Some time ago, just mentioning the term “AI” would lead your idea to be associated with massive breakthroughs and a frenzy of excitement.
However, things have changed. In addition to the drying up of venture funding, a wave of disappointment has swept through the industry as numerous companies found themselves unable to fulfill the grandiose promises and expectations that once fueled their ambitions
For those who have tangible projects — combining both a solid tech value proposition and a viable business plan — this is good news. As the industry evolves towards practicality, AI-powered startups can still find fertile ground to flourish. More sectors are adopting AI-related innovations, so robust, scalable solutions that deliver measurable value can carve out an attractive niche for themselves.
With that being said, if you find yourself at the helm of an AI startup, it’s crucial to remain mindful of key considerations that can make or break your venture’s success. This entails not only identifying the essential actions to take but also recognizing the pitfalls to avoid.
Stop.
AI entrepreneurs need to carefully discern where not to invest their time and resources.
You might feel strongly tempted to develop your own AI models. In my experience, this will lead to financial losses, since building your models in-house requires sizable amounts of capital. Instead, investors propose a different approach, one that brings to mind every gold rush — historically, investing in shovels has been more lucrative than chasing the elusive gold.
The same recommendation goes for attempting to compete with established AI models like ChatGPT, Gemini or Claude. The rapid pace at which AI technology is advancing signifies that even the most cutting-edge startups can be swiftly surpassed by existing solutions, and building yet another interface atop ubiquitous AI platforms offers little differentiation.
Additionally, you might need to shift your pitch strategy. If you’re starting your pitch with the now ubiquitous phrase, “We are an AI startup,” is an immediate red flag — it is equivalent to stating that you created a vehicle with wheels. The focus of your pitch should not be on your AI integration, because everyone is doing it. For instance, in MediaTech, the first major market for AI applications, there is no task related to media or content generation that AI is not addressing yet.
It’s time to disrupt, not duplicate.
Continue.
In a landscape dominated by tech titans, seize opportunities to collaborate. Instead of reinventing the wheel, emerging startups can focus on leveraging the resources of tech giants to solve specific problems. Developing these niche-oriented models can help aspiring AI entrepreneurs carve a competitive edge and mitigate the risk of being outplayed by larger players, all while capitalizing on the infrastructure that is already in place.
When you’re developing your AI models tailored to specific verticals, but for tasks requiring significant computational resources and overall power, leverage the models already built by tech giants, and combine them with your own. This is most vividly manifested in the MediaTech industry, where LLMs from major companies have already solved many industrial, content-related tasks, and snatched a piece of the market share.
Perplexity is an extraordinary example of this. The company is redefining the search engine industry by harnessing the power of models like OpenAI’s GPT, and then building its proprietary LLMs on top. This showcases the high potential of symbiotic relationships between promising startups and tech giants, a collaborative model that helps emerging ventures be more competitive and drives faster progress in AI-related innovations.
Other successful cases include Jasper, an AI-powered brand copilot that helps creators boost their marketing activities and optimize their content, and Filmustage, a cutting-edge, highly-niche solution that helps filmmakers cut pre-production costs and make the process more efficient.
Harness the might for targeted solutions.
Keep doing.
Additional advice is to keep leveraging the aggregation of various models, and enabling seamless integration with external ones. This will enhance accuracy and relevance. Furthermore, by standardizing your interfaces, you can create a universal product that instantly taps into the best available solution.
Remember that AI is a very fast-paced industry, and there is a lot that we don’t know. Staying adaptable is key to maintaining a competitive edge and facilitating scalability. AI trends evolve swiftly, so being able to quickly pivot and shift gears in tune with what the market needs can make the difference between staying in business and growing or becoming obsolete.
Stay agile, adapt faster.
Final thoughts: Integrating AI into the physical realm
We invest in products, not just in technology. Thus, spotlight your startup’s value proposition. A product must solve a problem or improve a task significantly. Investors back a venture’s value proposition when they see a genuine problem, identified customers, and their willingness to pay. How the problem is solved is the investor’s fourth question — only after primary concerns are addressed.
The most promising AI sectors now are mediatech (as noted earlier), healthtech (with research, diagnostic breakthroughs, and physician support), and B2B SaaS (given the 60%+ AI adoption at work). These sectors not only promise rapid impact and growth but also offer lucrative avenues for revenue generation, making them prime targets for entrepreneurial endeavors and investment ventures alike.
Finally, as we ponder about the future, we must discuss the eventual integration of AI into the physical world. Consider the trajectory of ChatGPT—what lies ahead as it merges with a robot’s cognition? What groundbreaking advancements await in digital vision technology?
Today, when you launch a new AI platform or tool, envision its practical application in the physical world. Tomorrow, strive to seamlessly incorporate these innovations into the fabric of our physical environment—whether it be within the circuits of a robot’s brain, the lens of a video camera, or the ears of a microphone. This progressive integration will be a pivotal advancement in the AI journey, and bridge the gap from the digital realm to the physical world.
Once AI becomes increasingly interactive with our physical environment, it is also important to ensure that we consider ethical and safety aspects, as this helps us harness these innovations for the betterment of humanity.
Ultimately, one hopes that we will move from mass hysteria and phrases like “AI will enslave humanity and destroy us all” to “AI is excellent at handling our practical tasks, and we are not delegating to it the responsibility for the survival of humanity”.
