Is this the AI replacing marketing professionals?

MarketJar

Wed, February 25, 2026 at 11:19 AM GMT+9 9 min read

In this article:

GOAI

+11.53%

Key Takeaways

**EVA Live (Nasdaq:GOAI)** has launched NeuroServer, a purpose-built AI system trained specifically for digital advertising rather than built on off-the-shelf AI models.

Its custom neural architecture is designed for multi-objective optimisation across real campaign KPIs, and was trained using reinforcement learning in an agency environment, positioning it very differently from typical AI-powered ad-tech tools.

With global advertising exceeding $1 trillion annually, even a niche position within the U.S. market could support meaningful scale.

Under a measured 2031 scenario of $500m ad spend, 15% fees and 25% margins, the valuation implies a 73% upside from today’s $4.30 share price.

The key risk is execution: adoption, proof of outperformance and customer diversification will determine whether the narrative holds.

These days, it’s more common than not to see companies claim use of AI in their processes.

Breaking down how they do it, you’d most likely see this: company connects to the API of an off-the-shelf (like OpenAI’s GPTs) or open-source AI model, give it task specific frameworks and instructions for the company’s use case… and voila, the company now has a custom-made AI agent that performs 10x faster than human professionals.

And there’s no dismissing that idea – it’s a smart way of leveraging existing AI tech to make your company more efficient, without the costs of researching and building it from scratch.

But doesn’t that beg the question… what happens when a company in an industry, with all the knowledge of that industry, actually builds AI infrastructure (you guessed it, for that industry) from scratch, for the sole purpose of creating the best-of-the-best in that industry?

I think EVA Live’s newly released NeuroServer is exactly that.

Introducing EVA Live and its key platforms

Founded in 2016, EVA Live is a US-based company operating in the marketing software industry, offering marketing and advertising agencies platforms they can use to optimize ad campaigns for digital marketing. It has two key platforms:

EVA’s platform, wittingly called “**Eva Platform**”, uses AI to intelligently match clients’ ad campaigns to specific ad spots to optimize for the clients’ goals, whether the ad is for brand awareness, lead generation, new customers, more sales, etc., for multiple channels all at once. Eva Live goes into more detail on how that works here.

EVA also owns the “**Eva XML Platform**,” which is built on the concept of advertising arbitrage. The platform automates the creation of thousands of ads and optimizes ad spend allocation across channels, thanks to its integration with ad spaces on networks including Google, Microsoft and Meta.

 






Story Continues  

EVA Live has always used AI to power its platforms:

But that’s not what this narrative is about.

Last month, the company uplisted from the OTC markets to the NASDAQ, the same time it announced NeuroServer. And the tech behind NeuroServer is the catalyst we’re looking at today.

NeuroServer’s AI: Purpose-built for digital advertising

Digital advertising is fundamentally a sequential, multi-variable decision system spanning creative, media allocation, fraud control, and revenue optimization.

NeuroServer is powered by Eva AI, which was purpose-built to address this challenge through a custom neural architecture trained with accelerated reinforcement learning inside a fully simulated agency environment. Let’s break that down:

Eva represents a custom-built architecture rather than a generic AI model. Unlike AI agents built on off-the-shelf models, Eva was built in-house with a custom-design neural network focused on sequential decision making, multi-objective optimization and cross domain skill transfer – the “personality” traits EVA Live’s R&D team deemed to be most fit for purpose. This build allowed Eva to evaluate tradeoffs between creative decisions and conversion performance.

Eva was trained through accelerated reinforcement learning inside a fully simulated global advertising agency. Reinforcement Learning (RL) is a machine learning method where an AI learns by trial and error, improving its decisions based on what works and what doesn’t. Studies have proven that agents subjected to RL outperformed over off-the-shelf agents. A more targeted study on RL on an agent operating within an advertising environment showed stronger campaign conversion rates.

Eva’s training was fully targeted towards hitting real marketing KPIs. For each simulation, Eva was scored and reinforced based on performance across ROAS, CPA, RPC, conversion quality and retention, fraud probability, and budget efficiency; all key metrics digital advertising is measured on.

Eva is the finalized model produced after training reached performance convergence. This is the point where additional simulations no longer made statistically significant improvements. The final model consistently outperformed baseline human benchmarks.

What EVA Live created is essentially an AI agent with the perfect personality for digital advertising, trained specifically and only for advertising, until it perfected advertising.

The difference between NeuroServer and other AI-powered ad-techs

To further support my narrative of NeuroServer being unique, let’s also review key players in the ad-tech space who use AI.

**Google Performance Max**: AI-driven optimization across channels using machine learning

**Meta Advantage+**: automated campaign optimization using machine learning

**The Trade Desk (Koa/Kokai)**: AI bidding & decisioning layers using deep learning algorithms

**Skai**: uses generative AI to provide users insights and recommendations (AI-powered chatbot)

**Albert.ai**: cloud-based AI platform that manages and optimizes campaigns using machine learning

**Adext**: uses prebuilt machine learning algorithms to optimize campaigns

As you can see, most of the platforms in the digital ads industry use machine learning frameworks to enhance performance, but I have not found any other ad-tech platforms claiming a custom-built neural network like EVA Live does.

