Collaborating with AI to Highlight Human Brilliance

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Suspendisse varius enim in eros elementum tristique. Duis cursus, mi quis viverra ornare, eros dolor interdum nulla, ut commodo diam libero vitae erat. Aenean faucibus nibh et justo cursus id rutrum lorem imperdiet. Nunc ut sem vitae risus tristique posuere.

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Suspendisse varius enim in eros elementum tristique. Duis cursus, mi quis viverra ornare, eros dolor interdum nulla, ut commodo diam libero vitae erat. Aenean faucibus nibh et justo cursus id rutrum lorem imperdiet. Nunc ut sem vitae risus tristique posuere.

At Neol, our guiding principles create the fabric of the interactions within our community. We fervently believe in the transformative power of small, impassioned groups, united by shared ideals and complementary skills, gathered around a compelling interest. This belief is also the cornerstone of Neol's purpose.

We recently spoke with Neol co-founder Akar Sumet, and Creative Leader Armağan Amcalar of MVPStrasse who built Neol’s AI Approach Builder. 

Being a fresh startup, our primary goal in 2023 was learning how we could build a product that would benefit both businesses and our community of creative leaders. We had a hypothesis that if we could leverage AI's human-level language skills, we could actually achieve that.

As we are a community powered platform, we understand that contextualizing very complex data is crucial to helping people build relationships that have a better chance of creating the desired impact. In utilizing AI constructively to highlight human ingenuity, we see a hopeful future.

Below is a condensed version of our conversation with Armağan and Akar, edited lightly for clarity. 

How did the idea of this project come about?

Akar Sumset
Two things come to mind. One is we are a community powered platform. So first what we did was build the community, and prioritize understanding them. Once we had the core community in place, we then started thinking about the mission of building trust at scale, which means contextualizing very complex data to help people build relationships fast. 

What we’ve been good at, and what we’re observing is that when we put together creative leaders with clients it clicks well, but we needed to find a way to do that without being in the room all the time. We can be in the room for 50 people, but not 500 or 5,000 creative leaders. We began working on this problem seriously during the Neol camp in July 2023 in order to learn by doing. We were asking ourselves, how can we solve this problem using AI? But if you go back to our investor decks, you will see that we’ve always seen AI as something that can make creativity much, much more valuable. We’re always looking for ways to use AI to show the value of creativity, of human brilliance.

How did you put it together?

Armağan Amcalar

This tool required a unique design. First of all, we've emphasized the user experience. We're not building it for the mass markets. It's a white glove service that humans normally do. It has to be super shiny, beautiful, and has to make you feel interested. We have an amazing design team that did a mind-bending version of coming up with a vision of what such a product could look like. 

We spent hundreds of hours just figuring out how it should respond. Thousands of conversations with the AI to really make sure that it's responding in the way that we expect. It’s been a big challenge for us to figure out things like, what is an industry to the client? What is a skill? What is a passion? Or even, how long should this conversation be? It's not a definitive science. You have to be lucky and you have to be cunning at the same time. So we put in a lot of hours to come up with something that really gives you an insight into our ecosystem of creative leaders.

Why an AI Approach Builder?

Armağan Amcalar
Our tool helps people clarify their minds by framing their challenges. There's a language which we use to frame a challenge and if you just know the problem and don't know what the solution space looks like, even from a language point of view, it will help you with that as well. It will help you identify different industries, skills and expertise that are necessary for what you want to accomplish.

Akar Sumset

This is a good example of how we are enhancing human brilliance, how we are not trying to replace human brilliance, but at the same time, remove friction and contextualize.

We are focusing on mapping you to the right person for your needs. With our creative leaders, and with creativity in general, we leave things to the right hands, which are the creative leaders to finalize the approach. 

We believe in small groups of people coming together with shared passions and complementary skills around something that's interesting to them. This is a very strong way to change the world. It's embedded in our purpose, it's embedded in our manifesto even. But as a platform our job is to do this at scale—building trust and connecting the right people with the right challenge. That's our mission, that's our job. 

Armağan Amcalar

The challenge here was never to find the solution. Clients are trying to find the right people, and we're here to help them find the right people. Sometimes they don't even know how to ask for the right person — understanding what questions they need to answer in order to find the right people for the job. This also goes both ways. Creative Leaders would be really unhappy if they were matched with random challenges that they're not interested in.

