Digital Horizons
Welcome to Digital Horizons, where today's strategies meet tomorrow's innovations in the ever-evolving world of digital marketing.
Dive deep with industry trailblazers and thought leaders as we uncover what's driving success right now and forecast the trends that will shape the digital landscape in the years to come.
In each episode, we'll decode the tactics, tools, and techniques that are transforming the way businesses connect with their audience, ensuring that you stay ahead of the curve.
From social media shifts to the newest SEO advancements, e-commerce innovations to content marketing insights, Digital Horizons is your essential guide to the future of online marketing.
Join us on this journey, and let's shape the future of digital, together.
Subscribe now, and never let the digital future catch you by surprise.
Digital Horizons
Why 95% of AI Projects Are Failing (and What’s Actually Working)
The AI revolution isn't playing out as expected. While tech headlines herald the transformative potential of artificial intelligence for business, a surprising reality is emerging: 95% of enterprise-level AI implementations have failed to deliver financial returns within six months, according to recent MIT research.
Meanwhile, small to medium enterprises are quietly succeeding where giants stumble. Between 30-40% of SMEs report achieving meaningful cost efficiencies from their AI investments. What explains this counterintuitive outcome? The answer lies in approach and execution. Large organizations typically pursue custom-built solutions requiring significant investment and organizational change, while smaller businesses leverage ready-made tools that solve specific problems immediately.
The most successful implementations aren't replacing creative functions but optimizing repetitive tasks—automating finance operations, streamlining data entry, and reducing time spent on menial work. This pattern suggests today's AI is best suited for process optimization rather than completely replacing specialized human roles. We've seen this reality check play out at major companies like Commonwealth Bank, where an AI call center implementation initially backfired, increasing call volumes and necessitating the rehiring of human staff.
Looking ahead, AI implementation is becoming increasingly accessible through platforms like N8N and Make.com, which create low-code environments where even non-technical users can build powerful automations. This democratization will likely accelerate adoption among smaller businesses that can quickly implement and iterate without massive investments. However, as businesses grow more dependent on these AI platforms, we should expect subscription costs to rise significantly—following the pattern of other essential business tools.
What repetitive processes in your business could benefit from AI automation today? The opportunity is clear: start small, focus on specific pain points, and leverage existing solutions rather than reinventing the wheel.
The Digital Horizons Podcast is hosted by:
James Walker - Managing Director WHD
Brian Hastings - Managing Director Nous
All right, welcome back to brand news.
SPEAKER_00:Welcome back.
SPEAKER_01:Actually, Digital Horizons, the brand news episode.
SPEAKER_00:That's right. A little bit rusty still. Today we uh obviously I've documented in one of our recent episodes around my journey within our agency to really focus and invest in AI implementation, which has seen so far some good results, but we've also had quite a few challenges in the process as well, because some of the things that we want to be able to do are just taken longer and it it's probably not as simple as what I was really hoping it would be when we set out and start. We've got working models, but they're not producing the output that is quite there. And so there's a lot more fine-tuning and looking at different angles and ways that we can implement it to get stronger results from what we've put into place. Um I think this ties in really well with the article that you're talking about at the moment.
SPEAKER_01:Yeah, absolutely. I think anyone who's tracking or looking at AI will have seen people doing posts and articles about this MIT research paper that has reviewed the AI implementations of a large number of enterprise level businesses in the US, specifically those who have done an AI implementation at scale to see what has the outcome been. And effectively, after six months, has there been an improvement of their on their PL, on their profit and loss? Have they seen a benefit financially from these AI implementations? And the headline that's grabbing everyone's attention is 95% of enterprise level businesses that have done a significant AI implementation have failed. A part of me when I first saw that was like, oh, maybe this is a you know bit of a bubble, maybe it's overpromising, maybe there's all this hope, like hope and expectation that it's just going to replace all of the hard work, and that's not actually happening with at the large end of town, so it won't funnel through to the smaller end of town. But looking at it, SMEs in Australia, when interviewed, I think there's a CBA report on 510 businesses, small to medium enterprises in in Australia, on their AI uptake. And there was something between 30 and 40 percent of businesses who have seen cost efficiencies realized at the smaller level of business. And I think this has got something to do with the approach to an AI implementation at enterprise level. You're gonna build your own thing.
