🚀 PR Futurist #114 🆕 GPT-5 Launch ⚖️ AI Ethics Clash 🤖 Virtual AI MP  11 August 2025

How McKinsey is adopting AI and what it means for consultants

The Wall Street Journal look at how the world's most prestigious management consultancy is adapting to use AI. McKinsey now has more than 12,000 AI agents its consultants use across an array of activities to do them faster and better.

Its AI uses include helping consultants to build PowerPoint decks, taking notes and summing up interviews, creating first drafts of research documents for clients. Another bot checks the logic of a consultant’s arguments, verifying their flow of reasoning makes sense.

Its most popular agent is the 'McKinsey tone of voice' that helps employees write in a classic style—language the firm describes as sharp, concise and clear.

All of these use cases are ones that can be adopted by PR agencies and in-house communications teams (and that we're helping them to do). The critical steps are: identifying the initial tasks and use cases, choosing and implementing AI tools, developing team bots, agents and prompts, and wide-spread training to ensure effective AI adoption.

Find out how CommsTransform™ can help you do this.

wsj.com  

Number 10 invites the influencers

A mainly positive, but slightly sniffy piece in the FT on Labour hosting its first Downing Street reception for between 70 and 90 'content creators' or 'influencers'.

It's a good example of why influencers aren't just for brands, but can be just as relevant to corporate communications. The FT quotes the 2025 Reuters Digital News Report and reports that in Britain, the share of people getting their news from social media at least sometimes has risen to 39%.

Mainstream media no longer cuts it, for any stakeholders.

Picture from the Instagram account of Love Island star turned revenge-porn campaigner Georgia Harrison. It shows how well planned the event was as presumably the sculpture of a twisted red telephone box was placed outside Number 10 for the event to help create Insta ready moments.

ft.com  

Unilever has built an AI studio to create content in-house

A case study looking at how Unilever has "been building a generative AI assembly line for its digital creative."

For the last year Uniever has been building the Beauty AI Studio: a bespoke, in-house system for its beauty and wellbeing business. It is now being used in 18 markets, including the UK and USA, to make content for paid social, programmatic display ads and e-commerce use across brands such as Dove Intensive Repair, TRESemme Lamellar Shine and Vaseline Gluta Hya.

digiday.com  

MP creates an AI version of himself as the UK's ‘first virtual MP’

You might have seen the stories about the Labour MP who has unveiled what he says is the UK’s ‘first virtual MP’. It's fair to say it has had a mixed reaction. It has been discussed on many of the multiple PR, public affairs, politics and policy groups I participate in. Try the AI Mark Sewards here.

I've had several people contact me privately to ask for my thoughts. It's a hard one to comment on as Mark is a personal friend. I know his family. I campaigned for his election. My wife is a councillor in his constituency. I've stood to be a councillor in a different ward in his constituency. I didn't advise him and didn't know the company behind it was based near me until this story.

My honest take is the idea and concept is brilliant. Using AI to assist with dealing with simple daily queries and questions from constituents is a great use case for AI. Personalising it with a clone of Mark's voice makes it even more useful. Finally, he did it for the best of reasons - to support a start-up business in his constituency, to show how a local MP can innovate and experiment, and to to provide a better service to his constituents.

The most interesting aspect of this is the start-up Neural Voice is focused on voice which means beyond this interesting experiment its technology has lots of potentially practical uses. The examples on its website include being able to make telephone calls to a sales assistant, customer support representative, recruitment assistant and travel consultant.

Most AI personas focus on chat or video. Utilising voice and focusing on telephone calls opens it up to different use cases.

But, you can probably sense a but coming on.

It could and should have been done in a different way. The backlash was inevitable and predictable. It couldn't have been avoided. It could have been massively mitigated. There are lots of complex issues to be navigated around social licence, transparency, trust, ethics, privacy, security, hallucination, risk, disinformation and more.

All of these should have been both considered and critically explained publicly as part of the launch.

If you don't have time to experiment with AI Mark you can watch this short video where Northern Agenda editor Rob Parsons puts it through its paces... with mixed results. It's evident from Rob's test that the experiment was probably launched a little hastily as its answers show it is lacking training data in its knowledge base that could have made it more powerful, without compromising privacy or security.

