The Future of AI for Brands: Agents, Applications, and Amalgorithms

Brands are at "Peak AI"

I got an email from a sales rep recently for a new startup called [Redacted] AI. While I liked the company’s value proposition, what left a bad taste in my mouth was seeing AI within the brand's very name. Their business is certainly “AI powered,” but what product or service isn’t, in 2025? Is "AI" really that big of a selling point anymore?

Anyone who has utilized AI in any sense before LLMs made their debut knows that AI has been a thing for a long time already. But in the last 2 years, everyone has started talking about AI more and more as a differentiator (and SERP clickbait), even when it’s not. Now we’re at the point where “AI” is making its way into brands' names as a shiny shorthand, like the prefixes and suffices i, e, .com, and ly that came before it. Simply put, we’re at peak AI. And just like we saw with the Internet bubble, that euphoria is exactly how we should know that this shiny new thing has now become a commodity.
 

Is the AI bubble over?

When open-source DeepSeek burst onto the scene earlier this year, it proved that LLMs in general will soon be a commodity. As the price of entry falls, processing burdens and environmental risks appear reducible, and open source availability expands access, AI is assured to soon become an omnipresent, expected, and even necessary part of our digital world, just like we experienced with wireless Internet, video calling, and relentless two-factor authentication alerts on our phones. 
 

The question now is, how to differentiate brands in an AI-saturated world?

The best way is to recognize and solve the biggest pain points with base-level AI platforms right now. Let’s consider LLM tools like ChatGPT and Midjourney. They are great within their domains and gave us our first glimpses into what GPTs and GenAI could be. But, there are also inherent limitations that startups everywhere are betting end users don’t want to put up with forever. 

The first limitation, prompting, is all about overhauling how we input requests to GPTs. 

Yes, you can study and learn how to prompt a GenAI engine, but wouldn’t it be great if you didn’t have to? Just like no-code platforms are broadening access to coding for laypeople, so will newer AI applications that help structure and point-and-click-ify complex prompts. At Genuine, we’ve put on AI workshops for our own team and client teams in the past couple of years to educate employees on promptology, but in my opinion, removing the need for that level of training altogether will help teams be even more effective in the future. 

The other significant limitation brands are tackling is how to overhaul GPT outputs. 

There are thousands of GPTs people are building on top of ChatGPT, Perplexity, and other generalist engines precisely because they want to be able to customize their results into specific formats, or consistently know the engine will include and exclude certain contextual aspects of searches. We recently did exactly this to streamline landscape research for an insurance client who was looking to gather over 20 specific data points about each of hundreds of sponsorship opportunities throughout the country. 

And these are exactly the kinds of customizations that are giving way to dedicated startups and products (aka “applications”) – that are going to make up the next wave of what we’ll see with AI. 

What does the next phase of AI tech look like?

APPLICATIONS

Well, like we said, applications is the clearest “next gen” category. Similar to apps as we know them today, these are any SaaS products that use an LLM to perform specific tasks. They are not fully autonomous, but more like assistants or tools, and will be the most common form of AI many professionals will interact with in their day-to-day work. 

AMALGORITHMS

The next, and most interesting use case, are amalgorithms. These are any system that uses multiple algorithms, techniques, or models to create a more powerful system. This can include examples such as hybrid learning systems, federated learning systems, or multi-agent systems. Multi-agent systems are especially interesting, and we recently sat down with the folks at Make.com to learn more about how their solution does exactly this. 

With Make, you can integrate multiple AI agents to work together, acting autonomously and interdependently to solve complex problems and make decisions. A great additional layer that Make enables, too, is “human-in-the-loop” interactions – these are moments within the amalgorithm decision chain where key decisions can still be sent to a human for approval before further action is taken programmatically.

AGENTS 

Lastly, agents are another example of autonomous AI systems. Whether a self-driving car, chatbot, or trading bot, these systems perceive their environment, process information, and take action to achieve specific goals set by the user.  They operate continuously and can interact dynamically with their environment. 
 

So, there you have it - the future of AI entails systems with better packaging, multiple models working together, and more autonomous decision making. If you’re wondering what the implications of this are for your business, or you just want to talk more about how to navigate this brand-new market of models, let us know and we’d love to hear from you.