Like many businesses out there right now, we’ve been racking our brains internally around the implications of generative AI on our craft. And truth be told, we’re all pretty excited about what the future holds for communications and PR.
But at the same time, we know that our clients will be wondering how our work with them is to be impacted. There’s a number of considerations for Bud in this area – whether it be practical, legal, or ethical – and we wanted to create a practical framework for how we’ll be moving forward.
During a recent company off-site we sat together as an agency and put the first steps in place to create our framework, which you’ll find below. The aim of the following is two-fold. First and foremost, it’s about giving clarity to our clients, so that they can continue to place their trust in our work, specifically around how we handle their sensitive information.
Secondly, it’s about starting a conversation across the agency landscape. We strongly believe generative AI is an incredible value-add to today’s top-class PR pro, whose ability to create meaningful relationships and tell powerful human stories has only been enhanced by this technology.
The team at Bud uses generative AI to…
- Act as a sparring partner when considering interesting story angles for a byline or Q&A interview, e.g. when creating a media pitch. In such instances, we create what we think is an engaging pitch and ask a tool like ChatGPT or Bard to devise additional ways we might frame the story. We do not input any information that’s confidential and thus not in the public domain at that time.
- Act as a sounding board when considering potential risks/challenges associated with a campaign idea. E.g. we’ve devised an approach for an internal communications campaign and are conducting a risk analysis, or we’ve proposed ten steps to restructure an area of our project/account management and want to sense check our process.
- Create a high-level outline for a piece of content, e.g. if we’re writing an op-ed about a specific theme and are considering the best way to frame our arguments. This involves asking ChatGPT or Bard to recommend an article structure in line with a specific word count. To create this outline, no information about our client is shared and the suggested arguments are always our own.
- Summarise online articles (one, or multiple at once) and help accelerate our understanding of a current news topic or trend.
- Condense the findings of a whitepaper (e.g. in PDF format) to create a set of takeaways for ourselves or our clients.
- Act as the first point of reference to obtain data points that inform our research, which we then verify via a third-party source. E.g. to find out the size of the APAC video game market, we would consult Bard then verify the output via a reputable online source (such as Newzoo or another video game research firm).
- Create suggested copy and caption texts for our own Bud LinkedIn and Instagram posts, or for our newsletter Bud in Brief.
- Create a list of headline suggestions for a client byline/op-ed, where we spar with ChatGPT on what phrasing is most coherent, succinct and engaging.
- Write the most coherent, concise and action-oriented email responses, whether to clients, partners or journalists. In such instances, we use the AI recommendation system in-built within Gmail, our chosen email provider.
- Craft a snappier headline if we’re about to go live with a client’s byline/op-ed, providing the headline contains no confidential information.
- Proofread/spell check our work. In such instances, we use the functionality within Google Docs (our chosen document creation software). We also have a subscription to Grammarly.
- Optimise specific sentences in a press release, interview or byline, if the sentence is unclear or could be written more succinctly. This is working on the basis that the sentence contains no confidential information.
- Support outline ideation for presentation decks, so that we can more quickly and efficiently present information to clients, e.g. in the case of media coverage reporting.
- Elevate the quality of visual communication within our decks, e.g. if we feel a presentation isn’t clearly communicating results to a client and want alternative suggestions on how to structure the deck. This is where we use the AI features within our chosen presentation software, Canva.
- For all of our media coverage reporting functionality, we use our chosen provider Cision.
We believe our framework is only the beginning and as such, this represents a living document that will update over time. However, the four key questions our team will always ask before applying any of the use cases above are:
Does this usage
- Respect the trust (including any formal confidentiality agreement) our client has placed in us?
- Represent our client’s best interest?
- Positively impact the overall quality of our work?
- Give us greater bandwidth to flex our attributes as top PR pros?
Onwards and upwards!