The conversation around artificial intelligence is changing.

At this year’s Ai4 conference in Las Vegas, more than 12,000 attendees gathered to discuss the technologies, strategies and challenges shaping the next stage of AI adoption. For Harvey Nash Tech Evangelist David Savage, the most interesting conversations were not necessarily about getting AI into the enterprise. They were about what comes next.

From August 4–6, David moderated a panel exploring the growing role of AI coding assistants and how AI agents could augment software development. He also recorded conversations with digital and technology leaders for Harvey Nash’s Tech Talks podcast.

Across those discussions, one theme kept resurfacing: organizations are spending significant time figuring out how to deploy AI, but far less time defining what success looks like once deployment is complete.

Here are David’s key takeaways from Ai4 2026.

 

An AI-savvy audience still had plenty of questions

Walking into a conference session, you never quite know who will be in the room.

During the panel on AI coding assistants, David asked the audience how many people were software engineers. Roughly 95% raised their hands.

That was revealing.

The session attracted an audience with substantial technical experience, yet the discussion showed that even people who are deeply engaged with AI are still working through fundamental questions about how these technologies should be used.

Research and conversations featured on Tech Talks have frequently highlighted a similar dynamic: engineers can sometimes feel more comfortable using AI in their day-to-day work than their leaders do. But technical familiarity does not eliminate uncertainty.

Even experienced professionals are looking for practical guidance, reassurance and opportunities to learn from their peers.

Being further along in the AI journey does not mean having all the answers.

One audience question captured that uncertainty particularly well. A teacher took the microphone and asked:

“What should I be teaching my kids?”

It was a simple question, but it shifted the conversation away from models, coding assistants and agents and toward something much bigger: the people who will live with the consequences of today's technology decisions.

AI may transform how work gets done, but human judgment, creativity and accountability remain critical. Organizations that reduce roles based purely on expectations of AI-driven productivity may also find that they still need people to provide oversight, make decisions and take responsibility for outcomes.

For the next generation, developing the skills and confidence to work alongside increasingly capable AI systems will be just as important as understanding the technology itself.

The real question starts after deployment

At the end of every Tech Talks interview, David asks guests the same question:

“What aren’t we talking about?”

It is designed to move the conversation beyond the subjects dominating the technology industry and uncover issues that may not yet be receiving enough attention.

At Ai4, one potential blind spot became particularly clear.

There was extensive discussion about deploying AI at scale. There was considerably less discussion about what organizations should expect once that deployment succeeds.

Erin Boyd, Chief Digital Strategy and AI Officer at The AES Corporation, summed up the concern:

“I do personally think we are not focusing enough on what happens when we actually successfully deploy agentic AI.”

Organizations are investing heavily in selecting models, building agent frameworks and creating guardrails. Those are important steps, but they are means to an end.

The bigger questions are organizational.

What happens when agentic AI becomes embedded in day-to-day operations? How might responsibilities shift? Will teams become smaller or simply change shape? How should decision-making evolve when AI agents are handling increasingly important tasks? And who remains accountable for the final outcome?

Those questions should be part of the strategy before AI is deployed at scale, rather than being addressed after the technology is already embedded in the business.

Start with the outcome, then choose the technology

Another recurring theme at Ai4 was the importance of connecting AI investments to measurable business outcomes.

Stephen Henn, Managing Director of AI Innovation at DLA Piper, described the difference between starting with technology and starting with the problem the business actually needs to solve.

“What do I want to get out of it? What is the big impact? We know that if we solve this problem, we’re going to get ROI. Then work backwards.”

That approach puts the business objective ahead of the technology.

Instead of asking where a new model or AI agent could be deployed, organizations can start by identifying a specific business challenge, defining the desired result and determining how success will be measured. Technology can then be evaluated against that outcome.

This may sound straightforward, but it represents an important shift in how organizations approach AI.

A technically successful deployment is not necessarily a successful business implementation. AI can be integrated into a process, perform exactly as designed and still fail to deliver meaningful value if the original objective was unclear.

The technology should support the outcome not become the outcome.

AI investment needs a financial perspective

AI conversations also need to extend beyond technology and operations.

Arjun Srinivasan, SVP of AI and Data Science at ShipStation Global, highlighted the importance of bringing financial leadership into the discussion:

“I would love to hear more from a CFO’s perspective. They’re the ones who sign the cheque. They’re the ones who are answerable to the board. I’m not hearing that.”

As AI programs move beyond experimentation and require larger, longer-term investments, CFOs and other financial leaders have an increasingly important role to play.

They need visibility into what organizations are trying to achieve, how investments will be evaluated and what business results are expected in return.

AI strategy therefore cannot sit exclusively within the technology function. The people responsible for approving investment and ultimately explaining that investment to the board need to be part of the conversation from the beginning.

Moving from AI adoption to AI outcomes

Ai4 provided a useful snapshot of where organizations are today.

Many are still working through the fundamentals: selecting models, establishing governance, building guardrails, developing use cases and determining how AI can be introduced responsibly at scale.

Those conversations matter.

But the next stage of AI maturity will require a different set of questions.

If an organization successfully deploys AI across critical workflows, what changes? What happens to teams and roles? How does decision-making evolve? How is accountability maintained? How will leaders know whether the investment is actually delivering the expected business impact?

These questions are particularly important as organizations move from experimentation into broader implementation.

The AI conversation has largely focused on capability what these systems can do and how quickly they can be deployed. The next phase needs to focus equally on organizational impact and business value.

Ultimately, the challenge may not be figuring out how to deploy AI.

It may be defining what the organization wants to become once AI is successfully deployed.

 

Hear more perspectives from Ai4 as David’s conversations with digital leaders are released on the Tech Talks podcast. Want to be a guest kindly reach out to David Savage or Harvey Nash.

Watch the video below for more of David’s conversations with technology and digital leaders at Ai4.