What is your reaction to this ad?
Word of the Day: valiantly
This word has appeared in 41 articles on NYTimes.com in the past year. Can you use it in a sentence?
Science Lesson Plan on ‘Breath Prints’
Explore what our breath reveals about us.
Outdoor Space
Where do you like to spend time outside where you live?
Word of the Day: certitude
This word has appeared in 19 articles on NYTimes.com in the past year. Can you use it in a sentence?
What Is Your Reaction to the Cease-Fire Deal Between Israel and Hamas?
Twenty living hostages were freed by Hamas on Monday under the cease-fire deal, and Israel released nearly 2,000 Palestinian prisoners.
Turning Curiosity into Innovation: How the Dartmouth MEng-CE helped Benson Liu Reinvent his Career

Curiosity has always been Benson Liu’s compass. From earning global credentials to navigating diverse professional fields, he approaches every challenge with a desire to learn and grow. That drive eventually led him to a place where his passions for finance, engineering, and entrepreneurship converged: Dartmouth’s online Master of Engineering in Computer Engineering.
From the moment he discovered Dartmouth’s program, Benson felt an irresistible pull. “We’re the first batch of students. It’s like Forrest Gump’s: ‘life is like a box of chocolates.’ I just had to give it a try!” For Benson, it wasn’t just about earning another degree; it was about “scratching that intellectual curiosity” and building the tools to make a real-world impact.
____________________________________________________________________________
Born and raised in Australia, Benson Liu also spent much of his childhood in Taiwan and Japan. He completed his undergraduate studies at Osaka University in Japan, specializing in pension finance within economics and stayed on to work there after graduation.
Benson’s early career was firmly rooted in finance. He held M&A finance positions and later earned a master’s degree in finance analytics from King’s College London, a technical degree that reflected the growing demand for quantitative skills in the financial industry. Yet despite his achievements, he felt drawn toward more technical challenges. “I wanted something to help me bridge my finance knowledge with something more technical,” he explains.
This curiosity led him from finance to software engineering and eventually into data science. Over the years, he progressed from junior data scientist to senior roles, and most recently, he took the leap to start his own company, helping clinics and hospitals in Japan develop smarter medical devices.
“The public [medical] sector is huge, but many operate on very old systems. Some hospitals even still use fax machines! My engineering background helps, and my finance background ensures a positive ROI. That’s how the different parts of my experience come together.”
Why another master’s degree?
Benson’s interest in Dartmouth’s MEng in Computer Engineering stemmed from his desire to explore engineering more deeply. A few Coursera courses in mathematics for machine learning had already sparked his interest, and when Dartmouth announced its new master’s program, he decided to give it a try.
Although he entered the program with some background in machine learning, he was unfamiliar with many of the courses. “I was like 30–40% sure of what the program had, so I said, ‘let’s see what the other 60–70% is about.’”
Learning without pausing life
One of Benson’s priorities was to continue advancing his career while studying. “I didn’t have to put my career on hold. I’ve always liked to study while I work. The good thing about me is I have a lot of freedom right now, so I wanted to use it as best as possible.” The only adjustment was managing time zones: his first module, Machine Learning, had office hours with professors held at times that were early in the morning for him in Asia. But the program’s flexibility and the responsiveness of Dartmouth professors made it manageable. “The professors at Dartmouth are very accommodating. When I email them, they usually respond in 5-10 minutes.”
Benson has also noticed the increasing international diversity of the program. In his latest module, Natural Language Processing, he is working alongside a teammate from Korea, with another Japanese student recently joining. “There are also students in Europe and Brazil. It’s amazing to see the global network forming.”
Why Dartmouth?
Benson didn’t seriously consider other programs. “It was Dartmouth or bust, in very simple terms. But honestly, I didn’t think I’d get accepted. Based on the UK’s education standards, it’s very hard to go from economics to hardcore computer engineering. But Dartmouth is a liberal arts-heavy institution. They like to look at people holistically. So even though I don’t have a formal STEM degree, I have STEM working experience, and they liked the fact that I was well-rounded.”
