How to Use AI At Work
How to Use AI At Work
+ Why you got down-leveled & more
In this post:
How to AI at work
Could the future of AI be underwater?
Why your offer got-down leveled
How to bridge the divide between stakeholders and users
How to Use AI at Work
Prompt engineering has gone from esoteric to essential. These are 4 skills you need to know to up your game:
Writing documents
Building tables
Understanding users
Analyzing competitors
Our tool of choice is going to be Bing AI. Unlike chatGPT, it's FREE. [Get on the waitlist now.]
But - it's different. You can't submit really long prompts; there's a character limit. And, you can't continue the conversation after 10 back and forths. So, Bing prompt engineering is a skill of concision. Here’s how to work it.
1. Writing documents
Bing can be your first draft's best friend. It will put down all the basics for PRDs, vision docs, strategy docs, team charters, and memos. It can save you time writing virtually everything important.
However, there's an art to getting its help. Telling Bing to write a type of document - eg, "Strategy Doc" - will push it to a generic template. Instead, YOU specify the key sections & details.
Here's my template:
"Write a document to add X feature to Y product for Z reasons. Include the following sections: A, B, C."
Tips:
Use creative mode
Ask it to 'Continue' if it stops
In followups, ask for more data and supporting evidence
2. Building tables
Bing can build you a table with information really quickly. It helps you "think" in 2x2s and multi-variable equations. It's a great method for communication.
The problem with Bing and tables is many will fail if you ask for difficult information. It will just say, 'Sorry.' So one of my tricks is to focus on searchable info.
Here's the template:
"Build me a table comparing X on A, B, and C."
Tips:
Make X what it searches for
Choose comparison variables it can easily find
Ask for information it can search for and is available
3. Understanding users
Bing can help you quickly read through user research, especially if yours is a product people talk about online. It's a good first pass on users before deeper research.
The trouble is getting it to give you quality versus surface-level. Asking it to give you user quotes is one way to get to testimonials.
Here's my template. Try it yourself:
"Describe with user quotes the top reasons X"
Tips:
Use follow-up prompts to ask for more quotes
Consider building a table to get dimensionality
Don't expect it to things that aren't easily searchable
4. Analyzing competitors
As a PM, it's hard to have time for every competitor move. For certain one's, you want very specific information.
Like tables, the hard part is keeping Bing confident. So specify things like pictures and user quotes.
Here's my template:
"Include pictures and user quotes to cover the top elements of X Company's Y move in Z time"
Tips:
Use balanced mode
The pictures currently land in the citations
Use follow ups to dig into details you need
Bing AI can help you (as a PM) reclaim time & focus. It's a competitive advantage. You can minimize time on low impact work to focus on high impact stuff like discovery & making good decisions.
Could the future of AI be underwater?
Microsoft’s Natick is a research project to build data centers underwater. Phase 1 was a success (pictured). MS successfully deployed and retrieved a shipping container sized data center in the ocean.
It’s an interesting prompt to explore where underwater AI will take us:
1. Underwater Robots:
In 2021, the UK Navy deployed a fully autonomous 7400 ton submarine. There could be a not distant future where humans are regularly deploying underwater robots outside of military contexts.
Stuff starts practical and gets whacky quickly.
2. Delivering Internet:
The internet is literally cables laid across the ocean floor. But deep underwater is highly unpredictable. AI robots can master it more safely, and eventually cheaply, than humans. They can help lay cables, maybe even move them.
3. Transporting Resources:
The sabotage of the Nordstream pipeline in 2022 highlighted how vital oil & gas resources can be hard to protect underwater. Underwater autonomous subs could help safeguard pipelines. They can also help repair them in case of attack.
4. Cleaning up the Oceans:
The Great Pacific Garbage Patch remains a problem humanity hasn’t solved. It’s a matter of time and resources. AI could help. Once the cost equation is right, machines will do our dirty work. As we have always asked them to.
5. Deep Sea Mining:
Humanity is already drilling deep into the ocean offshore for oil. But these jobs are dangerous and tough for humans. And the projects are prone to mistakes, as the BP Gulf Disaster of 2010 showed (Deepwater Horizon). Starting with offshore oil rigs and moving through to jewels and uranium, AI will help explode the extraction of resources from the sea.
The ocean has proven tougher than the air for humans and autopilot alike. But once the AI truly advances, it will bring huge amounts of resources.
6. Underwater Colonization:
Underwater AI could eventually help humans towards underwater permanent stations. Like their arctic cousins, these could be scientific. They could also be property claiming oriented. And robots could man them.
Alright, that’s the end of this particular trip down water AI. I told you it could get whacky quick.
Why you got down-leveled
“They came back with an offer, but it’s at a lower level.”
Many jobseekers have recently come to me with this conundrum. The joy of an offer is paired with the frustration of down-leveling. I’ve sat on several hiring committees where this was discussed or done.
Here’s why:
1: It’s a bigger company
When you move to a bigger company, the experience and pay bar for a certain title, like “senior” or “director” is higher. This is probably the case in 65% of situations. “It’s not you, it’s them.”
2: It’s big tech
Microsoft, Apple, Google, Facebook, & Amazon are notorious for this. They beat compensation even at lower titles. So, once they hire someone with X years of experience at a level, hiring committees compare to them. And you have the situation now, where directors get hired as ICs.
3: Your patience
It’s not always because of the company. It can be you. Junior roles are more common. Some job seekers are simply not patient enough for a role at their level. Sometimes, you have to wait for a role at your level that you can score.
4: Too narrow of a company list
Similar to lack of patience is being too selective with companies. There are SO many companies out there. If you want to maintain upward trajectory on your level, sometimes you have to go outside the target list.
5: Your years of experience
An alternative “paper” reason you may be getting down-leveled is your years of experience. Your prior company may have promoted you due to your accomplishments. But to the market, your years of experience are at a lower level.
6: Your gravitas
It could also be your interview performance. If your presence felt junior or weak, the hiring committee may be worried you couldn’t hang at the hiring level. In these cases, they often reason they are doing it “for your good,” come performance review time.
7: The altitude of your examples
A final reason you may have gotten down-leveled is the way you responded to questions. If it feels like you are more comfortable at the lower levels, they are likely to put you there. VPs work with VPs.
How to bridge the divide between stakeholders and users
Anyone who has built products for a while has done it: shipped something internal stakeholders love, and user's don't.
It's the "Curse of Knowledge."
This cognitive bias is a psychological classic: The more we know, the less we can relate to others who know less.
So how do you break this down in the product development process?There are 3 important ways to de-risk & build something users love:
Continuous discovery
Prototype testing
Narrative control
This is especially important for features that are large eng investments.
1. Continuous discovery
If you keep your solution focus process trained specifically on providing solutions to user problems, you stay glued to the user. Opportunity - Problem - Solution is the key pattern. This sequencing helps you build something for users, not stakeholders.
2. Prototype testing
Make the rubber hit the road before you go to engineering. Plan far enough ahead so you can put the prototytpe or mock in front of users. This allows you to easily spot views of an animals butts instead of their fronts.
3. Narrative control
The Curse of Knowledge will infect your process if you allow cross-functional and leadership discussion to be focused on design decisions.
Bringing it back to the:
A. Business goal and
B. User problem
Provides a more fruitful discussion with stakeholders. That narrative control helps you evaluate shipped solutions in the context of progress towards the business & user goals.
Otherwise, despite all the continuous discovery & prototype testing, there will be a gravitational pull towards specific implementations, catered to experts.
That’s it for this week! If you aren’t subscribed or part of the paid version - be sure to smash that red button below. See you next time.