Podcast

4 min read

AI conference season: key takeaways for SMEs

Conferences are packed with talks, panels, and networking — and buzzwords. Yet, there are always useful learnings beneath the noise.

If you’re an SME leadership team trying to make sense of AI adoption, conferences can feel overwhelming. Everyone’s talking about agents, sovereignty, human oversight and transformation. But what should you actually do differently on Monday morning? That’s what we’ve spent time filtering through.

TLDR: Skip the AI hype narrative. Focus instead on three things: (1) strategic clarity about what your business needs to become; (2) realistic cost planning for token usage; (3) a pragmatic, safety-first approach to automation. The backlash against AI is real — lean into business outcomes, not technology evangelism. And start thinking now about how AI skills will diversify your workforce.

The buzzword landscape

If we’d created buzzword bingo cards, we would have won immediately. “Agents,” “agentic automation,” “human in the loop,” “human at the helm,” “RAG,” “context,” “data sovereignty,” “digital transformation” — these words appeared across nearly every presentation.

The pattern tells you something important: the industry is collectively working through the same problems. But the repetition also means that if everyone’s building agents are doing the same things, there’s no competitive advantage in the agents themselves.

This insight cuts to the heart of the transformation question. Building five agents to automate routine tasks is table stakes. Building a genuinely different business model — one that uses AI as infrastructure rather than as a feature — is what actually separates winners from followers.

Takeaways for SMEs

1. Investment money is flowing to specific bets

At Panathena, it was clear where VCs and angels are placing their bets: hardware plays (robotics, defence, logistics, manufacturing) and infrastructure plays (trust, financial services, healthcare).

The pattern isn’t about AI itself — it’s about sectors where AI solves urgent, expensive problems with clear ROI. If you’re in manufacturing or defence, that’s a tailwind. If you’re elsewhere, you need to find your own equivalent problem. That’s where your transformation story starts.

2. Cost is real, and it’s changing

Everyone at both conferences was talking about token anxiety — the creeping horror of watching your AI costs spiral. The reason? AI right now is heavily subsidised. OpenAI, Anthropic, and others are burning billions on hardware while you’re paying a fraction of the true cost. That subsidy will end. Sooner rather than later, most people think.

This means: design your systems now with cost efficiency in mind. Focus your agents. Minimise variability. It’s often cheaper to run three specialised agents than one that tries to do everything. This isn’t optimisation theatre — it’s survival planning.

3. There’s a K-point in careers that’s becoming very real

One panel speaker made a point that stuck with me: as a business leader, using AI tools to upskill your team is one of the most powerful things you can do for their careers — regardless of whether they stay with you. Those who can use AI effectively will see their career arcs rise. Those who can’t will see it move down.

It’s not a flatline decline for the laggards. It’s acceleration downward. That K-shape is becoming the reality of the job market. Leaders who treat this seriously — by genuinely embedding AI into how work gets done — will retain talent and attract the best people. Those who don’t will face a skills crisis that upskilling programs alone won’t fix.

The backlash is real

This one didn’t come from the conference floor — it was happening during the conference. Eric Schmidt was booed off stage at a graduation for mentioning AI. Young people are voicing real hostility to AI adoption. The innovation community was deliberately overlooking this tension.

Your SME shouldn’t. The spotlight is shifting from “should we adopt AI?” to “at what cost, and to whom?”

The answer isn’t to shy away from AI. It’s to stop leading with it. Lead with what you’re trying to achieve. Use AI as a tool to get there. But don’t make the technology the story. Make the business outcome the story. That’s how you navigate the backlash.

What should you do?

  1. Get strategic clarity. Where do you want your business to be in five years? How does AI enable that shift — not as a feature, but as infrastructure that changes how you operate?
  2. Think transformation, not tactics. Upskilling people on ChatGPT isn’t a transformation. Rethinking your processes, your competitive advantage, your operating model — that’s transformation.
  3. Plan for cost control now. Token economics will normalise. Design your systems to survive that normalisation.
  4. Prioritise trust and safety. The startups doing this right aren’t rushing. They’re being pragmatic. Follow that lead.
  5. Take the K-point seriously. Your people need real, embedded AI capability — not training courses. That means changing how work actually gets done.

Conferences are useful. They let you hear what everyone’s thinking. But the real work happens when you filter out the noise to focus on what actually matters for your business.

We can help with that.

A deep dive into full-trust AI for SMEs.

Below the Surface: full-trust AI

Real AI adoption needs sensitive data, and commercial LLMs can't be trusted with it.

A full-trust AI stack is easier and more affordable than most think.

We compare 6 models, explain 5 layers and outline 4 steps to start.

Read the free guide
Giles Thurston and Theo Paraskevopoulos