Getting to the heart of artificial intelligence

With a data science and mechanical engineering background, founder Sarah Mitchell understands the hype around AI. 

She also knows that to understand AI, you have to cut through the jargon. Mitchell founded Anadyne IQ in 2023, a New Zealand-based AI consultancy, to do just that. 

“Our core philosophy is around clarity over complexity. I’m passionate about breaking down complex topics so businesses can get going with AI without all the confusion and fear,” she says. 

‍

‍Sarah Mitchell sits down with Caffeine to discuss AI, a tool she believes is “industry agnostic in terms of its ability to support different businesses”. 

‍

How do you use AI in your business operations?

I try to weave AI into everything we do. In terms of what that looks like, it’s automation, augmentation and insights. 

For example, on a day-to-day basis, I might use it for content creation, perhaps writing new content for my website or blog or generating images. I might also use it for ideation and brainstorming as a two-way conversation with a generative AI. It’s kind of like back-and-forth ideation, which is a lot of fun. 

I also use it for analysis and insight, leveraging the technical aspects of my experience to develop models and help support advanced decision-making. 

‍

How do you stay up to date with AI? 

It’s a lot of just trying to be embedded in the industry. Reading articles or the latest journals, experimenting with some of the new tools to see what they do and how they work. The most important thing is then being able to identify from that tool whether there’s potential for it to be applied for my clients. 

If anyone wants a regular feed of AI information, LinkedIn is quite a good way to do that. I enjoy following people like Allie Miller. She’s an AI voice in business. There’s also Cassie Kozyrkov, Google’s first chief decision scientist. She applies AI and decision science to help businesses make better decisions. 

‍

What has been your biggest challenge with AI? 

The pace. I'm very conscious of recommending solutions that won't quickly be superseded by the next iteration of Open AI for our clients. 

It's about staying on top of things, going back to basics, and solving tangible business problems, starting with low-hanging fruit and not getting too captivated by shiny objects, brand-new tools, and exciting things. There is absolutely a role for them, but my focus is on leveraging the latest advancements to deliver value rather than just wow factor.

‍

How do you manage privacy with AI? 

It is extremely important to be cognizant of how customer data is used. This means ensuring that it is consented, protected and used for the purposes you have asked them to use it for. 

For example, if I'm talking to clients and recommending tools, I always say to read the fine print of what you're signing up for. 

If you cannot opt out of them using your data for training, modelling, or other purposes, just be cautious about what information you put into that tool. Leave out your proprietary information or any personally identifiable information. 

‍

How does AI legislation need to be implemented? 

AI legislation should be enacted responsibly, but with pragmatism, ensuring the safe use of AI without stifling the industry and the advancements. I imagine we will follow something along the lines of the EU Artificial Intelligence Act. 

It's really about making sure people are aware of the inherent risks – because there are risks with utilising AI – but then mitigating various risks, and being conscious of what needs to be done. 

‍

What do you wish people knew about AI? 

That it is a force multiplier. What I mean by that is it can augment your workflows and help you do the things you don't enjoy in your job; anything dull, boring or repetitive. 

Start by looking at the bits about your role that you dislike the most. AI is ripe for solving those sorts of problems for you. 

‍

What trends are you following closely? 

Agentive AI, where AI – acting with a little less human oversight – can perform tasks and make more decisions on your behalf. 

At the moment, we still require quite a lot of direct input with AI. For example, if you think about ChatGPT, you’re asking questions and interacting with it directly. Agentive AI will act more independently; you’ll see a task, and it will go and do it. 

I can easily see a future where, if you order a product online and it is damaged, you can get AI to contact customer service and arrange the return or replacement. With Agentive AI, you’ll likely see your AI talking to the company’s customer service AI and arranging it all for you. Hopefully, it will remove some painful tasks that people don’t enjoy. 

‍