Four teams, four briefs, one hour – what did we learn from the Forum’s first ever AI hackathon?
The Financial Services Forum brought together a group of members this week to try their hands with some AI tools.
The session, devised and hosted by Sasha Dabliz, Head of Marketing at Waverton, tasked four teams with creating a product and a campaign from four pre-determined briefs.
- A savings product for under-25s
- An eco-conscious fashion brand
- A new dog food
- A running shoe
The teams, comprising attendees from different companies and of different seniorities, had an hour and access to a range of free AI tools on the internet. There was no competition or prize – the idea was simply for everyone to try out the tools, present what they’d been able to produce and report back on their experiences.
Here are some of the learnings that came out of the event:
- There are as many ways to use AI as there are ways of thinking
From observing the groups it was clear that each had taken a completely different approach to the task. Some groups brought in ChatGPT at the very earliest stages of ideation, asking it for gaps in the market and using it to generate as unique a product idea as possible. This was particularly evident in the running shoe group, who devised an app-connected trainer that allowed people to track their fitness.
Others quickly brainstormed to come up with their idea and then turned to the tools, spending the bulk of their time on the creative elements of the campaign, with the savings account having a particularly developed library of creative.
What was clear was that whichever approach they opted for, it was extremely easy to get started. Some attendees had never tried AI before but were able to easily get to grips with the free tools in an hour.
- ChatGPT 4 is vastly more powerful than the free version
One group were able to try more advanced tools as a team member provided access to a paid version of ChatGPT, which integrates not just the standard chat interface but also the ability to generate slide decks, pictures and videos. The capability of the tool and what it was able to generate in an hour compared to its free counterpart was notable.
- Quality control is definitely a must
Anyone who has tried AI will already know the results can’t necessarily be relied on. In an otherwise almost flawless presentation, the graphic explaining what was included in the savings account offered an unknown product called ‘bouls’. Auto-generated picture and video content also generated some noticeable howlers – for example, a voiceover said the fashion brand had been founded by three friends while showing four people on screen.
- The human touch may be needed to achieve cut-through
While all groups were able to generate a brand in time allotted, a common refrain was that the logos created by the AI were derivative of other major brands – the running shoe team, for example, noted that their logo looked similar to that of bike company Peloton.
As one attendee described it, there is a danger that different brands using AI generate essentially the same ideas and there is no cut-through.
However, a lot of the ideas were surprisingly passable as genuine advertising.
- AI is most useful as part of a process rather than in the final results
Relatedly, while some noted that AI had hugely reduced the time to generate creative, these were probably most useful as prompts for other ideas, or as a way to quickly mock up an idea, rather than to be relied on as the final results.
Some teams reported that using AI provided a quicker path forward through decision-making, removing some of the brainstorming time or ego attached to individuals’ ideas. The fact that the groups were able to produce such developed campaigns in an hour is a testament to this.
