Where now for AI?

Yet, despite all the ceremony behind the cutting-edge innovations that are making the headlines, the reality is that implementation often lags behind promise.

As head of AI at Tangent, I lead a team that develops practical AI solutions tailored for enterprise-level organizations, and this hands-on experience – paired with my own geekery in my spare time – has given me an understanding of where AI is truly delivering value today and where its challenges still lie.

To make sense of where we’re headed, it’s helpful to break the AI landscape down into four key horizons.

Present and near-term

AI is already driving tangible efficiency gains. Text and code generation tools are boosting workforce productivity, from developers to product managers. At Tangent, for example, our PMs now generate complex data scripts rather than spending hours manually formatting spreadsheets – it’s a simple but effective use of this burgeoning tech.

AI is automating repetitive tasks, enabling software solutions to guide users through data collection, and lowering the technical barrier for non-developers to execute sophisticated functions. We’re seeing businesses integrate AI not just as an add-on but as a core enabler of productivity and decision-making.

Looking ahead, AI’s role will become more sophisticated. We expect a rise in agentic systems – something that was a hot topic at this year’s SXSW in Texas and has cropped up in multiple projects for clients old and new over the year. AI models capable of performing multi-step operations across different modalities (text, image, and more) appear on the near horizon.

Additionally, I expect that specialized AI assistants will become more prevalent. This sentiment was shared in a recent article in The Drum spotlighting the insights of AI practitioners from brands like Schneider Electric, Lenovo, and Buzzacott – so it’s not just me who is expecting this shift.

Rather than relying on broad, unreliable chatbots, businesses will develop highly targeted AI tools that excel in specific domains, from advanced design assistance to customer support. This shift will open new verticals where AI can be deployed with greater reliability and precision.

Mid- and long-term

For AI to progress beyond its current limitations, improvements in reasoning, reliability, and memory are critical. AI systems will need to move beyond pattern recognition and demonstrate deeper grounding in knowledge, allowing for more complex decision-making.

Another key focus will be enhancing user control. As AI assistants grow more capable, both enterprises and their customers will demand greater transparency and customization. This could drive a surge in open-source AI adoption, allowing organizations to tailor AI models to their specific needs.

While today’s innovations are impressive, the most transformative AI capabilities are still in development. These breakthroughs – whether in cognitive computing, fully autonomous agents, or AI-driven creativity, to name a few – will reshape industries in ways we can’t yet fully predict. However, these advancements hinge on foundational technologies that are still being explored, meaning their mainstream adoption remains on the distant horizon.

…and for now?

For organizations looking to navigate this evolving AI landscape, the key lies in striking a balance between immediate wins and long-term vision. Today, AI can already enhance workflows, automate repetitive processes, and create intelligent assistants. Soon, businesses that embrace agentic AI systems and specialized assistants will gain a competitive edge. And for those with a longer horizon, investing in trustworthy, user-controlled AI will be a strategic differentiator.

In my day-to-day, I’m increasingly in receipt of briefs that are centered around understanding what’s possible now, what’s just around the corner, and how best to prepare for the next wave of AI-driven transformation – meaning that among many brands there is a desire for future-ready AI. While no one truly owns the future of AI, those who can bridge the gap between innovation and real-world implementation will be the ones shaping it.