The issues raised by the AI MP experiement is why we invest time in developing plans and policies for clients. The purpose is to ensure innovation that delivers practical results in a safe, ethical way.

We've been exploring similar ideas for clients and prospects, but need to ensure the correct plans, policies and safety is in place before doing anything public.

independent.co.uk  

Swedish PM criticised for using AI

After the AI MP another example of the necessity of social licence for AI is criticism of the Swedish PM for using AI as a sounding board.

Ulf Kristersson told the Swedish business newspaper Dagens industri: “I use it myself quite often. If for nothing else than for a second opinion. What have others done? And should we think the complete opposite? Those types of questions.”

It's similar to the flack directed at UK cabinet minister Peter Kyle for using AI. My comment was I'd be more shocked if the minister responsible for AI policy wasn't using AI.

All of these examples show how critical it is to understand social licence for AI and always have the right policies and processes in place.

theguardian.com  

Explore The Economist's annual The World Ahead using AI

The Economist has made the 70 articles and analyses in its annual The World Ahead available in Google NotebookLM.

Not only can you read the articles and analyses but you can also study them with all the usual NotebookLM functions: You can create an AI-generated podcast, read it as a FAQ, view it as a mindmap, create a study guide. Or just ask it questions.

It's a great example of making content more accessible to stakeholders. Corporate comms can upload and share their own content such as annual reports, research and white papers, ESG or DEI reports and more.

Instead of forcing stakeholders into your straitjacket you create a user experience they can personalise to their own needs.

purposefulrelations.com  

We need more corporate comms videos like this

Are you bold, innovative and effective? Vattenfall's corporate comms team clearly is. It teamed up with Samuel L Jackson to create this knock-out video will a killer script. And yes, Mr Jackson drops the F bomb.

The climate emergency is real, but solving it requires massive disruption to peoples' lives. We all have to stop fearing and starting embracing nuclear energy. We all have to accept the inevitability of unpopular wind farms dominating our landscape. I say unpopular because while wind farms are popular with people using their green energy, they tend not to be with the people who live where wind farms are built.

If we are to change minds so people accept these changes, we need bold, innovative communications like this. I'm not privy to the data, so can't comment on its effectiveness, but I'd put money on it working better than a bland, factual defence of wind farms.

creativeboom.com  

Hertz's AI system that scans for damage on rental cars dings innocent people

Who would have guessed that an AI system to check for damage on rental cars would make mistakes and penalise innocent customers? Anyone with a modicum of common sense?

Hertz is facing criticism because an AI system to check rental cars for damage is flagging false positives and penalises customers for damage that doesn't exist.

This is a good example of why corporate comms should always be in the loop when companies implement any AI. It's why we keep stressing the importance of social licence and ensuring AI is implemented in ways that win the trust of stakeholders such as customers and employees.

Apparently, Hertz failed to do this and is now feeling the heat?

Is your corporate affairs team always in the loop on how your company uses AI?

futurism.com  

📺 AI indexes 15 years of TV news in under an hour

Monitoring text and even images is easy, but video has always been more challenging. Google's Gemini 2.5 Flash has a near-human ability to watch video and analyse what it's about.

The GDELT Project at Vanderbilt University used Gemini 2.5 Flash Thinking to process 14,749 evening news broadcasts from three major US networks, identifying more than 154,000 distinct stories. Each story was tagged with geolocation, sentiment, framing, and narrative data—creating a searchable index of 15 years of coverage. The entire process took just 39 minutes and cost $154 in AI tokens, showing how AI can scale tasks that previously required decades of manual effort.

This has huge implications for media monitoring and analysis.What's remarkable is that no data was used to train, tune or otherwise contribute to any model. It used the standard off-the-shelf ability of Gemini 2.5 Flash Thinking.

It shows how AI can move beyond keyword search to story-level analysis. For communications professionals, it offers a new way to track media narratives, assess sentiment, and understand how stories are framed over time.

It creates a huge challenge for tradtiional media monitoring, analysis and monitoring companies as what was previously complex and expensive can now be achieved in-house.

gdeltproject.org