We take everything into account when we try to find the match, their passions, even down to their tone of voice. 

Akar Sumset

And this is not the ultimate product. I connect this to our way of working, which is learning by building. So we wanted to learn how this product would benefit everyone — clients and creative leaders. Making complex information comprehensible and building trust is very, very crucial. And we had a hypothesis that if we could leverage AI's human-level language skills, we can actually build that. In terms of success, we have seen that it is very, very helpful in making complex information comprehensible, in contextualizing. For example, it can immediately communicate this huge profile Arman has. 

This isn’t just a problem/solution tool for us—as a startup, we are trying to communicate what is meaningful and different about Neol. And for us this is a product solution, but even more it's a brand solution. How do we substantiate our brand which is about novelty and new ways of solving complex challenges.

Armağan Amcalar
It's a novelty in and of itself. It's like an invention.

The bigger picture is that we are trying to adapt to fluid talents and fluid needs. Classical agency models don't work because you cannot develop one-size-fits-all strategies anymore, because of the rapidly shifting needs of clients. Everything needs to be custom, unique, tailor made. That's also why AI can't really play as well—at least in the line of work that we do, working with highly experienced, extremely unique people.

For a really long time we've been trying to tackle questions like, “What makes a great team? How do you bring together a great group of people with different, but complementary skill sets to work on a given problem and give more perspective to it?” We thought about it for a very long time. And it wasn't easy at all.

Thinking hypothetically also doesn't solve any problems. It doesn't give you any data. It doesn't give you any actual real needs. You have to build stuff in order to walk in that path. And what we’ve built is an alpha version. I can say it surpassed my expectations and it also surpassed a lot of other people's expectations from an alpha prototype perspective.

It has a lot of room to improve, but that's mostly because we figured out what we really need to know about our creative leaders to be their voices. We were able to discover what we need to learn from them and how we need to represent them. We developed a lot of new strategies that we want to try out in the future to further contextualize their presentation based on different challenges.

But this had to start with building something and the first version we’ve built has been super helpful in identifying what our next steps will be. 

Akar Sumset

It's a very different paradigm we are operating on. We are not in the paradigm of answering questions.

Armağan Amcalar

We are in the paradigm of enriching the meaning or the depth of your search, your ask.

Akar Sumset

Right—it's adding context around your challenge, and at the same time, it adds context to who can help you.

Armağan Amcalar

We are bringing in some reasoning capability as well. We're not interested in coming up with an answer. We're interested in what our community can do based on an understanding of your challenge. So we're trying to understand your challenge first, and then trying to figure out what the community members can do individually for that challenge. And then it's up to you to decide who you want to work with. 

We’re always looking for ways to use AI to show the value of creativity, of human brilliance.

What have you learned from the initial use of it?

Akar Sumset

A large challenge is that the current system doesn't leave much room for feedback from users. You get a result and that's it. So we want to improve our feedback mechanism, both in terms of the results you get and also if you’re not clear about something like, why this person? Or the results simply not representing what you're expecting. Another challenge has been in managing the perception of this tool. It is very much still alpha level, something we are trying to learn by building.

Another challenge is an inherent problem of AI, which is that it’s so confident of itself. It gives you results, and you'll find sometimes, well, I feel like this is not a very good response or...I think maybe don't talk about me that way and this way. So how do we bring that across? This is generative AI at the end of today, and it is at odds with precision. It's a long-term challenge. We've been very careful not to make this an open-ended conversational tool, because it creates all sorts of problems.

People approach this tool with different intentions. They use this as a search engine, for instance, trying to search the entire network which is not possible just yet. People have different ways of using it, and they break it beautifully—when they break it and they complain, we learn what to do next.

There has been a lot written about AI vs human elements in our current and future world. How do you think humans and AI should work together? What is your ideal working relationship?

Armağan Amcalar

GPT on its own cannot do what we wanted to do, because we're trying to bring meaning and reasoning to a challenge. If we were just trying to come up with answers, then it would be okay because you could build patterns, you could give GPT a thousand examples, making the answers easier to predict. But we're not trying to do that—we're trying to add meaning and reasoning, which entails a lot of decision-making in it.

