Reclaiming your time: GenAI for better slides
Nobody goes into consulting because they love formatting PowerPoint. Here's the workflow I use to get from rough notes to a finished deck — and where it still creaks.
There's a particular misery in consulting called the slide time sink. The thinking is done, the value is there — and then someone loses two days to alignment guides and bullet indentation. High-value work, derailed by tedious formatting.
I wanted to see whether the iterative logic we use in software could fix that. Short answer: yes. With a conversational LLM for the structure and Nano Banana for the visuals, I can go from rough notes to a compelling deck in a fraction of the time it used to take.
The five-step workflow
- Start with a conversation. Describe the project to the AI and ask it to help you build a prompt template that matches your branding and visual style.
- Get the content right first. Before anything looks pretty, the message has to be right. Refine your points and put them in a clear hierarchy.
- Assemble the brief. Merge the prompt template with your headlines and bullets, and tell the AI what visual energy each slide needs.
- Translate for the machine. Have the LLM format everything specifically for image generation — corporate colours, style guides, the lot.
- Generate and polish. Feed the output into Nano Banana, then take the finished asset straight into your deck.
Chat, don't prompt
A one-shot miracle is not coming. The power of GenAI is the back-and-forth — the same reason you wouldn't build a whole software system in a single prompt. Break the slide into chunks: template first, then words, then image.
// Prompt template example
[Asset Type] Stylized Map Illustration of [Subject] a local region, [Action/Context] showing connection lines between a wind farm icon and a house icon, [Art Style] Friendly vector illustration, [Color Palette] Light background, Grey (#32373C) paths, Teal (#64C1C6) energy sources, [Composition/Lighting] Top-down view, clean and simple, [Aspect Ratio] 16:9.
The part that matters
When I was at KPMG I watched formatting swallow whole weeks. Under pressure, teams obsess over polish long after the value has been delivered. It's an exhausting, burnout-shaped cycle.
This workflow is what we always hoped AI would do — it gives you back the strategy and the big ideas, and leaves the tedious production work to the machines.
I ran this exact method on a renewable-energy project here in South Africa recently. There's still some window-hopping between tools, but the time saved by dumping my thoughts into a notepad and letting the AI do the visual heavy lifting was absurd.
Where this goes
Right now the workflow is a bit manual. It won't stay that way — smart organisations are already baking these efficiencies into their everyday tools. The people who stay relevant will be the ones who stop identifying as slide builders and start acting like the strategists they were hired to be.