Resetting Our Expectations About AI in Social Media Intelligence with Forrester

Don't forget to share this post!


Last week, we hosted a webinar with leading market research company Forrester to discuss the many uses - and limitations - of AI in sourcing compelling insights for brands and businesses.

If you didn’t get the chance to join us in real-time, you can take a look at our webinar in full here.

In this post, we’re here to provide a recap of our discussion on the state of play with AI in social media intelligence.

We’ll explain why human expertise is so crucial to unlocking the true power of AI as a tool to understand fans, consumers, and competitors. We’ll also take you through some use cases illustrating the potential for companies to combine AI with human expertise.

Ready? Let’s dive in.

Closing the opportunity gap with AI

As Jessica Liu, Senior Analyst at Forrester Research, explains, Forrester’s goal as a company is to work with business and technology leaders to develop customer-obsessed strategies that drive growth.

A big part of that strategy development is using AI to unlock helpful and influential customer insights.

But what does that really mean in practice? And how exactly can we use AI in marketing and social listening platforms?

Volume, velocity, and complexity: AI in practice

For Jessica, using AI is a question of necessity. “The reality is that marketers need help. Human cognition simply can’t cope with the volume, velocity, and complexity of modern customer engagement - especially when it comes to social media sources.”

“Marketers rely on technology to leverage massive amounts of data to make billions of daily decisions. The amount of information people need to process simply exceeds human cognitive capacity.”

“This creates a giant opportunity gap,” says Jessica. “And that’s where AI-powered solutions come in.”

If this all sounds a little ominous, don’t worry: AI isn’t about replacing human roles.

“We’ll all about complementing what marketers are doing right now,” says Jessica. “Alongside human expertise, AI can really help to close this opportunity gap.”

How AI closes the opportunity gap

AI helps to close the opportunity gap between data volume and analytical capacity by enabling:

  • Efficiency - AI helps marketers to process larger amounts of data more quickly
  • Smarter decisions - AI leverages data to help drive better decisions
  • Speed - AI helps businesses keep up with the pace and complexity of data
  • Continuous performance improvement - AI systems learn faster and improve in real-time
  • Customer journey optimization - AI helps to tailor customer interactions for specific users

As Jessica notes, however, AI isn’t a product or a solution on its own. Instead, AI technology forms a foundation for how a lot of marketing tools and products work.

To form this foundation, AI technology needs to be able to do the following three things:

  1. Sense - being able to quickly ingest and mine troves of diverse social data
  2. Think - utilizing algorithms to make inferences based on this data
  3. Act - taking action on the social data processed

With these three functions, AI has the power to help marketers interpret data, find new insights, and make better decisions.

"AI isn’t a product or a solution on its own." - Jessica Liu, Senior Analyst, Forrester

The true promise of AI

The real promise of AI, according to Jessica? Alleviating the burdens of us long-suffering humans.

“AI-enabled social listening platforms are really striving to alleviate the human burden of processing information,” she says. “These platforms unlock helpful insights, without needing huge amounts of time and effort from people.”

However, as Jessica notes, AI has limited capability to make effective decisions on its own. As we discussed in our previous blog post, the technology requires human input, oversight, and context.

AI is nothing without human expertise

A lot of companies are falling over themselves to weigh in on the transformative potential of AI, and are promising the world when it comes to techniques like machine learning.

In reality, however, we agree with Forrester: while AI technology can help save a lot of time and effort, it isn’t perfect on its own. What matters most is human expertise.

“The reality is, current social listening platforms aren’t fully-powered AI technology solutions just yet,” says Jessica. “Human users still need to establish the rules, train the system to use data, and maintain the platform. Without these things, the system can get things wrong.”

A great example here is machine learning, where algorithms are used to source and distinguish complex information. “In our experience,” Jessica says, “it probably takes 20-30 social posts with brand-specific information for a machine learning system to make accurate decisions.”

That’s why, at Linkfluence, we add a layer of human expertise and practical judgement to get the most out of AI.

By combining AI-based techniques and technologies with specialized social data researchers and analysts, we can find the most compelling insights for our clients, and can save significant time and effort in gathering and processing vast amounts of data.