The program stood out because it blended rigorous engineering training with an appreciation for students’ broader experiences. “There aren’t many programs offering advanced hardware degrees that also value how STEM intersects with fields like finance. It was a perfect match for me.”
A pivot to entrepreneurship
Benson’s ultimate goal has always been entrepreneurship. When asked about his initial goals for pursuing the degree, his answer was clear.
“When I was working for Coca-Cola in Japan, I was fascinated by embedded systems. Interestingly enough, half of Coca-Cola revenue comes from vending machines, which is basically one large embedded system. And vending machines aren’t as simple as putting a coin in and getting a drink anymore. Now it has analytics, dynamic pricing, and automations. Coming into the Dartmouth program, my expectation was to understand the hardware behind these systems.”
However, Benson has since pivoted his reason for pursuing the MEng-CE. As of two months ago, Benson ventured into the realm of entrepreneurship, and he credits his time in the program for giving him the push to finally get started.
Benson’s new company, fueled by his newfound knowledge, develops small, innovative medical devices for the public sector. “A year and a half later, I decided I wanted to build something that benefits society. It’s a bit of a deviation from my original goal, but that’s life.”
Advice for prospective students
In speaking with Benson, he does emphasize the rigor of the coursework. Brushing up on calculus and linear algebra is a must, and the need for time management is critical to balancing study, work, and personal life. When summarizing his advice into three key points, Benson says:
- “This program scratches your intellectual curiosity. Be prepared to work hard, and it will pay off.”
- “Once you’re in, pace yourself. This is a marathon, not a sprint. Keep your mind clear and don’t overdo it. Even as an active learner, this is one of the hardest programs I’ve done. But if you keep a balanced lifestyle, you can do it.”
- “Be active in the community! I attended a Dartmouth event in Japan and met alumni and even got to meet the president of the university. The power of the Dartmouth network is real, even for online students.”
He also highlights the connections he’s made within the program itself. “I still talk to classmates I met in the first module (Machine Learning) on a weekly basis. Even being 7 courses in, we still keep in touch and keep each other going. This program has a human touch. We’ve all promised to meet on campus at graduation, and I know it’ll be a great time. I think it’s safe to say this program has allowed me to make lifelong friends.”
For Benson, Dartmouth’s online MEng in Computer Engineering has been more than a degree; it has been a catalyst for curiosity, growth, and real-world impact. Whether you’re seeking to deepen your expertise, pivot your career, or launch your own venture, Benson’s journey shows that Dartmouth’s MEng-CE empowers students to pursue their ambitions while fostering connections that last a lifetime. Learn more about the program →
AI’s Role in Simplifying Use Case Development
Scire responsum non significat intelligere quaestionem.
Knowing the answer doesn’t mean understanding the question.
I tasked that inquiry to AI, using a specific language model for learning.
Why is it relevant to creating a use case and then sending it over to the reader or the person tasked with responses?
They see the use case. They supposedly know their system (this is highly questionable as product dev doesn’t always tell them what is in the pipeline to roll out OR the responder often the salesperson doesn’t know the system inside and out and thus relies on someone else, say a solutions consultant, to do their job for them, without any Q/A (Quality Assurance) to validate.
What ends up then is a use case seen by someone who ‘knows’ the answer, yet in a way unbeknownost to you, doesn’t really understand the question.
Your use case is the context, but what if someone new to the industry, and new to learning systems, is presented with this use case?
Will they really understand the question?
A high probability says no.
What then is missing in the use case to provide the answers that the person, instead, is presented with the question and in return presents the answers accordingly?
Nobody likes writing use cases. They can be an experience of dread without procrastionation taking hold.
Your Subject Matter Experts present feedback that is usually jargon and statements only they know the answer to, again, without understanding the question.
Then you have other departments, or worse, a committee.
Each individual states what they need, knowing full well (or at least they should) that this is not a buffet where everything they want goes into the use case.
The problem, though, is that the person writing the use case may shove all those items into the case, which then becomes a new book written by James Joyce (good luck with that read).
Off goes your use case.
Into the hands of a person who reads it and would never say they lack understanding of what you need, or again, in a real-life scenario, some vendors have ready-made responses as a plug-in and send back (common with those RFPs you send thereafter or before your use case).
