Watterson Streamlines Media Analysis and Sentiment with AI

Solution Overview

The Client

Watterson is a boutique public relations and marketing communications consultancy based in North Sydney, Australia. It has supported clients in the technology, sustainability, business, building and consumer spaces for 30 years and has been a long-time supporter of the technology industry, having been a founding and ongoing sponsor of the IT journalism awards for 22 years and counting.

Watterson monitors coverage of its clients, their competitors, and related topics in the media to highlight how they’re being perceived, the sentiment of their clients versus their competitors, and also flag if negative coverage has emerged. But the process in doing so was becoming cumbersome and expensive, particularly amid rapid growth for the agency, held back by manual process and data entry. 

It began exploring whether Generative Artificial Intelligence (GenAI) would be able to assist.

The Challenge

In the past five years Watterson’s client base had grown from 14 to more than 50, with each requiring different levels of media monitoring – some need broadcast coverage, others online only, some print and others wanted share of voice monitored to track themselves against their competitors.

Watterson leveraged a combination of external data streams (manual searching/Google alerts along with two third-party suppliers) to ensure it achieved a holistic view of coverage for each of its clients. But with that came the issue of varied formatting, and manual data entry was required to encapsulate all clips and sentiment into a single source of truth, often an Excel spreadsheet that would then be exported to a PowerPoint for reporting processes.

As the company expanded, it was clear that this task needed to be automated to allow Watterson’s consultants to divert this precious client budget to value-adding consultative tasks.

“We realised that if we could find a way to reduce this administrative work and associated third party expenses for our clients, we would have more time and budget to deliver value for clients and media,” said Cameron Wells, Director, Watterson.   

One of Watterson’s clients is Decision Inc. Australia, which had recently put together a succession of AI applications for clients in the banking, media and retail space, leveraging AI, LLMs, Machine Learning and Automation. So, Watterson began speaking to Decision Inc. about working together to come up with a way of not only automating its process but build something intuitive that they could own and manage internally.

Furthermore, it expanded the remit to leverage an LLM to help summarise stories and analyse the sentiment of a client in the article, be it positive, negative or neutral. It also developed a function to quickly develop exportable reports including sentiment analysis, share of voice and annual coverage comparisons.

Within weeks, the company had developed a PoC, including design, which Watterson approved.

The Outcome

Within months, Decision Inc. Australia had developed a cloud-based browser application leveraging AWS Bedrock Services including Cohere, a foundation model with access to the internet. Watterson could now reduce the burden of data entry required for its 50-plus clients, allowing them to maximise budget and offer clients more bang for their buck, while reducing the reliance on – and subsequent cost of – third parties to report coverage.

“It’s early days and we’re still in the iterative phase of real-world deployment, but we should be able to expand the scope of work by an additional activity or more per month with this tool taking a ton of heavy lifting off our plate,” said Wells. “That’s a huge upgrade in value for any client, regardless of the size of their PR budget.”

“Watterson can also now do whatever they want with this application, rather than be tied down by third requirements and the additional costs that come with it,” said Decision Inc. Australia CEO, Heke.

For instance, third-party suppliers often have restrictions on the number of clips or the type of clips they track and pass to the agency; any coverage which went beyond that would incur an additional cost. They had limits on the number of clients Watterson could ask them to monitor for. 

“Now, they can ramp up a new line of tracking and analysis for anything really,” said Heke. “They can analyse coverage and AI-driven sentiment of not just competitors, but of general trends – such as GenAI or even the downturn in the building and construction sector – and offer that to clients with the click of a button and without impacting their budgets.”

Additionally, the speed to report coverage will now be at the discretion of Watterson – often, to get the full breadth of coverage, Watterson would have to wait for the third parties to deliver an email listing each clip, which was often the next day and missing key pieces of coverage. Now, with a click of a button, Watterson can get a coverage update whenever they need it.

“The AI aspect of the tool provides a detailed visual of the coverage, an easy-to-read analysis of each story, and an incredibly accurate accounting of sentiment,” said Heke. “It’s an aspect of client work that will soon just get done without Watterson having to worry about it, helping them achieve the outcomes for their clients and media faster, without headaches and at a reduced cost.”

LaTEST AI Case Studies