In 2025, it’s time to prove AI is worth the spend
After two years of excitement and experimentation, 2025 will be the year companies get serious about proving AI’s ROI. The days of flashy demos and vague promises are over. Instead, businesses will need to see hard evidence that AI delivers real results, whether that’s faster processes, cost savings, or something else tangible.
Budgets are growing, but so must results
While budgets for AI are growing – one in three large businesses are planning to invest at least $10 million in AI in the next year – so is the pressure to show that money is being spent wisely. Getting a grip on ROI from AI is about measuring both indirect strategic benefits and immediate productivity gains – and weighing them against the costs of ongoing investments.
What that ROI looks like will vary from business to business, but ROI models that pertain to AI should be flexible to account for both direct and indirect benefits of AI. Productivity gains and time saved through automation are fundamental benefits, but this might also look like tracking how much time employees are spending on creative and strategic tasks that involve deep thinking versus admin. Success will require looking beyond traditional ROI metrics to consider the broader strategic impact of AI initiatives while maintaining rigorous financial oversight.
From the perspective of AI startups, this means the ones that thrive next year will be the ones who can prove their impact to buyers and finance teams early, stepping up and showing that they’re creating real value, not just riding the AI hype wave. Companies that do focus on hype rather than results will fade fast as buyers become more cautious and sophisticated.
The Savile Row of AI
In 2024, AI models became faster and more efficient. In 2025, not only will AI be able to do more, but it’ll do it better and become more specialised, with tools designed to solve specific problems or tackle niche pain points. Not all AI solutions need to be based on massive models trying to do everything at once, instead it needs to be as tailored as possible towards what businesses are trying to achieve.
This shift from general-purpose AI models to specialised, task-focused AI will have significant implications for businesses. Specialised AI tools will go beyond just identifying issues – they will be able to solve them too. This will enhance collaboration between AI and experts within businesses and ensure that AI is designed to work alongside people, not just for them.
In legal tech, for example, we’re focused on making lawyers’ lives easier, not replacing them. By targeting clear challenges, like streamlining contract reviews or managing data breach responses, we can deliver solutions that truly make a difference.
The future of AI isn’t about being the biggest; it’s about being the most useful.
Overcoming hurdles in legaltech
Right now, legal AI accounts for just 1% of the $1 trillion global legal market – so there’s enormous room to grow. But the opportunity isn’t just with law firms; businesses of all types need AI to streamline their legal work.
The challenge for legaltech startups in 2025 will be proving their worth in a crowded market. It’s not about building tech for tech’s sake; it’s about creating solutions that genuinely help people do their jobs better. For legal teams, that means giving them tools to work smarter, not harder – securely, efficiently, and without cutting corners.
2025 is shaping up to be the year AI gets serious. The hype has been fun, but now it’s time to deliver.