Yes, a use case has to be specific, but it needs to understand the question, even if you, the writer, know the answer.
The answer isn’t the system itself – that is why the use case probe is being sent – rather the answer is the use case that you wrote based on what you know, assume to know, and therefore should be clear to anyone.
Can AI then assist in a manner that eliminates the extensiveness, and present the findings in a clear and concise format, which in today’s world recognizes the speed read approach.
And in return, can the person who is basing their chat with you, on the use case itself, if the questions are presented in a way that really adds depth?
Let’s be real, there are people who ask AI to create their use case for them, based on a set of criteria, or maybe the output becomes a template that the person overseeing the use case pulls a plug in and saves time and goes format.
The vendor, in turn, may take that extended use case and stick it into any of the free AI tools, request a summary, and tada – fast and efficient.
Ignoring the many mistakes aspect, hence the need to do a re-check.
While there are plenty of sites and people providing AI prompts for anyone to use, and even teaching people how to use prompts – format, structure, etc. – none of these will really help you with your use case.
All this goes back to the Latin statement – Scire responsum non significat intelligere quaestionem.
You do not need to create an agent or use an agent (you may have heard that term) to generate various responses that can be sent in place of your use case.
Ask more, and see more.
View the Entire Booklet – The Prompts I used, the results and more – Knowing (available for download)
Your Use Case
Recognizing that it is all about speed these days, and limited focus, I wrote a prompt with a series of words (this is after I uploaded my use case – you can find a multimodal (basically it accepts files, etc.) which in turn outputs a response.
Summarize use case, including specific requirements by user role, key challenges, explicit and implicit goals, expected outcomes and essential questions to address with potential LMS vendors.
The focus here is to get right down to the nitty-gritty.
Based on my prompt, the response starts with a very high level and short executive summary (remember this is going to the vendor)
Prompt: Summarize the use case in a digestible format, identifying the challenges, risks, outcomes, and impact of learning for the employees, customers, and the company itself.
Present results in a bullet format with headers for each section, reading time no longer than 15 minutes. The reader will have limited information about our company.
For these results you will see information that can be sent to the vendor, information for your internal processes and purposes, and data points for you to track.
What types of items do I place into my prompt or prompts?
Here is a short list that I used with my prompts, drilling down where needed
- What is the overall goal – outcome?
- What are the challenges? – From IT to Implementation to Context of learning?
- What are the key takeaways?
- What are the security and infrastructure issues?
- What do you want the new system to solve?
- How can a vendor understand what you are presenting?
- How can they respond, whereas they are responding and not using AI in any manner in conjunction with their follow-ups?
- Will their answers spur you to think of additional questions to address those inquiries?
Do I or Do I not?
There is no doubt there will be people who wonder whether they should send their use case over to the vendor anyway, even with these specifics.
My response is yes. Think of it as a way for them to get additional context.
The goal of the use case prompts is to present specific statements, summaries, questions and other information they may not be able to extract.
Or if they do, to ensure everything is crystal clear.
If you have specific implementations, state that in your use case.
“We use Workday, ADP and WidgetSystem, which will need to be integrated into the LMS (or any other type of learning system”.
The Use Case Analysis
My prompt
Present use case in short bytes for a quick overview for the reader, design in a micro-learning approach, with objectives followed by preferred goals, stakeholders involved, challenges overall and specific job roles, and outcome required
Then I was able to break it down further, including changing the format of specifics say outcomes vs challenges in a compare and contrast with a table using icons where applicable.
You can really do quite a bit, as long as you have the initial information – which you do – its that use case – and then tapping into the ideal prompts to get you to the next stage
Salesperson Questions
Remember AI today, cannot provide deep thinking, thus to ensure that a team of folks at the company (the vendor’s system) will need to work together to respond accordingly, you can set up a prompt to develop the questions (always review before accepting) and expectations.
I added some details including level of product knowledge from the salesperson, and total amount of time to read.
Before going on to the demo, knowing the answers to these questions would be extremely important. Make it a requirement they must respond by X date (typically two to three weeks) for consideration.
