Sign In

Peter Manniche Riber

A Head of Digital Learning @ Novo Nordisk

Peter Manniche Riber is a professional in the field of digital learning and technology. He currently serves as the Head of Digital Learning at Novo Nordisk, a position he took on June 1, 2019.4 With a background in learning and development (L&D), Riber is passionate about making L&D accountable and ensuring that people receive the right training.6

Professional Focus

Riber's work revolves around several key areas:

AI in Learning and Development:: He has been involved in experimenting with AI coaches to provide accurate recommendations for employee development and internal career paths.15

Digital Education and EdTech:: As the Head of Digital Learning, he likely oversees the implementation of digital tools and strategies for education within Novo Nordisk.

Personalized Learning:: Riber has shown interest in learning personalization, indicating his commitment to tailoring educational experiences to individual needs.3

Expertise and Interests

Peter Manniche Riber demonstrates expertise in:

  • Upskilling and digital education
  • AI applications in L&D
  • EdTech implementation
  • Career development and guidance

He holds a Master's degree in Education from Harvard3, which likely informs his approach to digital learning and educational technology.

Professional Contributions

Riber actively shares insights and experiences from his work on LinkedIn. He has discussed:

  • The need for simplification in learning and development processes1
  • Experiments with AI-driven coaching and career guidance2
  • The importance of context in learning initiatives2

His posts often reflect on the intersection of technology and learning, showcasing his commitment to advancing the field of digital education.

Highlights

Aug 15 · linkedin

Reflections from AI Career/Development Coach experiment with a limited audience - 2nd edition.

Human vs. Machine. Expectations to value of response, recommendations and guidance. Can the AI Coach stand alone? It's very interesting how we tend to be tougher on technology's ability to accurately guide or help us, than we are towards a human being guiding us. We rarely fairly (I should write lyrics :) evaluate the alternative.

In the experiment our AI had the overview of more than 1.200 job positions, with all the content in them, thousands of training and learning options and details of those, external databases of advice from a massive pool of data, etc. And in the user's completely private space it would calculate against your own career history, internal development plan, job description, skills, etc. Then come up with recommendations, with argumentations behind them. Imagine if it knew all job descriptions across the company, organisational history/movement, combined with department strategies, etc. on top of that 😅

Which human being can store or process that width of knowledge so quickly? Even at all? We are naturally limited to our bandwith, network internally, "top of mind" and I guess we'll have to admit, that our recommendations in theory should be limiting the people, we guide compared to what an AI could.

So why do we immediately discard the machine generated recommendations, when we feel they don't hit the spot? Maybe some of the answer is hidden in context again (see previous post). The AI doesn't know enough about the person, company, culture, environment, gut feelings.. to actually make it accurate enough. Maybe we are not ready to stand corrected (think differently) by a machine? :)

Our experiment has seen the majority of participants express a feeling of "surprised impressed" as to the immediate accuracy of the AI generated recommendations. Fantastic! But we also got also the opposite feedback of "that's a crazy recommendation" - especially because it recommends things, we think it "should have known". "Should have known" is a huge and extremely complex contionous coding exercise, that requires updated and accurate databases, comprehensive training and fine tuning, etc.

Some of you asked in the last post thread about "replacing human interaction" and let the AI be the coach. I believe at least our experiment - and also intentions btw - concludes that a wonderful combination between a machine with a broad data overview and a human being with contextual sense is a dream combination. At least until AGI ;-)

Next week 'll put something together about building career plans in a coach tool like this. Hope this is useful, and would love to read your reflections below.

Aug 6 · linkedin

For the last 6 months, we've been experimenting with AI in a closed environment with around 50 people, to see if we could improve career development recommendations to people, with the help of an AI Coach. Really help people navigate accurately and get what they need (advice, a coach, training, job recommendations, career paths, etc.).

We created a prototype, which can do 3 things basically (but in reality the output could be more variations depending on prompt):

  • Skills development: Provide advice, links to programmes, training, articles, videos, provide development plans, etc.

  • Job Recommendations: Match people with open job opportunities internally or advice on what positions might be relevant for them in the future, based on their profile and desires.

  • Find an expert: Find someone to link up with (who accepted that role, of course) internally, within a field of expertise.

I'll share a few learnings over the coming weeks, from our process, which could maybe help the community:

  • Context is so incredibly important.

The agent must know or get to know the context of you and your workplace. You know that feeling, when an automated chat feels irrelevant? Not sure whether to laugh or cry? That we must avoid at all cost, because that's where people start zoning out and stops taking things seriously. This is the difficult part and where AI is not just plug and play. A workplace, your role, you, your expectations.. are complex things, which constantly change and consists of A LOT of data points. The trick seems to be finding the right balance in what you preload (data) and which questions the agent can ask to get enough context to answer accurately.

I thought it would be enough just to preload as much as possible. The issues with that though, are data quality (accuracy and update ratio), GDPR/Privacy concerns and .. that the data actually doesn't always give the right context. It will get you some of the way, but to avoid some of that head shaking among users, we need get more situational context from people.. "What are you passionate about currently?", "What projects/tasks are you working on right now?" "What are you curious about?" "Dreaming of?"

A couple of examples of data points:

  • Is your resume actually you? Does it really accurately tell what you master? What magic you can bring to the table, based on your experiences through life? Probably not the full picture. Which means it's a good starting point to kick off the conversation without being completely blank - but it does need more context. And ask some of the juicy question - for people to get the juicy (valuable responses).

  • Job descriptions/adverts: do they really describe what you need to master in the position? In the context of how you do things in your company? Or are they polite and fairly generic descriptions? Again, the agent might need to ask some great questions - and be trained to interpret our corporate language and make it human :)

Character limit reached :D

Related Questions

What are Peter Manniche Riber's main responsibilities as Head of Digital Learning at Novo Nordisk?
How has Peter Manniche Riber's career evolved since joining Novo Nordisk?
What are some key projects Peter Manniche Riber has led at Novo Nordisk?
What is Peter Manniche Riber's educational background?
How does Peter Manniche Riber incorporate AI in his digital learning strategies?
Peter Manniche Riber
Peter Manniche Riber, photo 1
Peter Manniche Riber, photo 2
Add to my network