Adopting AI effectively

AI opens new doors for SMEs to compete. But, argues Mark Rotheram at BCN, complacency could close them just as quickly

 

AI is no longer the preserve of large, multinational enterprises with deep pockets. A decade ago, building even a moderately sophisticated AI tool would have cost in the region of half a million pounds or more. Today, the same capabilities are often far more powerful and can be delivered at a fraction of that price. Historically, this would mean that only the largest enterprises would reap the productivity rewards of AI, but now SMEs can compete directly.

 

The cost of AI implementations has fallen dramatically as the technology becomes more accessible than ever. For small and medium-sized enterprises (SMEs), this provides an opportunity to level the playing field with larger companies.

 

Ignoring the opportunity comes with major risks. For those companies that fail to innovate with AI, it’s a case of disrupt or be disrupted.

 

 

AI’s disruption potential

AI’s potential to disrupt business models is already being felt. For smaller companies, the real competitive threat isn’t just similar-sized companies with similar business models; it’s the hypothetical startup that could launch tomorrow, built entirely around AI from day one. Such a company would have no legacy processes or outdated technology investments slowing it down, and it could undercut its incumbents on speed and cost.

 

Small and medium-sized enterprises need to act quickly and strategically to become a disruptor, rather than the disrupted. To keep up with future competition, there is a framework we use with four AI maturity levels to help SMEs gauge where they are now and where they need to go.

 

Level one: Experiment. Encourage employees to play around with AI tools such as the free AI copilots included in Microsoft 365 to safely explore AI’s capabilities within a secure ‘bubble’ rather than using random online AI services. This will help avoid data leaks.

 

Level two: Augment. Many SMEs are now licensing tools like Microsoft 365 Copilot to boost worker efficiency. The key here is to target the right use cases to those who will get the most value.

 

Level three: Automate. Identify a workflow that’s draining time or resources, that’s often something that’s repetitive or data-intensive; then create a bespoke AI or automation solution to handle it.

 

Level four: Reimagine. This level has the most profound impact. For businesses whose value is mostly in advice and consultancy, AI can be extremely disruptive. Letting AI handle the heavy cognitive lifting can reinvent how a business delivers value.

 

Many SMEs start by experimenting, then augmenting existing work as confidence grows. The reimagine stage is the most transformative - and the most disruptive - but ignoring it entirely could leave a business exposed to faster, leaner AI-native competitors.

 

 

Overcoming the obstacles

Two major barriers tend to hold smaller organisations back from AI adoption: data readiness and internal skills.

 

Data readiness means having well-documented processes and clean, consistent data. Many SMEs struggle to automate effectively due to information scattered across spreadsheets or a lack of clear process documentation. Without a solid foundation of data, even the most sophisticated AI will struggle to deliver value.

 

Skills gaps are equally common. Business leaders may be intrigued by AI but unsure how to apply it, while employees may lack familiarity with the tools and the speed at which they have evolved. Education and guidance can bridge this gap. Often, all it takes is a demonstration, such as AI summarising a complex report or drafting a customer email, to trigger a lightbulb moment that reveals AI’s potential.

 

 

Real-world wins

AI isn’t just a theoretical advantage; it’s delivering measurable gains in real organisations. One SME automated the management of a shared customer service inbox. Instead of staff manually triaging and responding to each enquiry, complaint, or request, an AI tool now scans incoming messages, gauges sentiment, flags urgent cases, and drafts tailored replies. Human agents review and send, meaning that hours are saved every week, which they can redirect to other higher-value tasks without the customer losing the personal touch.

 

In another example, a financial reporting business slashed the time taken to draft bespoke financial reports. What previously took analysts several days can now be completed in minutes. The AI system pulls the relevant data, drafts the content, and adjusts the formatting, leaving the analyst to check the details and finalise the report. The return on investment was delivered in a matter of weeks.

 

These stories demonstrate that you don’t need cutting-edge AI research labs to see transformative results – you just need to target the right processes.

 

 

Avoiding common pitfalls

AI is powerful, but if it’s not used correctly, things can go wrong. Three things can undermine projects. First, investing in an AI tool without investing in training staff or integrating the tool into workflows can be a costly mistake if it’s underutilised. AI’s capabilities are evolving monthly, so regular training and change management are critical.

 

Another issue is data security and governance. There are AI services on the internet that don’t offer a secure environment. Employees who knowingly or unknowingly input sensitive client data into these systems could breach client trust, violate GDPR and data laws or leak IP. Set clear policies, use AI tools that are enterprise-approved, train staff on what’s appropriate to share, and consider technical safeguards to block or monitor the use of external AI sites from work systems.

 

Finally, jumping into AI without a clear goal or expectation is another serious pitfall. Some SMEs get caught up in the AI hype and try to use AI everywhere without asking, “What problem am I solving?” “What value will this bring?” If you don’t have clarity here, you can waste time and money on AI projects that don’t move the needle.

 

 

A strategic path to AI adoption

A phased, goal-driven approach is key when rolling out AI into an organisation. The process should begin by identifying current pain points, both within teams and the wider business operation. These might include slow customer response times, time-intensive proposal writing or bottlenecks in reporting.

 

From there, pick a quick win: a project that can be delivered in four to eight weeks and has a clear and measurable impact. This first success is critical for building momentum and confidence internally with teams. Treat it as a learning experience and refine it as you go, whilst offering necessary staff training.

 

Once the pilot proves its value, expand its scope or move on to tackle a more ambitious project. Over time, the goal is to integrate AI capabilities into multiple parts of the business, with knowledge and skills spreading beyond a handful of early adopters.

 

Sooner or later, every business will need to adapt to AI. So for small and medium-sized businesses, it isn’t a question of whether or not to adopt AI; it’s how quickly and effectively they can do it – before someone else does.

 

---

 

Mark Rotheram is CTO at BCN

 

Main image courtesy of iStockPhoto.com aND Suriya Phosri