The question headers (specific questions under each header)
- Integration and technical architecture
- Implementation and Reality
- Scability and Performance
- User Adoption and Change Management
- Competive Differantion
- Risk and Failure Mode
Let’s look at one question under the user adoption section
Manager Dashboard Abandonment Pattern
Based on your customer data, what percentage of managers log into their dashboards more than once per month after the initial 90-day launch period?
What distinguishes high-engagement manager populations from those who never return—and what intervention strategies have proven effective to combat dashboard abandonment?
The use case presents this as an issue that needs to be resolved – but while specific it opens up the vendor’s response in terms of generality.
This question goes further in, again, requiring the vendor to respond in a real world environment – not a general everybody does this retort.
Let’s say though you want a different approach and style of questions to be sent over to your salesperson.
My Prompt
Present a series of questions, based on the response that requires the reader to provide in-depth answers, validating that they understand the information presented to them.
The questions should be clear and concise. Show each question in a bullet point format with spacing between each question and a header.
The findings – i.e. the questions, I felt were better than my initial sales questions depending on your variance and choices.
Example of another type of sales questions (think a different approach)
My Prompt
Present a series of questions, based on the response that requires the reader to provide in-depth answers, validating that they understand the information presented to them. The questions should be clear and concise. Show each question in a bullet point format with spacing between each question and a header
Output Example
Question – Explain how the absence of HR system integration creates a cascading effect of administrative challenges throughout the LMS. In your answer, identify at least three specific downstream impacts beyond just manual user creation.
This is going to the salesperson – they receive the information – the questions and thus need to respond.
Another example – once again, from your use case
Employee Impact Trade-Offs
Question – While the LMS provides numerous benefits to employees, identify two potential negative impacts on employee experience that are not explicitly mentioned in the document. Explain how these could undermine adoption despite the positive features.
Prompt: Summarize the use case in a digestible format, identifying the challenges, risks, outcomes, and impact of learning for the employees, customers, and the company itself.
Present results in a bullet format with headers for each section, reading time no longer than 15 minutes. The reader will have limited information about our company.
For these results you will see information that can be sent to the vendor, information for your internal processes and purposes, and data points for you to track.
Additionally you can dive further in, and extrapolate the info of use to you, and the vendor themselves.
Sending the use case with questions for the vendor to respond to.
Impact of Learning
Prompt : With my use case what will be the impact of learning to business?
Pretty specific, right to the point and uses the recommended approach of Impact of Learning, rather than the usual ROI, which doesn’t apply well enough for showing the pluses (which you always want to do)
I followed up by asking ‘Summarize the text, specific bullet points for each category – Takeaways, no more than a five minute read – highlight the big wins.’
As you can see, I asked for specific information that I wanted extracted and a maximum time to read it.
Big Wins
- Transforms training into a revenue generator and strategic differentiator
- Drives measurable improvements in efficiency, compliance, retention and client value
- Enables rapid scaling and innovation without proportional increases in headcount
- Supports data-driven talent and business decisions
Now ask yourself what would you rather have a senior exec see – A general ROI with no specific outcomes or a “Big Wins” tied to the business itself?
What else did the prompt provide me?
Quite a bit, but I am only going to highlight an area, any business leader would buy into
Talent Development and Employee Retention
- Personalized Learning Paths – Course recommendations based on roles and skills reduce turnover by 10% to 15% (500K to 1M saved/year)
- Skills dashboards – Employees see their growth, increasing internal promtions and reducing recruiting costs by 20% to 30%
- Just-in-time Learning: Faster onboarding (from 90 to 60 days) and immediate access to sales/product traianing
- Knowledge Retention – Spaced repetition and manager follow-up double retention rates and life sales quota attainment by 15% to 20%
- Peer Learning and Gamification – Community features and competitive elements boost engagement and participation by up to 60 percent.
Again, all of this tied around the impact of learning and using specific prompts – all from your use case.
These prompts are not in the usual hey do this or here is a list of generality – rather it is knowing your subject, and inquiring to see actionable results and possibilities.
Can I do this with customer training/client training?
Absolutely. This isn’t limited only to employees, or a business that has internal and external (as my use case example does).
If you plan to charge X price for the content state that, and follow up with the impact of learning prompt.
