Interactive’s head of data and AI, Lizzy Jones, has urged marketing and advertising teams to create workplaces where employees feel comfortable experimenting with AI, making mistakes and sharing unfinished ideas.
Speaking to B&T at the recent TechLeaders conference in Bowral, Jones said businesses needed to move away from cultures where employees feel they must have the “perfect answer” before speaking up, arguing that experimentation is essential as AI tools continue to evolve.
For Jones, creating an “AI-curious” culture is not about mandating the use of AI, but giving employees the time and psychological safety to experiment with it.
At Interactive, employees are given dedicated time to explore AI through fortnightly “AI Playground” sessions, where staff from across the business demonstrate how they are using the technology.
The sessions run for around 30 minutes on Friday lunchtimes, with employees typically spending 10 to 15 minutes showcasing an AI experiment, tool or process.
“Some people have slides. Some people do a bit of a demo. Some people walk through the process they’ve gone to get something up and running,” Jones said. “Some people come and show stuff that’s not quite working yet, so it’s really cool.”
Jones said the approach had helped employees think more laterally about how AI could be applied to their own work, rather than simply learning how to use individual tools.
“It’s how do you think, in a diverse way, laterally about how you could use it and how you could optimise for it?” she said.
The approach is also designed to tackle a cultural barrier Jones said could prevent employees from experimenting in the first place: a fear of being judged.
She said people at Interactive could sometimes be reluctant to speak up unless they were confident they had the right answer, particularly given the company’s mix of experienced technology professionals.
But AI has created an environment where “not everyone knows everything”, making it easier for employees to admit they do not know whether something will work and try it anyway.
“I think the AI thing is really cool because not everyone knows everything,” Jones said. “And so that’s been a really big shift for us, and it’s something we’re still promoting.”
Jones believes leadership also has an important role to play in creating that environment.
At Interactive, senior leaders were encouraged not only to learn about AI, but to demonstrate that they were using it themselves and openly discuss what worked and what did not.
“If you know people at the top are trying it, maybe it’s working, maybe it isn’t, that sort of gives you licence to try things as well,” she said.
Rather than creating additional AI-specific meetings and requirements, Interactive has also incorporated AI discussions into existing team meetings, making it a standing agenda item.
Jones said this helped keep AI “top of mind” without adding another layer of work to employees’ calendars.
Why teams shouldn’t be locked into one model
While Interactive has a company-endorsed AI tool, Jones said she did not believe creative and marketing teams should necessarily be restricted to one model.
The company endorses Microsoft Copilot for business use, particularly from a security and governance perspective, but employees are not prevented from accessing other models such as ChatGPT or Claude through their browsers.
Jones said organisations should continue testing different models because they are evolving at different rates and can produce different results.
She pointed to Microsoft’s efforts to bring multiple models into its Copilot environment, allowing users to compare outputs from different systems.
“I think it’s good to keep testing because they advance at different levels, and so different models are going to be better at doing different things,” Jones said.
However, once a team identifies a model that consistently delivers strong results for a repeatable task, Jones said there was little value in constantly switching between tools.
“You can just go ‘Great this model works really well, I’m not going to even think about it again for another sort of few months,.'”
The key, she argued, is avoiding a “model monopoly” without creating unnecessary noise for employees.
And as AI becomes more embedded in agency and marketing workflows, Jones also expects greater transparency around how the technology is being used.
She suggested agencies could increasingly spell out their AI practices to clients, while AI usage could also become part of contractual discussions.
Some clients may welcome AI if it creates efficiencies, while others may explicitly prohibit its use.
Jones said Interactive already works with some large government clients that have instructed the business not to use AI.
At the same time, she believes agencies could eventually build AI-driven productivity gains into their commercial arrangements, potentially reducing costs as certain processes become more efficient.
“You could start to build into some of your deals is like a productivity gain and you say, ‘Well, maybe we’ll use AI, and our services won’t be as expensive over time’,” she said.
Ultimately, however, Jones said organisations should focus less on forcing employees to use AI and more on giving them the opportunity to learn.
“The message we try to give people is this is your opportunity to learn and the company’s got your back, and we’ll actually create space for you to learn and try,” she said.
“This technology is not going anywhere, so it’s an opt-in thing. We don’t certainly mandate people do it or use it, but we’re trying to inspire them to get on the train, because the train has definitely left the station.”