And as Forrester have found, this combination of AI and human expertise has a wide range of potential uses.

Forrester’s AI social listening use cases

As Jessica explains, AI technology is a versatile tool for any marketer to have at their disposal.

But how does this technology help in practice?

According to Jessica, there are a bunch of different use cases AI can augment and improve:

  • Public relations: AI can help flag potential brand crises, helping brands to proactively manage risks to company reputation
  • Brand health: AI-based platforms can track brand health over time, applying machine learning to map social conversations to brand pillars or corporate values
  • Risk management: AI technology makes it easier to detect bots and fake users designed to deliberately manipulate online exchanges and conversations
  • Customer service: AI can make it faster and easier for customer reps to find customers seeking help, and to know exactly the kind of help they need
  • Trend detection: With AI, product and R&D teams can uncover new trends in product innovation, and can then take advantage of these trends
  • Media planning and buying: AI makes it easier to identify audiences and consumer segments to speak to

Out of these use cases, we want to highlight a specific use case of trend detection we've done with our client, Danone.

Trend detection: An AI + human success story

As our CEO Guillaume Decugis explains, getting the most out of AI is all about combining machine technologies with human capability.

“We believe AI delivers a lot of value,” he says. “But there’s also a last-mile component that requires human expertise.”

We can see this in our use case with Danone, one of the world’s largest food companies.

Finding the next star ingredient

For a client like Danone, ingredients are everything.

“The question about ingredients affects a lot of things,” says Guillaume. “Food has a long innovation cycle, and ingredients are one of the key drivers of customer preferences.”

With Danone, we wanted to find a better way to predict trends when it comes to popular foods. This lead to our key question: Can AI help us to identify tomorrow’s star ingredients?

For Danone, answering this question accurately could lead to significant advantages in terms of sourcing, supply chain, and production. So, we looked for ways to combine social data with AI and detect emerging trends within consumer conversations and exchanges.

“We believe AI delivers a lot of value,” he says. “But there’s also a last-mile component that requires human expertise.” Guillaume Decugis, CEO, Linkfluence

Discovering trends and patterns with AI

To start, we needed to refine the problem to make it more workable. By starting with a list of over 800 ingredients, we could use social listening to monitor the volume of conversations and data over time. We could then use machine learning to extract trends and patterns.

From there, we selected a shortlist of ingredients that were likely from a statistical standpoint to become more popular in the future. Once we had this shortlist, we could validate these ingredients with a qualitative review, extracting a smaller number of promising contenders.

At the end of this process, Danone could focus on a shortlist of ingredients, rather than looking at over 800 options. From there, they had a more manageable set of decisions to make.

So, which ingredient came out on top?

And the winner is...

One of the ingredients included in the initial list was Ashwagandha, a plant product used traditionally in India. With our model, we were able to pick up meaningful consumer discussion about this ingredient, suggesting that the future market could be significant.

Ashwagandha, identified by Linkfluence as a potential superfood. Source: Banyan Botanicals

As a result, explains Guillaume, Danone have now launched new products based on the plant Ashwagandha, and are marketing this to customers.

AI + human expertise = :)

As with Jessica’s use cases, the key success factor in the Danone example is about overcoming the limitations of AI by combining it with human market expertise.

“This will be the model in the coming years,” says Guillaume. “AI algorithms will improve and become more sophisticated, but we think the best way to leverage this technology and put social media data to use is by combining it with human and market expertise.”

“Companies can either find this expertise in-house, or can enlist help from a company like Linkfluence to structure and interpret this data.”

Find out how to put AI to work

As we’ve seen, AI is an amazing tool - but it still needs a human in the driver’s seat.

For all the time AI technology can save in selecting and analyzing vast amounts of data, if you’re not asking the right questions or setting the right parameters, you won’t get what you’re looking for.

That’s why, like Forrester, we’re so interested in combining AI technology with human expertise. We think that’s the key step in unlocking valuable and actionable market insights to help our clients serve their fans and customers.

If you’d like to know more, be sure to check out our webinar here.

Forrester webinar: What AI Won't Tell You

Thanks again to Forrester, and to Jessica Liu for joining us for this great discussion!


Don't forget to share this post!