Already, boardrooms are beginning to demand more than just novelty from AI.
As we settle into 2025, a new reality will dawn for chief technology officers (CTOs) and innovation leaders, says Alex Smith.
The days of experimental AI projects and playful applications are over. From now on, they will need to demonstrate tangible, bottom-line results from the AI investments made in 2024.
At the same time, many organisations that have mistakenly concluded that search is “yesterday’s news” in a world where AI seems to be getting all the buzz will rethink this position.
As the understanding of AI and search technologies advances in 2025, organisations will recognise AI and search are not competing technologies, but complementary tools, each with their own strengths and applications – and there are concrete steps they can take in each area to best leverage each one within their business.
From hype to impact: AI will need to prove its worth
Already, boardrooms across industries are beginning to demand more than just novelty from AI. They want to see AI drive efficiency, boost productivity, grow sales, enhance profitability, and open new revenue streams – at least in some combination.
As a result, CTOs will find themselves challenged to transform AI from a groovy toy – a chatbot, perhaps, or a tool that spits out text in iambic pentameter – into a powerful tool reshaping business operations.
The advice here is to find problems where people care; problems where money is involved, or client relationships are integral. From there, tackle simple, low-hanging fruit that will have an immediate impact. For example, maybe what your team or your client really wants or needs right now is an accurate summary of a contract. Start there; start small.
From there, organisations can aim “bigger”, but if AI is going to be game changing and truly impactful, they need to do the hard work of looking at process within their business and then figuring out how they can “break it down” and/or reprocess in a way they can do some big things with AI, such as completely changing the way they do business or deliver a service.
Teams are capable of organically reprocessing some elements of their workflows, but they need to bring in actual change management professionals to make the correct decisions and ensure those changes stick. You can have the world’s most successful trial project, but if no one changes their process accordingly and falls back to exactly what they did before then, then what was the point?
Once they’re able to do the small task, then, organisations need to figure out how to deprocess and reprocess, otherwise they’ve not really changed anything – and the AI is too expensive just to be a nice little add-on. Only AI initiatives demonstrably impactful will survive and thrive. Those CTOs who falter in this endeavour could well see their AI budgets slashed: a technology that isn’t game-changing may simply prove too expensive to sustain.
“Already, boardrooms across industries are beginning to demand more than just novelty from AI. They want to see AI drive efficiency, boost productivity, grow sales, enhance profitability, and open new revenue streams – at least in some combination”
“Search” will make a comeback in an AI-driven world
While organisations are taking a close look at how to make AI deliver true impact, they’ll also be getting reacquainted with a technology more associated with decades past: search.
If you surveyed a cross-section of professionals and asked them “what do you want AI to do?” a lot of them would likely respond: “Find stuff – and find it more effectively”.
For instance, whether you’re in Microsoft Copilot or some vendor’s AI-powered system, if you say, “Find every contract our senior partner has written to and then create a summary of everything they’ve written”, that is in essence a search. The end user can’t get AI to create a summary of a list of documents if AI can’t find those documents.
Software vendors will incorporate new developments such as semantic search and retrieval augmented generation in their solutions, coupled with nifty natural language capabilities (such as the ability to be more natural in interactions with a search engine rather than having to learn arcane keywords or commands). Together, these advancements help knowledge workers quickly and efficiently find the resources they need to do their best work and deliver the best outcomes.
Knowledge graphs – which underpin things like recommendation engines – also play an important role here. Ultimately, there are a lot of problems that can be solved by search.
What this highlights is the importance of the underlying information architecture (IA) when it comes to both search and AI. Unless data has been properly curated, managed, and tagged with appropriate metadata, it is virtually impossible to leverage it in any meaningful way. It’s very difficult to search this data and pull up relevant results if it’s primarily just an amorphous blob of data – and it’s even more difficult for AI to make any sense of it.
The guidance is clear: organisations hoping to see good results from search and AI can’t neglect the IA. Fortunately, AI can lend a hand here in helping to “clean up” massive amounts of data at scale by going through it and then classifying it appropriately. For example, “this is a services agreement,” “this is an employment contract,” and so on.
The better a handle we have on this foundational information architecture, the better positioned they’ll be to unlock the full potential of search and use it in conjunction with newer technologies like AI.
It’s time for a transition
CTOs and innovation leaders must transition from experimental AI projects to implementing solutions that deliver measurable results. By focusing on enhancing information architecture and integrating advanced search capabilities with AI, businesses can achieve greater efficiency, productivity, and profitability. The ability to effectively harness these technologies will help determine the ability to successfully innovate and deliver better business outcomes in the year ahead.