What we're doing is just like a brain that has multiple centers for different tasks, like vision, understanding language, auditory cortex and so on. The brain has these specialized sections in it, and this is how we had to build an AI application—it's not a single prompt to understand, like, “Hey GPT, paraphrase this in this nice form.” It doesn't work like that. We had to build three different Autonomous agents who have their own tasks, their own brains, and their own memories, which does not exist in GPT. GPT does not keep a memory. 

But we invented a way to build memory into AI, because they need to remember the entire conversation,all the previous steps that they tried, what worked and what did not. And they need to update their running memory. You need to have a snapshot of the most up to date information from your perspective, from your expertise point of view. So our agents have memories of their own—they also share knowledge with each other. 

For example, we have this “account manager” agent who is talking to the client. The account manager, just like a person in real life, keeps the full transcript of the chat, but only transmits portions of it to what we call a planner, which comes up with an understanding of your challenge—the definition, the industries, the skills required, the budget, the details. And that builds up an understanding of the entire challenge to go back into the community and search who could be the best fit. And that information is then fed back to the account manager, so that the account manager can present this to you, to the client.

This is like mimicking an actual team and their intelligence and their communication, their information exchange. And we have to build these agents in totally separate ways to have them think and apply reason to what they're doing. This is not a single prompt. It's the result of autonomous agents that are working in the background on their own timelines. And they take different amounts of time to process this information and come up with an answer. So they collaborate and then the “account manager” decides when it's time to talk back to the client and what to tell the client.

We had to build "intelligence" into it, I don't typically like saying intelligence because that doesn't really have a definition, but I can say memory and reasoning capabilities or incremental memory — like updating your current knowledge based on previous events. GPT does not learn. Even if you have a conversation with it, when you start a new conversation, all that information is gone. There is no methodology in it to learn about your project, what you're talking about. The only thing you can do is to keep all of that information in the context of a chat. But that's not organized information, that's just a bunch of text. Therefore, you cannot control what it actually will add up to, and you never know what your GPT agent will say. That's not good for us because we need reasoning. We need a more structured way of keeping, persisting, and transferring information. That's why these agents set their own memories. There are currently no architectures out there in the world that have this much complexity of retaining information and transferring information between autonomous agents. And we had to do this in a business use case. 

Doing creativity in a different way and helping people develop better questions, especially in our line of work. When you talk about equity, how do you make sure that you're not prioritizing your friends or prioritizing the people that you already know? It is quite impossible with the community that we have, it's impossible to really know get to know, in detail, what each of these creative leaders does best. So we are also on to something much bigger here. We're trying to make it super, super equal for everyone—to give them a voice to be heard and to build tools to best represent their skill sets and their expertise—to match them equally with any incoming client challenge. This is something humans cannot do because we're going to run out of time. We don't know all these people. We cannot talk to them. And information crunching is what AI is great at. 

A lot of people, when they first hear about AI, the first thing that comes to their minds is, oh, this is going to come with its own biases. In fact, we are trying to build an approach, an application with AI that will destroy human biases by creating more equity for everyone.

An example of that is the profiles that we have. Some people know how to optimize it in terms of SEO, adding all the tags and like keywords. If you do a basic search, you almost always get the same people, because they know they're marketers. But if you add AI to the equation, a creative leader who's not a marketer can still be highly visible, and you can still bring them to a level playing field. 

And this is at the core of our efforts with AI — we want to build this in a way to bring equity and diversity to this business of matching challenges to different creative leaders. It's not like who has the shiniest profile, but who has the most meaningful profile for the job.

Until now people have been using artificial intelligence to segment and classify people, like for credit companies or credit scores, etc. Eliminating the “bad seeds” and finding who is worthy of certain things. It's been a destructivist approach. What we're doing here is constructivist, where we try to bring out the best in everyone and then try to find good matches. If we use AI in a constructive way to bring out human ingenuity, I think that's where our best hope for the future lies. So I think this is a very stark difference between the past and the future. 

Have a timely challenge? Start developing a solution. 

Pouring your energy into answering the wrong question is a frustrating experience—and the returns of asking the right question are transformative.

We invite you to submit your most pressing challenges to our Creative Leader community at

You may also like

Do you like our stuff? Subscribe.

Expand more icon.

Five main reasons to sign-up for our newsletter

Let's chat.
8 Back Hill
Herbal House 5th Floor
London UK
Businesses & Organizations
Create an account
Creative Leaders
Apply to join
Freelance Talent
Join the network
Join the team
Open positions