The big US tech companies are responding positively to pressure for more transparency about how and when artificial intelligence is being used. Now small and medium-sized businesses need to decide what approach is appropriate for them and their customers.
How do small and medium-sized businesses (SMBs) make more use of artificial intelligence in a way that inspires confidence rather than suspicion from customers? The US tech giants’ recent introduction of AI watermarks underlines the importance of this debate, with both Anthropic and OpenAI recently unveiling initiatives to improve the transparency of their AI tools.
Anthropic led the way in August with the announcement that text generated by its Claude AI model will henceforth include patterns that will be detectable to the company and to those who have a key to decode them. While this watermark won’t be immediately visible to readers, it provides a means to check whether a piece of text has been produced by Claude.
OpenAI subsequently announced something very similar, though it chose, in the first instance, to limit the roll-out of the new tools to the European Union, where new AI regulation mandates transparency. The watermarks will be added to ChatGPT and Codex outputs worldwide over time, OpenAI explained.
In practice, this will have limited impacts in the short term. Only certain stakeholders – regulators and independent researchers, for example – will be given the keys required to detect the watermarks, though access will widen over time. And the watermarks only tell you AI has been used to generate content, not how the AI reached the output or decision. Still, the changes meet regulatory demands in the EU and represent a shift towards more openness.
Customer confidence depends on transparency
These initiatives also speak to an important debate that SMBs will increasingly need to consider for themselves. As they make use of AI that is embedded in the tools and processes they use every day, the fact that AI is involved won’t always be obvious to customers when they’re interacting with the business. Research suggests this is a problem, with customers demanding more visibility.
In one recent study by cybersecurity company PingIdentity, for example, 57% of consumers said that knowing when AI is being used is the most important factor when interacting with a brand. By contrast, only 27% said they were more interested in getting the fastest and most accurate result, regardless of disclosure.
Tina McKenzie, policy chair at the Federation of Small Businesses say SMBs need to take concerns about transparency seriously. Its latest research shows that while the use of AI by small and mid-sized businesses has risen significantly over the past two years, so have concerns about AI-related risks – 92% of SMBs now express anxieties, up from 73% in 2023.
“There is a healthy realism among small businesses when it comes to AI,” says McKenzie. “Business owners can see the potential, but they are also asking sensible questions about how their data is used, who is responsible when things go wrong, and how they can adopt the technology safely.”
For some SMBs, that may mean flagging up every use of AI to customers in very visible ways and explaining how outputs have produced; it may also mean publishing more detail about the safeguards they apply and to what extent AI tools are subject to human oversight. Other SMBs may be more circumspect – perhaps still highlighting their use of AI but directing customers to other resources if they want to know more.
Adopt AI that empowers customers
The right approach will vary from sector to sector – and according to the nature of the SMB’s customer base. But providing customers with a means to make informed decisions about their exposure to AI is imperative, argues Kamila Miller, programme director for AI at Henley Business School. “For organisations to thrive, they must understand that trust is not a simple asset to maximise, but a nuanced relationship to cultivate and maintain,” she argues. “This means building systems that empower customers through AI's capabilities while preserving their sense of ownership over choices.”
Sage’s own approach to transparency offers an example of one possible route forward. Last year it launched the AI Trust Label, as part of its commitment to responsible technology development. Customers who use Sage Intacct, Sage’s suite of AI-powered finance software tools, see the AI Trust Label when using these products. Each label sets out key information on trust indicators such as compliance with privacy and data regulations, how customer data is used, the presence of safeguards to prevent bias and harm, and the systems in place to monitor accuracy and ethical performance.
In addition, users have the option of accessing additional information from Sage’s Trust & Security Hub about the deployment of AI within the tools.
Aaron Harris, Sage’s chief technology officer, says the initiative is designed to provide comfort and reassurance for businesses looking to make use of AI but concerned about what’s going on beneath the surface with the technology. “Businesses should be able to understand how AI works and what it means for them without needing to be tech experts,” he explains. “This is our way of making that possible.”
Harris points to recent Sage research showing that when businesses lack that confidence and understanding, they’re much less likely to embrace AI. Among SMB leaders who say they trust AI, 85% are active users of the technology in their business; among those who don’t yet have that trust, the percentage drops to 48%.
Certainly, customers have reacted positively to the AI Trust Label’s introduction. For example, Tony Phillips, assistant controller at Boston-based Income Research & Management, says: “In my role, I need to know that decisions and outcomes generated by AI are accurate and fair – I use Sage Intacct’s AI features regularly so it’s reassuring to know how the technology is built.”
What does transparency sound like?
There may also be other steps businesses can take to help customers get comfortable with AI usage. A recent Sage initiative used music to bring the principle of explainable AI to life. The Financial Symphony turned real – but anonymised – data from businesses in the UK, US, South Africa and France into a piece of classical electro-acoustic music, performed by a string quartet. The music is accompanied by a published methodology to make the translation of anonymised business data into music understandable and verifiable.
“Every note, rhythm and change in volume has its origins in real business data, and the journey from data to music is transparent,” explains Harris. “As AI plays a bigger role in how businesses make decisions, it’s a reminder that trust is earned by showing the working, not just the answer.”
SMBs don’t have to hire their own string quarters to make their point. But they can still embrace the principle – the idea is simply to communicate about the use of technology and AI as openly and clearly as possible.
The bottom line? SMBs implementing new AI tools need to think carefully about when and what to tell customers about how AI is being used. That way, SMBs can take their customers with them on the journey as they improve products and services through the use of AI, rather than risk losing their trust.
