AI is set to transform marketing, or so it says. But what progress are successful scale-ups making to develop personalized relationships at scale and at a limited cost? Can they automate tasks and gain deep customer insights through predictive analytics, resulting in higher efficiency and conversion rates? And create high-quality content without missing the mark?
In the run-up to the SaaS scale-up event B2Brocks I asked a number of speakers about AI. I am curious how they apply AI and what their experiences are in practice.
From content plans to translations
First I speak Julia GoellesVP Marketing at Parloa. Parloa is a leading Conversational AI SaaS platform that empowers businesses to have better conversations with customers.
In such a company, it will come as no surprise that all team members – from the content team to field marketing and performance marketers – are using AI to support content creation. For this, as in many companies, ChatGPT and Open AI are used. This starts with generating ideas for content plans to optimizing texts or checking translations with the help of AI.
Big impact on content
At Parloa, the IT team is deployed to use generative AI safely and reliably for content creation, by ensuring it is not used to train data. While it may seem obvious, Julia sees content as one of the areas where AI has a big impact in marketing: after all, content fuels every marketing function and project.
While ChatGPT and OpenAI are used for textual content, ElevenLabs and RaskAI are being tested for video translation and subtitle creation. Julia points out: “We have also explored the dubbing capabilities of these tools. There is still a lot of manual work involved, but I am sure that these features will improve significantly in just a few months and make marketers’ lives much easier.”
To make marketing workshops and campaign planning meetings more efficient, Miro is widely used. A handy AI-based tool that allows you to everyone’s input can cluster and put team members’ thoughts into actionable categories. “This has proven to be incredibly helpful,” Julia says.
Pitfalls in AI
When I ask Julia about 3 possible pitfalls when using AI, she mentions the following points of attention:
- Formulaïsche content
More and more companies are using AI to generate content, causing similar, repetitive content to flood marketing channels. - Data Privacy and Security
Unfortunately, sensitive data is sometimes used to train models. A pitfall that Parloa itself is very aware of. - Bias and lack of diversity
Julia sees this as one of the major risks of generative AI, as it reinforces existing stereotypes through its reliance on training data. “As marketers, we are responsible for our company’s public communications, and therefore it is crucial that we address diverse perspectives and promote pluralistic opinions.” To address this issue, Parloa has taken several initiatives, including a highly engaged Responsible AI Board, in which Julia actively participates.
Predictive analytics
I will speak next Antoinet van Dalenoriginally Dutch, but working in Paris since 2007. The past 6 years at Cegid, a large SaaS company that offers various solutions. Where Julia focuses more on new business, Antoinet mainly looks at existing relationships.
Customer Success Management (CSM) involves proactively guiding and supporting customers to ensure they get maximum value from a product or service, resulting in higher customer satisfaction and retention. The primary application of AI that Antoinet uses for CSM is the use of predictive analytics.
“AI can provide insight into characteristics that lead to certain functional applications of our SaaS services. By finding similar characteristics in customers who are not yet using these applications, you have a powerful starting point to let those customers do more with their SaaS product.” This can relate to using the current tooling more productively, but it can also be a great opportunity for up-selling or cross-selling.
Better insights into churn
The same goes for gaining better insights into churn. What characterizes customers who stop using SaaS services? For example, do they not use certain functionality? Are certain integrations missing? Are there major differences in the use of the solution between departments? These are all hypotheses that can be tested by AI and with which you can potentially reduce churn.
Antoinet indicates that they use the current functionality in tools such as Salesforce to make such analyses. “But we are still in the early stages, the tools are still in full development and offer more and more possibilities”. An example of a tool with AI analysis possibilities for CSM is Staircase AI van Gainsight.
As a point of attention, Antoinet, like Julia, indicates that the results from a predictive analysis do not offer 100% certainty. It remains a prediction, and therefore a starting point for entering into a conversation with a customer. “AI certainly does not replace the personal contact with customers that is part of customer success. This is and remains a field that revolves around personal relationships”.
Combining data with a story
Thirdly I speak with Heather McDanielshe specializes in data storytelling. Data storytelling combines data with a story and visualizations to convey insights clearly and captivatingly. This involves more than just showing numbers. By adding a narrative, the data becomes more understandable and relevant.
A good example of data storytelling is the Goalkeepers Report from the Gates Foundation, where the lives of Bill and Melinda are contrasted with the life of a girl born in the Sahel.
You might expect data storytellers to be wildly enthusiastic about AI. But at the start of the conversation, Heather immediately shares her reservations about the possibilities of AI. It is an important new technology, but it is not going to change the essence of marketing overnight. “In that respect, we are in the hype phase, as we saw earlier in the late 90s with the rise of the Internet.”
I ask Heather how AI can help with data storytelling. “With data storytelling, you naturally deal with a lot of data. AI can be very practical and helpful in that, for example in interpreting data, cleaning data and making predictive analyses. But you have to be very careful with AI when adding the narrative.”
Recording stories remains human work
Capturing the personal stories behind the data that fit your audience remains human work. Because although AI can help compare characteristics and features of a large sample of profiles, in-depth interviews give you much more context and the personal story. And that is essential for creating the story that fits the narrative.
Once you have recorded the narrative and personal stories, you can use AI to see how you can enrich this with data. After all, you can have AI search the data from a certain angle.
Points of attention
In the various conversations, the central finding is that AI is complementary, not a replacement for certain marketing functions. Marketing must first have a clear goal and then see how AI fits into that specific goal.
“If you have limited resources, I would recommend startups to use AI primarily to create efficiencies and reduce repetitive work,” says Julia. Think of automations, personalizing outgoing emails, summarizing all types of conversations or creating documentation. A tip to start with this is to explore the AI features in existing tools you already use, such as Notion, Slack or Google.
The various conversations also consistently show that users need to be aware of the pitfalls surrounding the reliability of information and creating authentic and creative communication. In addition, it is important to carefully consider the ethical and social implications, such as the reproduction of prejudices, that their use of AI can have.
AI is not a marketer
From the conversations I conclude that AI can really enrich the life of a marketer, but that you should not see AI itself as a fellow marketer. Marketing remains human work, communication is about developing personal relationships. And AI is intelligent, but remains artificial.
Source: www.frankwatching.com
