Annabel Kerr, Director – Technology
For businesses across Scotland’s technology sector, the next six months could be an important period in turning AI ambition into meaningful action.
That was a key theme at ScotSoft 2026, where discussions moved beyond the excitement surrounding AI’s capabilities towards more practical questions about its value, governance, security and commercial potential.
The challenge for businesses is no longer about whether to adopt AI, but understanding where it can make a genuine difference, managing the risks and turning investment and innovation into lasting business value.
The Clark.tech team attended sessions across the conference programme, hearing from technology leaders, security specialists and industry experts.
Here are five questions that stood out to us – and our take on what they could mean for businesses over the next six months and beyond.
1. Are we solving the right problems with AI?
As organisations face growing pressure to embrace AI, several speakers challenged audiences to think more critically about where the technology genuinely adds value. AI isn’t a universal solution and, in some cases, established technologies such as traditional machine learning may still be better suited to a task.
Matthew Leeper of Silmaril AI argued that structured, deterministic workflows give organisations greater predictability than autonomous agent execution. With AI token costs capable of escalating quickly, this predictability brings both operational and commercial benefits.
Meanwhile, Mark Rodger of Telefónica Tech challenged businesses to avoid using AI just to make existing processes incrementally faster. Instead, he suggested asking whether those processes need to exist at all. The bigger opportunity, he argued, is to redesign how work gets done rather than chasing marginal gains by adding AI to what you’ve always done.
There was also recognition that being first isn’t always an advantage. A “wait and see” approach can give organisations time to learn from others’ mistakes, make adjustments and potentially “leapfrog” those who jumped in too quickly.
The overall message? Start with the problem, not the technology.
2. Who is accountable when AI makes – or influences – a decision?
In her keynote, BCS Deputy President Emma Logan disputed the idea that AI “makes” decisions independently. People decide what data goes in, when humans intervene and how much weight to give its outputs – so accountability needs to reflect the human decisions made throughout that process.
She drew comparisons with professions such as medicine, accountancy and aviation, where trust and accountability have been built over time through layers of professional standards, competency, regulation and independent scrutiny. We trust an aircraft to get us safely from A to B not just because we trust the technology, but because we trust the people, processes and safeguards surrounding it.
Emma believes that AI needs a similar framework: professional standards and accountability that sit with the individual and move with them throughout their career, rather than existing only at an organisational level.
3. How will regulation change the way businesses use AI?
The EU AI Act, GDPR and Data Act featured prominently in ScotSoft’s legal track, accompanied by an important reminder that businesses don’t necessarily have to be based in the EU for European rules to affect them.
One of the most practical messages from these discussions was that before organisations can determine which rules apply, they need a clear understanding of what systems they use, what data they process, who provides them, what decisions they influence and where they are being deployed.
As adoption grows, understanding the rules surrounding AI looks increasingly less like a specialist legal concern and more like part of everyday digital literacy.
4. How do we put a harness around AI?
As businesses explore the use of AI agents capable of taking autonomous action, the question of control is becoming much more important.
On ScotSoft’s security track, speakers highlighted the importance of AI “harnesses” – the systems, rules and controls that govern what an AI model can access, what actions it can take and when human intervention is required.
William Wright from Closed Door Security neatly summed it up: put the control in the harness, not the prompt.
He and his fellow security experts repeatedly explained that rather than relying on instructions telling an AI system what it should or shouldn’t do, organisations need to build safeguards into the infrastructure surrounding it.
It’s an important distinction: giving an agent access to business systems or sensitive information creates a very different risk profile from asking a chatbot to draft some copy.
5. Where should Scotland focus its AI ambitions?
Scotland’s position in the future of AI was a recurring theme, from ScotlandIS CEO Andrew Weir’s opening remarks through to the closing industry panel.
Scotland has strong AI research capabilities and valuable datasets. But how can it turn those strengths into technologies that are adopted, commercialised and scaled, particularly when countries around the world are pursuing the same opportunities?
One potential answer discussed by the panel was specialisation.
Rather than trying to compete everywhere, Scotland could focus on the intersection of data, its deep sector expertise and AI – identifying areas where it already has distinctive knowledge or datasets and developing specialist AI models that solve a challenge specific to those areas.
Doing that will require stronger connections between universities, industry, investors and public services, alongside the expertise to assess not only the risks associated with technologies like AI but whether they are genuinely the right solutions.
What comes next for Scottish tech?
From adoption and regulation to security and commercialisation, ScotSoft highlighted an industry grappling with a hard set of questions about how we put AI to work in the real world.
For technology companies developing AI-powered products and services, finding answers will be essential if they want to lead the way, demonstrate the value of their solutions and earn the trust of customers.
But they can’t do it alone. Perhaps the clearest message from the conference was a very human one: progress depends on bringing the right people together – across industry, academia, investment and the public sector.
As Scottish tech firms navigate these challenges and opportunities, how they communicate about AI will be just as important as the technology itself. With that in mind, here are our key comms takeaways:
- Lead with purpose, not AI hype: Show how AI is solving real business challenges and delivering tangible value
- Shift the narrative from efficiency to transformation: Emphasise AI’s potential to fundamentally rethink processes and unlock new ways of working
- Build trust through accountability: Emphasise the importance of human oversight, professional standards and clear responsibility for AI-driven outcomes
- Make responsible AI a core message: Highlight governance, security and regulatory awareness as essential to successful AI adoption
- Differentiate through specialist expertise: Highlight how your company’s unique datasets and deep sector knowledge give you a competitive edge in delivering tailored AI solutions
If you’re looking to partner with informed, experienced tech communications experts, get in touch here.