these building blocks and and that, you know, not, you know, starting over each time or whatever, whatever last question I asked it. And so by having those building blocks, it it really improves the process as you go on over a long period of time. Does that clear things up, I hope? Yeah. Yeah. Excellent. Thanks, Neil. Great. Alright. Well, I'll jump in on some of the takeaways, so from today's presentation. So, first thing is process over tools. We did have a couple questions people asked about what tools we we recommend. At Red Team, our thought is that we, you know, we wanna help our clients with the process about how do you go about growth, how do you go about, you know, in this case, preparing for an orals presentation. And that, you know, you could use multiple different tools. I mean, there's many good AI tools out there, both GovCon specific or using the Frontier models in in different ways. So, you know, many of these tools, you prompt them in a very similar way, and that's why we had today's presentation be more on, the process versus a specific tool. And I would encourage companies out there to think that way as well before deciding on a tool or implementing something. Think, like, what does that process look like? How are we gonna get value out of that and, you know, evaluate it that way? The second thing is using AI to coach and evaluate. You know, we hit on that multiple times. I don't wanna belabor this point. There's tremendous value if you structure that process right. And then don't underestimate your own judgment. You know, this is something with when AI turns out content, I think I'm I'm guilty of this too. I'll read it at first blush and think, wow, that sounds amazing. That's, you know, it's a great let me just run with that. But then if I come back a couple hours later and I read it again, I kind of I I don't feel as good about it. I I kinda pick it apart. I'm more critical, and I'll iterate through it and come up with something better. And so some of that is, the judgment that comes through in your experience. And so, you know, make sure you don't underestimate that because, iterating through it as Neil talked about, having those building blocks and iterating through there using your judgment, matters a lot. And that's why I think when companies, you know, when AI first came out, there's question of, well, if everyone's using AI for their proposals or if everyone's using AI for their orals presentation, isn't everyone's presentation gonna look the same or every proposal's gonna look the same? You know, that's not the case if you follow the right process and, you know, put in like, I talked about the rag earlier, put in your own custom data, and prompt it the right ways to get the the right outputs. And then real quick on who we are, what we do, and then we'll go to q and a. So as I mentioned, you know, Neil and I are with Red Team Consulting, and we partnered with Deltek to present this webinar today. And Deltek, we've been customers of Deltek for over twenty years, huge fan of their products. And if you're not on the Deltek platform yet or don't have GovWind yet, I would definitely encourage you to to make that investment because it's definitely well worth it. But Red Team, what we do is we help companies grow and scale in the federal market, And our services are really we have our grow framework that we've rolled out, which is helping companies longer term overcome their bigger strategic, growth objectives, which might be breaking into new markets or accounts or turning services in the, into solutions or, helping to increase your enterprise value before selling. So that's where we bring a team of growth advisory and capture and proposals to help overcome that challenge. Then we also have our different lines of corporate strategy where we help a lot with buy side m and a and then business development, capture proposals, and we do a lot of trainings. I know this is more of a public, webinar training, but we do a lot of custom tailored trainings for our clients as well. So now we'll go to questions. I know and there's our contact info. If you have any other questions or wanna talk more about about this topic or red team in general, happy to, you know, have those conversations. But, Neil, I'll I'll throw an initial question to you about, there was a question on if and and Michelle asked this, and it's not a dumb question at all. She said this might be dumb questions. It's not. It's that's why I'm leading with it. It's a great question. So if slides should be developed as the proposal is also being developed, what content gets fed into your AI tool and how often do you upload the updated information? So I think the question is getting at if you're iterating through your proposal and trying to, you know, get a final proposal ready for submission, but at the same time you're iterating through and developing your orals deck, how how does that work together and how do you avoid, you know, potential issues there? Okay. This this is this is the the best question, to tell you the truth. And it's a it's an example of how AI really helps you. I