SPEAKER_00:They've got the money to invest. Where I imagine the people that are buying the small to medium size are probably buying stuff off the shelf. Exactly. They're just getting the stuff that's being tested, works, they know, and it's it's probably not huge cost savings, but it's probably all time savings. But it's probably going, you know, we've we've just implemented an AI chatboard on our website, and we now have to spend less time servicing chats on the website. Or maybe we've implemented an AI voice caller who's dialing out and booking calls for or booking meetings for us or something like that, which has already been developed and a lot of tons of investments gone into creating these things, and they're just paying a monthly subscription fee to access it.
SPEAKER_01:And I think the advantage SMEs have is their nimbleness, their ability to change a past process pretty quickly and realize the benefit. Whereas the enterprise level businesses, they're trying to, maybe they're not even trying to change too much, but they're trying to commit to a wholesale change across an organization that has all of these barriers and blockers stopping it from happening. Maybe they'll eventually get there and find um improvements. But where we're seeing the improvement isn't so much in the generation of creative or the content. It's more in the optimization of repeated processes. It's back of house stuff, the automation of the finance function, the taking menial tasks and streamlining them so that people are spending less time on those menial tasks. It's not as much in the creation of the actual service or product itself. We've seen trials in voice AI customer service start to have its failures or dropbacks. I think we reported on CBA, I think it was, um, and their AI inbound call center. They laid everybody off and introduced an AI call center. The call volumes actually increased as a result of this in a recent sort of admission from CBA, and they're having to admit that they got to hire back some staff. So that's again another example of an enterprise-level implementation. I still think they will get there. Like the potential is too great. The opportunity is too high to get a customer support nailed by AI and not have to have hundreds and hundreds of people in a room working at a call center. But I think the early winners are going to be businesses like ours, small to medium enterprise, who can just take an off-the-shelf platform, leverage a chat GPT or um a generative AI tool, create some automations like you're doing, and remove some of that bloat and that pain in doing our day-to-day jobs, do more of the work that we need to get paid for and get paid for it.
SPEAKER_00:Yeah. The way that even I think a lot of platforms, even Google and even ChatGPT, setting up the way to be able to create AI agents and making it very accessible, platforms like NADN and make.com, they're making it's kind of like almost using Zapia. Like it's it's it's becoming more and more accessible to people without the big investment. So I think that the ROI on just spending a bit of time learning and implementing these things is going to have a massive impact for small business.
SPEAKER_01:Yeah, absolutely. And I think it's becoming more of a tool. Automations have been around for a long time, but the addition of some of the LLM capabilities to add some logic and summaries to that is going to make it easier for businesses to do this themselves. Yes, there'll be the opportunity for AI automation agencies, but I think as it gets easier, you're probably not going to need these external advisors for the simple stuff for your basic automations.
SPEAKER_00:I think so though. Then it's like kind of like then, well, Google's got easier and easier over the years, but people still need to get, I guess busy people are willing to pay for results. Right. And so I think that there's still a place for it. But I do believe, as you said, that there's going to be a yeah, they're they're making it so simple.
SPEAKER_01:Yeah, I think where advice and support will be needed is well, what should you automate and how? Yeah. Not necessarily the difficulty in clicking the buttons. Obviously, if you took it to that point, I'd rather just get someone to do it if they said they can do it. But yeah, I think the time required to do that output for even the third-party agency is going to shrink and shrink as these platforms get better connected.
SPEAKER_00:Yeah, I saw a big report that as well, or as businesses get more and more reliant on it, the platform costs are gonna go up. Yes. Just as just as G Suite. I mean, we spend$5 per user. I think I'm paying$25 because it's so ingrained in your day-to-day. And I think at the moment it's new, people are starting getting on board to it. But I think that these platforms like N8N and different ones, it's just gonna the ship price is gonna go through the roof.
