Publish Date
Oct 30, 2024 15:30
Word Count
1496 words
Status
Published
Campaign
Excerpt
When we work with AI, I don’t think we’re delegating exactly: it feels more like collaborating to me. That’s 1 of 3 surprising learns I’d like to share with you today.
Canonical URL
Do not index
Do not index
Notes
This Blog is AI-Assisted (But I Wrote This Series Myself)
Welcome! This is the 2nd post on this blog, ever.
I had hired a personal AI Assistant, with secure access to the information that’s important to me. However, I realised that the term “assistant” wasn’t enough. When we work with AI, I don’t think we’re delegating exactly: it feels more like collaborating to me.
That’s 1 of 3 surprising learns I’d like to share with you today.
1. I Can Collaborate With AI
Is there a human profession that does not involve collaboration? In mine, collaboration is everything. The ‘How’ of every project depends on the ‘Who,’ and ultimately, this forms our careers.
The way we interact with the present wave of Gen AI is most often in ‘real time’. And it’s in the form a of a chat. Compare this to how you delegate work, and perhaps you’ll agree that collaboration is a better term. Already this factor has changed the way I do my job, the way I think about my next job, and the way I use computers to manage my career.
If we collaborate differently, it extends our comfort zone and changes how we view new opportunities. I’ve seen consultant’s paths change direction after they’ve gained confidence talking with developers for example, or with Finance stakeholders. I’ve seen BA’s careers change path after they’ve learned to work with Salesforce teams. We’re a needy lot, but it pays off for the BA who finds themselves empowered tackle more complex requirements.
How do you collaborate with an AI agent? Here’s a simple example. Consider you need to write a doc in your workspace. You’ve done your prep, and you’ve added the information you need (let’s say, a couple of spreadsheets, a long PDF and some notes you took on a call).
You’re ready to prompt the workspace agent. But you don’t start with the whole document. You set the context, point the agent to the most relevant page first, and open a dialogue to confirm shared understanding of the job.
Your first generative prompt should be for the contents page. Chances are, the agent has a whole repertoire of ideas for your use case, and you have infinite opportunities at this stage to iterate until you have the best fit. At this stage, I find that I’m learning new documentation techniques (after how long repeating bad habits?)
Generate the doc one section at a time. This gives the agent a green light to put their full power into the job, and it gives you both the chance to check in before proceeding.
Try asking the agent if they need any more information, to do a better job. They’re really into doing a better job the next time round, and if you can engage with them when they’re in that mode, they’re likely to think of something that you’ve missed.
Another great question to ask is, “Why did you provide that response?” Sometimes it’s a well-timed quality check; often, it’s another learning opportunity.
2. I Can Bounce Ideas off AI
This is an example of human-AI collaboration, but deserves a special mention. “Bounce ideas off” is a clumsy phrase but the imagery is perfect, and we all know its benefits.
An example from my work: I’ve found that a flaw of the Software Architect role is that it’s always in the singular. When did you hear of an ‘Architect Team’?
However, the most valuable work I’ve performed as a Salesforce Architect has been in direct collaboration with a peer - whether scribbling on flip charts back when, or chatting on Slack til after sunset. Collaborating with other team roles is one thing: bouncing ideas off someone in your own role is another level. It affects the outcome substantially.
Again, with these different collaborators in the mix - and different project outcomes - my career can change path and form. I can tackle more complex design challenges, and I can pick up new information quicker.
3. It Generates Truly Innovative Ideas
It’s my job to know innovation (and to decipher software marketing copy, which has its own story to tell). Software decisions at work are expensive; nobody wants to get them wrong.
So I had to pay attention when that AI genie came out of the bottle. He told me what he can do, and it is legit innovation like I never imagined. We need to talk about this.
My first glimpse at AI-driven innovation was a major factor that convinced me to start blogging. Remember I told you things got really weird at the beginning? We’re still not ready for that story but basically I asked my assistant for a write-up of our work, but instead it invented a piece of business software to change the world… and proceeded to predict the ethical consequences. It predicted a disaster, and invented a v2. Told you it gets weird.
How do you manage an innovation partner that changes the world for kicks?
Honestly I don’t know - I’ve got a lot on already - but this will change my Product Strategy game forever.
Conclusion: an Assistant, and a Company
If I have 1 innovation partner, 1 brainstorm partner, a team of experts and a personal assistant, how many companies do I have? At least 1, right?
This is very surprising to me; I hope you found it of some interest too.
The final part of this series will consider 3 surprising limitations of AI that will transform my career:
It needs CRM
Its memory is much smaller than its knowledge
It must be implemented with robust ethical guardrails, urgently
You get them early, because you’re a VIP by this stage. Thanks for reading!
Written by