when you're doing that, and why AI is so amazing is I I would treat the, orals deck as another volume in your proposal. And AI is just great at keeping continuity among your volumes. And, you know, we have a methodology very similar to what we talked about here in for, orals with proposals where we use these building blocks and cross volume continuity and an interview method with the tool and, you know, moving through the proposal systematically where on each turn, on each iteration, really, almost every day, we would be creating a draft of all of the volumes for review, both by the team and by the tool. And so every day, you'd have a new version of your slides at an at an increased level of maturity for review. And so the idea is review, iterative, review, iterate, review, iterate. With each day, of your cycle, the material is getting better. And you would ask questions of of your people and of the AI. And then you would, you know, you would work on improvements both manually and within the AI. And so it's a it's a very different way of of of working that is much, much better because the AI keeps track of a lot of your configuration management and a lot of the document, creation for you, until you get to the end when you might have to take it. You know, maybe the last two days or so, you might have to take it out. With the orals deck, you might have to put professional graphics in and stuff like that. But I hope that answers your your your question is is that it's just another volume, and you keep it in sync the same way you keep your volumes in sync. Excellent. And, yeah, a couple other questions because I know we need to leave a minute or two for Nora and Alex to jump back on. But, question here on what AI tools are you expecting us to be using for your tips and tricks? I think that was maybe on one of my slides. You know, there's a lot of different tools out there like I mentioned. I'll personally say that I use Claude quite a bit. We actually just rolled out an enterprise license. I know there could be some issues for people who work with Department of War, but, there's a lot of really good features there with, the way it just integrates with Microsoft Office suite and things of that nature. So, definitely really helpful there. And then someone else asked about, it seems like AI is a great tool to write a proposal or make a presentation. Can we prepare oops. Questions are popping around. Can we can we prepare a proposal for NASA or Department of War using AI will be taken negatively? Neil, do you have thoughts there? Well, I mean, you have to follow the, you know, instructions of the, you know, your your customer. But I I I think that, you know, you can't put into AI material that, is isn't allowed to go into it. But you should not be able to tell that your content was in an AI, you know, when it comes out. Mhmm. It so from that perspective, they shouldn't be able to know. But, Yeah. you know, you can't use it if you if it's if it's, not allowed. Yeah. And then, one last slide someone asked about the best or one last question. Someone asked the best AI tools for generating slides. I would say Claude is really good at that. What I do is at Red Team, we have a, Red Team brand guide, which is a PowerPoint deck that has all, you know, our beautiful logo and our fonts and all that stuff in it. And so I'll give that to Claude. It it has it actually in my rag. And I'll say, you know, when I ask it to prepare a slide deck for me or tailor a deck for me, it builds it with that template, that it that it already has. So that that's good that, honestly, ChatTBT had issues with doing that in the past, but maybe other GovCon specific AI tools would be good at that as Yeah. The field is changing so quickly. What works today changes tomorrow. You have to stay on top of things, and process is what's important. And what you gotta do is find tools that support your process. Excellent. And I know Alex and Nora are gonna jump back on here. And someone asked a question if gov, gov one will be coming out with a tool to help with this. So I don't know if you wanna take a stab at that, Nora or Alex, but I'll, I'll pass it over to you to wrap up. Well, first, thank you so much, Blake and Neil. That was awesome. I loved it. I learned a lot of new stuff, and, I might be giving you guys a call. We are Deltek announced, I think it was last year, that we're developing AI, embedded proposal tool. And so look for more information on that this year. We would love to if you're a Delta customer, contact your CSM if you're interested in learning a little bit more about that. So before we conclude today, we just got a quick reminder that you will receive a recording of today's webinar by email within twenty four hours. And, again, if we're not able to get to your all of your questions today, we'll be sure to follow-up with you online. Also, if you could please fill out a short survey that you'll see at the conclusion of this webinar, we'd love that. We'd greatly appreciate it. So thank you so much again for joining us today, and please visit deltek.com for more valuable Deltek events.