AI for Quality, Innovation & Efficiency

Read on to find out how we’re using AI internally at Propellernet to drive rapid innovation and improvements to the quality of our work and processes for, and with, our clients.

AI is driving change around processes and innovation virtually every day at Propellernet. It’s enabling us to deliver a higher quality of work by helping us access and make use of new data, insights and ideas, and it’s incredibly exciting.

 

Having invested early in AI training, everyone at the agency is skilled and able to build GPTs and AI bots to help them with their day-to-day work. Handing over repetitive, time-intensive, un-skilled tasks to what is effectively an AI assistant is freeing up time for our team to spend more of their day doing the high-value stuff for clients – and reducing process friction and frustration at the same time. But, more importantly, it’s helping us to innovate in ways that have never before been possible. We can’t wait to see what the (near!) future has in store.

Here are some of the ways we are using AI in our processes…

How we’re using AI in the Tech team

Faster coding

AI is helping us to quickly sense-check, fix and improve code or mark-up for clients. We'll always give it a human once-over but it's like having a junior coder on the team.

More efficient Excel use

We've found AI to be great for quickly generating complex formulas for use in Excel - something that previously took some time to do.

Taking data from images

Where clients or online sources have important or useful copy within image files, we're able to use AI to extract the text - removing the need for time-intensive transcribing.

Generating Regex rules

Mastering Regex is key for getting bespoke data out of tools like Google Search Console, and AI is helping us craft our rules quicker and more accurately than ever.

Keyword categorisation

Generally, the bigger a (qualified) keyword set we have for a client, the more impact we can have. However, manually categorising thousands of keywords can be a gruelling task. AI can do a great job of a first pass.

Enhanced insights from cloud datasets

AI helps us uncover deeper insights from complex cloud datasets, quickly surfacing patterns and trends that previously required extensive manual analysis.

How we’re using AI in the Content team

Generating content skeletons

Where appropriate, we're using AI to automate the generation of outlines for content pieces we're briefing or writing for clients. Frequently, we'll build on and update the skeletons - but using AI gives us a fast, solid start point.

Research and data sourcing

We find certain LLMs to be really useful in finding data and important information to work into copy we're writing. Perplexity, in particular, is great because it's always very clear with its sources, meaning we can spot hallucinations and verify the data's accuracy.

Identifying potential duplication of content

When we start working with new clients, we often inherit a wealth of editorial content that's been planned and produced by a wide number of people who may or may no longer be in the business. We've built a bot to help us spot content that's potentially being seen as duplicate, so we can make recommendations as to how to tackle and streamline it.

Spotting opportunities to update content

We work with plenty of clients that operate in fast-moving marketing where important news, advice and legislation changes often, which can result in out-of-date or stale content on a regular basis. We've built a GPT to help us spot instances where this has happened at scale, whilst quickly helping us identify what current or new information should be incorporated.

How we’re using AI in the Digital PR team

Custom client-trained assistants

We're always keen to reduce friction between our clients and our teams - and so for each client, we're building a custom GPT that's trained on their tone of voice, existing content and any regulations we need to consider. It's incredibly useful for an initial stress-test of ideas and for checking press releases and content.

Custom image creation

It can be difficult to find images to visualise and sell our creative concepts sometimes. We're now using LLMs to generate custom images to demonstrate our thinking, to great success.

Improving content quality

We've built an EditorBot to help us refine press releases, pitches, and content drafts. We also use off-the-shelf tools like Grammaly to ensure proper grammar, clarity, and identify readability improvements.

Research and data for campaigns

We're using Perplexity to research sources for campaigns, as well as summarising lengthy reports and complicated data sources, and tools like Instant Data Scraper for scraping reviews from websites like TripAdvisor, Reddit and others.

Campaign idea development

We're using AI to brainstorm campaign concepts and angles. For clarity, we use AI in this way to assist with lateral thinking, rather than outright asking LLMs to give us ideas.

Trend identification

We're currently developing bots to analyse online conversations, media monitoring, and news for PR opportunities.

Automating simple data tasks

AI is helping us wrangle data much quicker than ever before, from cleaning spreadsheets, to categorising qualitative data and more.

Refining media lists

We're employing AI to analyse and update media lists for topical relevance/journalist current role/publication status/likelihood of coverage being linking or non-linking, helping us to ensure quality outreach.

How we’re using AI in the Paid Media team

Proprietary platform AI

The major ad networks we work with, including Google, Meta and TikTok, all integrate AI directly in the platform. AI has existed in many forms for a long time here with early adoption of Machine Learning being the driving force behind Automated Bidding and Ad Delivery. Primarily, this enables automated budget & bid delivery, ad delivery and ad creative delivery, based on pre-defined goals and targets.

Writing code

We've been using ChatGPT to write bespoke scripts in JavaScript, to support the management of our accounts. This includes agency-wide budget trackers, security alerts and opportunity analysis

Ad creative prompts

We've used LLM's to provide recommendations of ad creative using our client's website copy as a foundation. We take these recommendations and re-write them using Google's best practice.

Competitor research

We've been asking LLMs to provide insights into client's competitor landscape to help us quickly get a broad understanding of key activity and market trends.

Other ways we're using AI in the agency

Motion AI for efficient workload planning

Our Director of Strategy, Sophie, finds real value in using Motion AI to manage her workload and structure her day. "It's like a live to-do list that's integrated into my Outlook calendar and live updates based on whether or not I've completed a task and it's been a real game-changer for my ADHD brain."

Generating notes from calls and meetings

Much of the team is now using AI tools to summarise what's discussed in internal or client calls and capture actions.

The Propellernet Brand Bot

All of our brand guidelines, previous content and tone of voice documents now power a Brand Bot which anyone in the agency can access quickly to ensure that their output is on brand, both visually and in terms of tone.

A final note on how we're not using AI

We know that AI can hallucinate and that AI-generated content is not always correct, useful or engaging for readers. As such, there have been plenty of instances of Google handing out manual ranking drops to sites that have published a large quantity of AI-generated, poor quality content at scale. We will always review AI outputs and will never publish AI-generated content that’s not been reviewed and edited by one of our team, on client sites or our own site.

 

Equally, we will not share confidential or sensitive client information with Language Learning Models. We will not use voice clone generators or create deepfake videos, which use AI tools to create ‘fake’ videos or audio, such as face swapping or attributing fake speech to a real individual.

We don’t just use AI for efficiencies and innovation – we’re also focused on Generative Engine Optimisation (GEO) and driving visibility for clients in Google’s AIOs and other AI search results. Read up on how we do that and get in touch if you have any questions or would like a hand in getting ahead in the age of AI.