It looks to me that NeuroServer or Eva would be the first in the industry to have this specific build and level of reinforcement learning.

The market opportunity for NeuroServer

Global advertising revenue was ~$1.14 trillion in 2025, with analysts estimating this to grow by 7.1% for 2026. Over one-thirds of this ($431B) revenue was from the US, where EVA Live operates.

85% of brands and agencies are either using or planning to use AI in an advertising-related capacity.

92% of brands and agencies in the industry consider AI’s biggest opportunity to be in improving existing processes.

Valuation

We will use the agency-based revenue model for our calculations.

EVA Live currently records revenues through a principal-based model, where it receives its clients full ad spend or marketing budget and recognizes that as revenue.

However, EVA seems to plan to move towards an agency-based model, where it instead earns and records revenue by charging clients a platform fee, which would be a percentage of total ad spend.

Key assumptions:

**Ad spend processed of $500 million by 2031.** This is a significant jump from EVA Live’s LTM revenue of ~$15.6m (under the principal-based model) but assumes EVA Live captures a small but meaningful position within the U.S. digital ad market. $500 million in managed spend would represent a niche but credible footprint among mid-sized agencies.

**Assuming 15% of ad spend as fees.** Agency-based ad platforms typically charge 10-25%. I’ve selected a slightly lower-than-midpoint percentage for this computation.

**Assuming a 25% operating margin (EBIT margin), typical for SaaS. **NeuroServer just went live so we don’t have metrics on margins just yet. 25% tends to be the average number for SaaS companies.

**Using a 10% discount rate.** The current risk free rate is just over 4%, but I’m adding a higher risk premium given NeuroServer is new to the market.

**Price-to-EBIT multiple of 20x.** Which is a typical benchmark for growing software companies.

EVA Live appears undervalued if NeuroServer gains even moderate commercial traction. If the company processes $500 million in ad spend by 2031 and earns a 15% platform fee, that would translate to roughly $75 million in revenue and $18.75 million in EBIT at a 25% margin.

Applying a 20x earnings multiple implies a 2031 market cap of about $375 million, or approximately $11.96 per share based on 31.34 million shares outstanding.

Discounted back to today at a 10% rate, this suggests a present fair value of around $7.43 per share which implies a 73% upside, highlighting that even measured execution could justify meaningful upside from current levels.

Risks

While EVA Live’s story is undoubtedly exciting, key risks include:

**EVA Live fails to deliver on its vision for NeuroServer.** NeuroServer sounds exciting, but we are yet to see technical proof of its outperformance over similar AI-powered ad techs. The biggest risk of this narrative.

**Low product-market fit.** Even if NeuroServer was a technical success, there may be less demand from the market as expected.

**Customer concentration risk.** At the moment, EVA Live’s receives over 80% of its revenues from three customers. While EVA Live is working on commercializing NeuroServer, the loss of any of these major customers will materially impact EVA Live’s financials.

**Competitive risk.** Digital advertising is dominated by Big Tech such as Meta and Google, all of which have the budgets to invest heavily into AI. Similarly, there are many other smaller players in the industry.

The Ultra-Long Term Narrative

The thing with Eva AI that I want to highlight before ending this narrative… its neural network is profession-agnostic.

As I’d mentioned earlier, Eva’s neural network was custom-designed for sequential decision making, multi-objective optimization and cross domain skill transfer.

While Eva in this context was specifically trained for EVA Live’s key target market – digital advertising agencies and marketers – the base Eva (i.e. the one with none of the reinforcement learning for digital advertising) can actually be trained in any profession that requires decision making and has multiple performance metrics.

Yup, that means Eva can potentially be trained in areas like accounting, operations and logistics, and cybersecurity… a story perhaps for another day.

Conclusion

To me, EVA Live feels like an asymmetric opportunity. The market opportunity is large and the company only needs to secure a small, profitable foothold to create meaningful value. If NeuroServer proves it can deliver better results and agencies begin shifting budgets toward it, the revenue model could scale quickly. The investment case now really comes down to execution, which we’ll hopefully see early signals of within the next year.

Have feedback on this article? Concerned about the content? Get in touch with us directly._ Alternatively, email [email protected]_

Simply Wall St analyst MarketJar and Simply Wall St have no position in any of the companies mentioned. This article is general in nature. We provide commentary based on historical data and analyst forecasts only using an unbiased methodology and our articles are not intended to be financial advice. It does not constitute a recommendation to buy or sell any stock and does not take account of your objectives, or your financial situation. We aim to bring you long-term focused analysis driven by fundamental data. Note that our analysis may not factor in the latest price-sensitive company announcements or qualitative material.

Terms and Privacy Policy

Privacy Dashboard

More Info