Skip to main content

Gen AI - the new frontier for marketing

Driven by the gen AI Hype Cycle, we’re seeing many of our clients experimenting with generative AI in creative content production and improving customer experience (CX) through more personalised and conversational customer interactions.

Take a look at the latest Deloitte research on Gen AI in content marketing

While existing foundation models that are trained on natural language processing (NLP) and image generation can be deployed with little computation and cost, fine tuning the models is required for more specific use cases and to avoid the associated risks of hallucinations, bias, and security breaches.

When approached by a client to implement a Gen AI solution, we ask these five questions to ensure that the solution we propose is best suited for their business requirement.

  1. Which key benefits are you hoping a generative AI program will bring to your marketing organisation?  Each organisation’s needs are different and you want to be crystal clear on what success will look like in your business.
  2. How can generative AI help you keep the customer first? A well-thought-out strategy adjusts and interprets customer behaviours and desires in real-time.
  3. How can generative AI help you enhance and support the creative talent in your organisation? Use Gen AI is good as a starting point for creative ideation and first draft production, but it can’t replace human judgement and empathy.
  4. How can you ensure that generative AI outputs will be trustworthy and ethical? By focusing on AI bias and emphasising AI ethics, companies can protect customer data while building brand equity and customer trust.
  5. How will you measure generative AI’s effectiveness in connecting with customers? A successful implementation will need to carefully think through what KPIs will be your signal of successful integration.

Read more here

By considering these questions, we are able to help our clients configure a clear solution and implementation roadmap to achieve their intended results.

Did you find this useful?

Thanks for your feedback

If you would like to help improve further, please complete a 3-minute survey