When I spoke with Guillem Cardenal Basté, Social Selling Lead at Merck Life Science, on our LinkedIn Live, one thing stood out: AI is not here to replace people. It is here to help them show up with more confidence.
Merck’s journey began modestly, a Teams group sharing LinkedIn Sales Navigator tips, but has grown into an ecosystem where employees can experiment, contribute, and publish. With custom GPTs now baked into this ecosystem, people have tools to get started, not excuses to stay silent.
Slack’s latest Workforce Index found daily AI use among desk workers surged 233% in six months. Those using AI daily say they are 64% more productive and 81% more satisfied in their roles. Given those trends, Guillem and his team have been experimenting with AI in a way that makes advocacy and social selling easier, safer and more human.
From breaking down the barriers that stop employees posting to exploring how AI could personalise every customer interaction, Guillem’s experience shows what is possible when you put people at the centre of your AI strategy.
How AI can enhance, not replace, human-led social selling and employee advocacy
At Merck, the focus has never been on tools alone. Guillem captured it well: “A lot of people are hesitant to post. They are not sure what to say, how to say it or whether their content is correct. Not everyone is a native English speaker, and that creates a barrier.”That friction was the seed for Merck’s internal GPT initiative. Rather than relying on off-the-shelf tools, Merck built its own GPT platform where employees can try prompt-based drafting, experiment with tone, refine ideas and get over the blank-page fear.
These tools were never meant to do the posting for users. They are meant to give people a leg up, to help them get started and feel confident about their voice.
How AI accelerates engagement and sales outcomes
Merck’s results are already becoming tangible. Guillem’s first GPT, built initially for personal use, caught on quickly once a small group tried it. It then expanded across teams. From that foundation, three primary tools emerged:
- A post generator that made it easier to turn ideas into engaging LinkedIn content
- A personal brand GPT that helped people define how they wanted to be seen, who their audience was, and even map out a calendar of topics
- A LinkedIn profile optimiser that guided employees to strengthen their profiles so prospects could immediately see their value
As Guillem put it: “With those three tools, I think this gives them the ability to start being more proactive on building their personal brand, or at least in another way that they have no excuse not to start.”
For sales, the impact goes beyond confidence. The vision is that AI will help compress sales cycles through more timely outreach, hyper personalisation and predictive qualification. Integrating GPTs with CRM or Sales Navigator data means sellers could anticipate where a lead stands in their journey so they can tailor messaging precisely and intervene when momentum is needed.
This optimism mirrors broader workforce sentiment. Slack’s Workforce Index shows 73% of desk workers believe the fanfare around AI is justified, expecting it to have a substantial impact on their work lives.
Best practices for blending automation with authenticity
One of the most common objections is: does AI drown out my voice? Guillem’s response is thoughtful: “Sometimes people say if I am using AI all the time, ultimately it is not me. But for non-native speakers like me, it helps me communicate more clearly. You still have to double check, fine tune and make sure it reflects how you want to sound.”That blend of machine suggestion and human judgement is key. AI can propose structure, tone or direction. The real control stays with the user.
I like to say: it is not AI content versus human content. It is good content versus bad content. Content that is clear, reliable, well structured and authentic builds trust. Content that is confusing, irrelevant or carelessly assembled does the opposite.
Why ethical use of AI matters in advocacy and selling
Merck’s regulated environment demanded built-in guardrails from day one. Their internal GPT includes strict terms and conditions, for instance no personal information, no sharing of private data and clear usage policies.
Those protocols matter because they provide comfort to employees experimenting with AI, and they ensure the company maintains compliance. Around that structure, Merck layers training, prompt coaching and support so people feel safe using the tools, not policing them.
Trust is not optional, with findings from the 2024 Edelman Trust Barometer highlighting that when innovation is well-managed, people are 12 points more likely to embrace and 17 points less likely to reject AI.
If people find AI outputs unreliable or opaque, they will reject the tool and the program behind it.
How to build a culture where AI supports employee voices
What I found most inspiring in my talk with Guillem was how Merck has transformed AI from a tool into a cultural practice.They’ve done this through regular Digi Hours, focused teams, open forums and internal office hours, Merck's employees get a safe space to explore, test and share. Over time, Guillem’s tools have evolved as he added tone controls, topic brainstorming, article linking and more flexibility based on real user feedback. That step-by-step approach not only keeps the tools relevant but invites the community to own them.
From where I sit, it’s clear that AI is not here to replace people. It is here to make it easier for them to share what they know and connect with the people who matter.
When employees feel confident to speak up, advocacy stops being a programme and becomes part of everyday culture. And when sellers can focus on meaningful conversations rather than struggling with the basics, social selling starts to deliver real results.
That is the role of AI in employee advocacy and social selling, to help people find their voice and use it in ways that build stronger connections.
