AIP Podcast EP 75 - Smarter Faster More Collaborative RFPs that Win with Trampoline

This episode’s guest, Edouard Reinach, Founder and CEO of Trampoline, joins host Anne to explore one of the most overlooked yet mission-critical challenges in enterprise operations, which is the chaos of RFP responses. Edouard shares his unexpected journey from teenage coder and global communicator to AI founder, shaped by curiosity, diverse international experiences, and a drive to solve real operational bottlenecks. He breaks down why RFPs are the ultimate stress test for AI and explains how Trampoline transforms a messy high high-stakes, and highly collaborative process into a streamlined workspace powered by advanced AI that reads complex RFPs, extracts requirements, engages subject matter experts, and drafts high-accuracy responses so teams can focus on strategy instead of manual effort. Edouard also addresses common concerns about generative AI by explaining how Trampoline eliminates inaccuracies through expert validation, continual knowledge capture, and a workflow that augments teams rather than replacing them while also creating structured data that can reveal capability gaps and market opportunities. 

Follow AIP Affiliate, Trampoline
Website: https://www.trampoline.ai/
LinkedIn: https://www.linkedin.com/company/trampolineai/

Follow AI Partnerships
Website: https://www.aipartnershipscorp.com/
LinkedIn: https://www.linkedin.com/company/aipartnershipscorp/
X: https://twitter.com/AIPartnerships

The AIP Podcast is hosted by Anne Cheng, on behalf of the AI Partnerships, a Railtown company

Transcript

SPEAKER_00 0:00

Girl Dad, whiskey collector, executive consultant to brands, Frenchman who left France at age 16, off to an adventure in Brazil. AI founder? Well, today we dive into the colorful world of Edward Reynak, the CEO and founder of Trampolin AI, an AI that solves for an almost boring and mundane problem, RFP submissions. I don't know about you, but more than a third of all government spending in the US is spent via public tender, which typically is issued through a document that frequently numbers in the hundreds of pages and gives you nothing more than two to six weeks to complete a response. Talk about crazy, but what's even more curious that is that the man behind trampoline AI, a solution that is built to attempt to solve it collaboratively. Well, Edward, thank you for being on our show.

SPEAKER_01 1:06

Thanks for having me, Anne. Very happy to be there.

SPEAKER_00 1:10

Thank you. Well, before we begin, tell us more about you and your backstory. And also, more importantly, what is one whiskey that everyone should try at least once in their lifetimes?

SPEAKER_01 1:24

Um it depends on your tastes, uh, but definitely a whiskey from Compass Box. If you want to taste a creative expression of whiskey, uh hedonism is a very popular one. Uh, if you really want to go for the core expression of whiskey, you go for a 21-year-old bottle of um Parliament from Glandronac. Uh, it's a really interesting expression of whiskey. Uh, and if you already know whiskey, try uh the very new distillery called Arnamercane and try one of their single casks. And they're one of the best on the market. So that's for whiskey. Uh, as for me, I grew up in France. Uh I started coding websites back in uh '98 because my dad had a Mac, and you cannot play video games on a Mac, so you're stuck basically fiddling around, and that's how I discovered coding. Uh I moved to Canada when I was 16, um, and I started making money coding websites that paid for my school, uh, college, basically university. And then I worked uh for a long time in communication and marketing for multiple agencies. I worked for the United Nations in Kenya. Um and eventually I started my own business. Uh, we're doing exceptionally interesting websites like web experiences for Microsoft, Curig, and other businesses like these. Uh, and then I became a head of growth where I mixed design coding skills and data analytics skills. Uh, so I grew a few businesses pretty fast uh during that time. Then I became an innovation consultant for a few years. So I was very interested in why more businesses cannot do what I do. Uh so I wanted to help them do what I do. Um, and then I went back to entrepreneurship at the be uh a bit after COVID started. Uh, a very good friend of mine uh wanted to get back on that. He was already an entrepreneur, and we always knew we wanted to work together, so we decided to do that. Uh and yeah, that that's how Trampolin uh got started along the way.

SPEAKER_00 3:32

That's amazing. But okay, we have to get serious right now because the world of RFPs is a really difficult one with most large-scale enterprises dedicating entire teams of hundreds of people literally to do one thing, to read RFPs, and then cross-functional teams working together just to respond to them. So, I mean, why did you decide to solve this challenge?

SPEAKER_01 4:00

Uh so basically RFPs are the perfect stress tests for AI in the real world. Um, because it's a very messy and noisy process. Uh, and I've worked with RFPs uh long enough to know that they're extremely painful. Um, and yeah, they're the ultimate example of what's a messy collaborative process. Um, and it's really something that I think most AI tools actually cannot handle well unless you're part of those businesses who are just churning templates, templated answers. But if you're part of the businesses who are taking RFPs super seriously and you really want to project that you've read it carefully and you have a very specific answer, um that's where like most tools just cannot help you. Um so yeah, like if you have those big teams, uh basically your Avengers trying to answer uh those RFPs, um, you'll find that even if they use ChatGPT, like a lot of the work is very manual. It's really like a lot of project management and engaging with internal stakeholders. Um, so for us, the tool we've created basically enables you to work with multiple experts uh at the same time. Um, make sure that everybody understands the contest, the context in which they have to do that work. Uh so you're not just reprocessing past RFPs, but you're also adding expert knowledge and expert brain power back into the tool and back into the proposal. Uh if accuracy is non-negotiable, uh that's also a problem with AI for many, many companies. So in our case, we found a way to make sure that things are always accurate in the proposal that you send. Um and yeah, at the end, like finding information in the business is not just a search query, like it is a strategic endeavor where you need to make sure that the information that you're using um works to that specific uh problem that the company uh has. So yeah, that's uh it's a big problem with huge stakes. And for us, it makes it make sense to say we have the technology to do that, and if it works for RFP, it's gonna work for a lot of other things.

