People don’t like change. Some colleagues may have been completing the same processes for many years. Everyone has been told to use new tools, which have caused resentment and disengagement. It was 10 years since Slack launched in February. It is one of a very small set of tools that have been able to fundamentally change culture and how we communicate at work. It’s rare for new tools to make a positive impact, but Gen AI has huge potential to do this (we would say it is already doing this). However, Gen AI isn’t a single application, it comes with more complications on how to engage with it and is currently saturated across a variety of different platforms and models. Training and documentation are important to set up your team to embrace new workflows and understand the benefits.

Large language models or multimodal models, such as GPT-4, which powers ChatGPT, or Google’s 1.0 pro model, which powers the Google Gemini web app, are trained on vast datasets. However, no matter how comprehensive the training data is for these models, there’s always potential for missing data or additional context that could contribute to Gen AI providing more accurate answers to your specific business or industry questions. The processes of Retrieval-Augmented Generation (RAG) and LLM Fine-Tuning techniques allow you to make the AI model more bespoke to your business needs. This is an important step to be aware of and incorporate into your Gen AI strategy to get more relevant and accurate outputs.

We are seeing many businesses experiment with off-the-shelf models and not achieve the best results. One of the greatest values for marketing teams is customising or fine-tuning models with your own data to address your unique business challenges effectively.

You don’t need to start from scratch. All the major software providers are integrating Gen AI into their tools. For example, Microsoft has introduced Copilot Pro, integrating Gen AI features across Word, PowerPoint, Teams, and SharePoint. Adobe has already launched new AI tools and integrated Gen AI features into their core products. Being aware of the roadmaps and what is already available from software you are already using is a great way to introduce Gen AI quickly and effectively into your marketing workflows. However, it is quickly becoming a saturated market with many new Gen AI tools launching with similar features. There is, of course, also a cost associated with having access to these tools, so it can be hard to understand the strengths and weaknesses of different tools and which tool(s) will work best for your business challenges.

1. Document your challenges: Collaborate with your team to document current challenges, time-consuming tasks, limitations, and areas where you can’t scale.

2. Map out current processes: Create process flows of your major workflows (and tie them with your challenge document) to understand the steps team members are currently completing and where blockers and time are being spent.

3. Create your solution wishlist: Document where you want to make improvements. Don’t worry at this stage if you don’t know if it’s possible or know all the steps to get there. What’s important is prioritising your aims. We have found that AI isn’t always the solution; taking a problem-first approach allows you to have a holistic view of your aims and not get bogged down in trying to map out a solution yet.

4. Audit current adoption of Gen AI and review security: Understand how your team is currently utilising Gen AI tools. There may already be pockets of innovation that need to be scaled. It’s also important to understand any data security risk where private or sensitive data is being inputted into models. Speak to your current tech suppliers on what they are doing with any data you're giving them for AI purposes ensuring it follows your privacy policy commitments.

5. Create prompts for quick wins: Start to create a standardised set of prompts for your regular tasks identified in point 1. Start simple with the strengths of out-of-the-box AI models, for example, idea generation, research, proofing, drafting content, metadata. Review prompt libraries such as AIPRM to give you ideas for different prompts. This can be as simple as documenting prompts in a shared Google Sheet or Excel

6. Get a training programme organised: There’s lots of online training available to raise awareness of what Gen AI is and its use cases. Amazon has a great intro to AI and AI learning for decision-makers, Google has a free Gen AI course too, and Udemy has lots of costed options. These courses are a great starting point for general training. We would then recommend rolling out more bespoke training for your specific AI requirements and applications through a specialised partner.

7. Identify important internal content & data: Start to to identify internal content and data which could to be elevated or made more useful by connecting it to a Gen AI model. Taking into account your priority team challenges and what assets you have available.