I’d even ask something like to generate $250,000 for the year, what should my mark-up be (increase between the actual cost to you for the content – think build) and what you charge.
On a personal note, I can state that you must break-even. Thus, if the vendor is charging you $25 per learner (per year), then you can charge $39.99 and generate revenue.
The problem folks have is that they are either greedy or unaware and thus assume a higher price point will boost better numbers.
Sure, but you want to build mass here – use the Blue Ocean Strategy – I did, and generated nearly 1M in nine months – of pure profit.
What did I charge per course? $49.
Remember when Apple launched their music store? What did they charge per song? 99 cents.
What strategy did they use? Blue Ocean.
What if we have requirements with our use case, or just a set of requirements in an Excel file?
Not a problem.
Based on the prompt you can generate results and export as an Excel file.
Then use Copilot if you have it or create it in a visual style and format. Or have your Gen AI or Agentic AI create it as a PDF file, or retain the table for download.
In today’s world, more and more reorganizations are impacting L&D and Training departments. The system ends up going to HRIS (often) or another department.
The person or people that will oversee the system, lack any knowledge around L&D and Training.
You still have the use case, or in this example an Excel or Google Sheets requirements.
What can you do?
Prompt: Shorten response design it to be read by someone who is working in HRIS with zero experience in using or buying an LMS. Be clear and concise. Focus only on relevant information in a summary that includes bullet points. Each section should be no more than 10 words. Write it in English
Let’s say though you want the response to be in Latin American Spanish.
Simply replace “Write in English” to “Write in Latin American Spanish”.
Proposals
Understanding and comparing proposal results can in itself be a tiresome and confusing process. Even savvy folks who have purchased systems before easily can be fooled into thinking one way, forgetting that the real (what you pay), is something completely different.
Then there are people new to the whole learning system space (learning tech too – after all the prompts can be for any use case or specs), and are required to receive these proposals.
Which one do they pick if they lack the knowledge of how a system works, what are the benefits directly to them, and the costs.
I used three actual proposals that I received. The proposals are from 2016 (thus do not tie the actual results, i.e. costs and recommendations based on the findings). Again, this is for reading/example purposes only.
Prompt: Summarize with recommendations, nor more than one minute read (high level)
Prompt: Based on the EXAMPLEXXXS use case requirements and three vendor proposals, provide a comprehensive analysis comparing costs, capabilities, and alignment with our business needs.
Bottom Line
Using general prompts you can find on the internet only works in our industry if you know your use case, present it, and seek specific insight which will be sent to the vendor.
Always remember time is essential these days, speed a necessity and mistakes can happen – which impacts your response.
Make sure to review your results prior to sending to verify that the information presented is accurate.
Despite what you hear, no Gen AI or Agentic AI is 100% accurate all the time. They may make mistakes.
The impact to you if you ignore that, is what happens when an employee is taught or provided with a list of prompts.
A real possibility that the output isn’t real
The prompts, yes.
The answers, no.
E-Learning 24/7
Word of the Day: ambivalence
This word has appeared in 140 articles on NYTimes.com in the past year. Can you use it in a sentence?
Motivation Effects and Efficiency of Retrieval Practice over Lecture
Taken together, what does this tell us?
1) We need to be paying attention to the interaction of learning strategies on motivational factors. Motivation matters both for student participation in course material but also for self-directed learning.
2) Jumping straight to practice is (maybe) ok to do. When I first read this article, I was thinking about how skipping the lecture was like skipping the explicit instruction and jumping straight to an inquiry approach. And in some ways, it is. But practice with feedback also includes important elements of explicit instruction. Participants are being explicitly told the answer after each problem. Still, there are a lot of instructors who probably use something like this as their inquiry approach.
However, I caution educators not to ditch explicit instruction altogether as I’m concerned there may be some interactions here yet to be uncovered. For example, there could be an interaction with cognitive load. Maybe these problems were relatively easy and more complicated material or students with lower prior knowledge wouldn’t show the same effects.
3) Practice matters. Across all studies, the conditions that included retrieval practice far outperformed lecture only conditions. It is not enough to just lecture. Students need that knowledge reinforced through retrieval practice.