SPEAKER_00 6:22

That's great. Um, you know, collaboration is key to getting RFPs over the line, right? But you know, with such little time allocated to responding quickly, collaboration is almost virtually impossible. How do you shift the dial for cross-functional teams in light of these timelines?

SPEAKER_01 6:43

So we basically flip the way you enter an RFP. Okay, so before AI, you're trying to template your answers as much as you can. With AI, you actually do kind of the opposite, right? So the way we work is that our the AI gives the teams a shared workspace, okay, where they can instantly get structured around the RFP work. So we have an AI that reads the document, highlights all the relevant parts, and turns them into action items. And then every of every single task gets the right subject matter experts and pull them in. Uh, and then we have that AI that drafts like 90% of the answer, right? So the experts arrive in our platform and they see the question, they see the context of the question, and they see a suggested answer. All they have to do is approve, correct, or actually start collaborating because they're like, oh my god, this is not a good answer. And that's interesting because if this is not a good answer, it's probably not because the AI has been smoking crack, it's probably because, well, you know, there is something the AI is either missing, or maybe this RFP is not the right fit, stuff like that. So it means that within one day, your whole team can focus on the parts where the AI is not able to help you so much. So the disagreements are the basically where you will win this RFP, is where you really need to get creative and think of that deal. Um, so yeah, like with Trampolin, you are focusing on the part that is highly collaborative on day two instead of two days before deliver, like pushing the proposal. So that that's where it becomes really interesting. Um, and also like the whole workspace is built so that you can work asynchronously. So you don't have to wait for a meeting to happen. The meeting is happening every time somebody uh logs in to Empolin. So yeah, like that's that's how we solve it, basically.

SPEAKER_00 8:44

That's great. One of the things that faces most large-scale enterprises and RFP responses is that many companies insist whether for ego or or or real you know survival, that they must win certain RFPs. So perhaps for financial reasons or reputational reasons to the company, but there is a risk to responding to RFPs. Um, and one of those large biggest risks of responding is that that some of the capabilities required to um perform the RFPs either are made up or actually a leap of faith, right? So, how do you deal with that?

SPEAKER_01 9:30

Um so I think it's fine to take on a customer if you know that you will have to figure a little bit of stuff along the way. I think that's how you grow your business, you grow your capabilities, uh your case study folio. Uh I think as long as you're not lying to yourself on your own capabilities and you have executive signing off, you know, on the proposal, it's I don't think it will hurt. That said, um, I think RFPs are generally made to buy the best product or service at the best price. So if you know your competition can respond to that RFP better than you can, it's actually wise to stay the fuck out of it. Like to just realize that, okay, maybe we're not the right fit and save your time for better opportunities. Um, the one thing I would say is that when you're using trampoline, what's interesting is since we've converted that those RFPs into a very structured database, it means that you can actually pipe that data into your systems and do very interesting analysis of you know what does the market require of you and where did you meet those capabilities or where were the gaps. So yeah.

SPEAKER_00 10:44

That's great. You know, there are plenty of companies out there that have declined the use of Gen AI related solutions for RFPs, citing obviously a lack of subject matter expertise and more so uh because Gen AI models sometimes respond with data that is either made up or not valid. So, how do you manage data quality and governance with your solution?

SPEAKER_01 11:11

Yeah, I think they are absolutely right to refuse to use an AI solution that can't provide accurate answers. Um especially in their domain of expertise, that's their reputation that's at stake. Um, we are made to address exactly that problem. Uh, because we think that our AI should not necessarily replace your experts. Uh we actually think it's a little bit ludicrous to think that. Uh, we think AI should augment them. So the idea is that we accelerate the whole process, but we don't try to automate it completely, which means that for every RFP, there is a new information capture from your experts. Okay. Um, and as I said, like it's a very, very fast process. So it we are still asking them less time uh than before, uh, but we always ask them a little bit of time. And every time they provide new information and new knowledge, they're basically contributing to an automated knowledge base. Okay, so that's how we fix the accuracy problem because the AI has a lot of context on finding past answers and understanding why this answer was formulated this way and why almost for the same question we formulated a very different answer. So the AI understands that. And then for the part of the governance, well, the entire system makes it so that you cannot put like you cannot write the final proposal if you didn't get your experts validating every answer with a click. Okay, so that's how you get an AI that works perfectly for your team, and that still saves you 80% of your time.

SPEAKER_00 12:53

That's brilliant. Well, wow, what a what a complete whirlwind. But sadly, Edward, our time has come to an end. You've got such a colorful personality, and with this wonderful challenge that you're really solving for, which seems almost mundane, and I surely wish you much success ahead. To our audience, once again, thank you for joining us. And remember that liking, share, following, or sharing this episode with someone who might benefit from its content will definitely help us keep going to bring you more interesting episodes with AI Founders globally. Once again, my guest was the colorful Edward Rainack, and my name is Ann Ching. On behalf of the AI Partnerships Corps, he's wishing our audience and all of you out there happy holidays.

SPEAKER_01 13:44

Thank you, Anne.