Artificial intelligence’s (AI) rapidly changing tides have presented opportunities for startup founders this year, but the constant evolution also brings the challenge of keeping pace. Next year, being ahead of the game will be critical for success.
We have heard repeatedly from investors that they are not looking to channel their money into more of the same and are specifically looking for strongly differentiated and defensible technologies. Only those who push the boundaries of functionality will be able to catch the early opportunities and outsmart the market. Unfortunately, the reverse of this is that we will see companies disappear.
The AI landscape is going to shift significantly again in 2025, propelled by new technologies and changing attitudes towards investments. My predictions for next year are industry movements that founders in the space should be aware of to secure their place at the front of the pack.
The rise of o1
OpenAI’s o1 model is going to have a transformative effect on the AI, data and tech scene. It will be one of the biggest disruptions to AI since its inception. Its capacity for deeper reasoning and ability to solve advanced problems is exciting, and the race that is sure to ensue as competitors grapple to develop comparable solutions will be the industry’s next great battleground. Currently, no one else has any kind of model that compares, and am I looking forward to seeing how the startup community grasps this opportunity.
o1 will open up new areas of AI application but I don’t think the full extent of its impact will be realised in the next 12 months as users get to grips with how to best apply its enhanced reasoning capabilities. I also don’t think that o1 is the end game in terms of large language models (LLMs) with deeper reasoning abilities. o1’s outputs are slow and the model comes at a high cost, so there is certainly space for others to innovate here.
Proving ROI in production
Over the past 12 months, companies have allocated budgets and resources for AI exploration. Next year, however, as enterprise AI moves from proof of concept (POC) to production, the return on the investments made so far is going to be more closely scrutinised. There will be little leeway for failing projects as snags that weren’t clear during prototyping come to light, so only solutions that can prove worth to businesses at scale will flourish.
To prove worth, AI providers will need to help organisations overcome data management issues which, along with privacy and data updates, is the most pressing issue when it comes to scaling AI. For enterprises to extract maximum value from an LLM, they need to learn how to harness constantly evolving private company data. The environments that businesses operate in are complex and rapidly changing, so models have to have the ability to contextualise, index and retrieve relevant data at operational speed. Because of this, as businesses look at better approaches to take LLMs into production, there will be a greater focus on Live AI in 2025 than ever before. This will not only overcome common limitations around data management, but will allow businesses to deal with the freshness of data in both structured and unstructured forms in smarter ways that lead to more powerful outcomes.
Prioritising use cases
As AI enters its next stage of evolution in 2025, there will be less hype around the power or capabilities of specific architectures, and the push will move towards finding techniques that best deliver. Businesses need to focus on locking in use cases, doing extensive persona work and considering the end user’s journey as they interact with the technology to maximise their investments in AI.
Banks, for example, deal with sensitive data hosted on-premise and it has taken years to find ways around firewalls so that external systems can interact. Now, there needs to be more movement in AI transformation. Businesses will therefore increasingly focus on identifying feasible use cases facilitated by the data they currently have and then work to expand this by developing governance policies which enable external data to be used securely.
Zuzanna Stamirowska is the CEO and Co-founder of Pathway, a leading AI company revolutionising data-driven decision-making through cutting-edge live data processing solutions.
The post Pathway CEO and co-founder predicts 2025 AI trends: Will your startup survive the shift? appeared first on Tech Funding News.