A Look at 3 Ways ChatGPT Impacts Strategy, OKRs and Product Management
A Look at 3 Ways ChatGPT Impacts Strategy, OKRs and Product Management
The Generative AI Edition – Prompting as the next human-computer interface
Welcome back to the Upstream Full-Stack Journal, where we go up- and downstream to explore how Strategy and OKRs can make you radically more effective.
In this edition:
Different perspectives on how Generative AI is reshaping:
Strategy
OKRs
Product Development
I offer my perspective
POLL: How interested are you in making AI part of your Strategy, OKR, or Product work?
AI and the future of work
When I grow up, I hope to be a “promptive.” From Doonesbury.
https://www.washingtonpost.com/doonesbury/strip/archive/2023/5/7
ChatGPT, Strategy, & OKRs
Brett Knowles over at PM2 Consulting kicks things off with a series of videos on tying together OKRs and Strategy using ChatGPT:
In Brett’s words:
“ChatGPT and any other Generative-AI tool can provide incredible value to our strategy development and strategy execution processes. OKRs (Objectives and Key Results) are the “wiring” that enable this process.”
Brett Knowles, Why OKRs are the Secret Sauce to GPT-Driven Strategy Development and Execution
Read more & watch the full playlist on Brett’s channel in YouTube here.
How Conversational AI Can Boost Product Management
Yana Welinder, CEO of Kraftful, used Generative AI to cull Product Twitter’s best minds, and turn them into quickly-accessible & actionable insights:
https://twitter.com/yanatweets/status/1617930615549984781?s=20
Yana has a full writeup on the kraftful.com blog of her experience powering her app with the best & brightest minds in Product.
Now in Beta, one of kraftful’s “killer” features rapidly pulls together customer feedback across multiple sources to create natural-language Product Research responses.
Perhaps even more impressive is the list of product-driven organizations that have already been using the Alpha version of this feature: Google, Microsoft, Meta, Netflix, LinkedIn, among many others.
Strategy to Execution Powered by OKRs and AI
From my friends over at Quantive, they’ve recently gone all-in on implementing Conversational AI throughout their Strategy to Execution platform.
Overcome the OKR “cold start” challenge. Users often need help defining objectives from an overall company strategy or high-level goals. With Quantive’s integrated AI capabilities, users will get suggestions to help sharpen objectives, consider new or complimentary key results, and even orchestrate tasks to help them get started faster.
You can read more on their blog here.
A Different View of AI and Product Management
Former Googler Itamar Gilad uses the Business Model Canvas as the main example in his piece “On GenML, Artifacts, and Product Management.”
After ChatGPT fills out his Business Model Canvas with “potentially useful content,” he digs into the implications of outsourcing hard cognitive work to AI.
While he concedes there may be potential benefits, he cautions they don’t come without risks:
“Obviously, no one is planning to delegate judgment to GenML systems (right?). We see them as a convenient shortcut, but we’ll still expect to think for ourselves. The problem is that the bots offer a compelling temptation.
Research, analysis, estimation. review, and decision-making require what Psychologist Daniel Kahneman calls System-2 thinking —slow, effortful, intentional, and tiring contemplation.
In the book Thinking Fast and Slow Kahneman explains that most of us avoid using our System-2 thinking when there are convenient shortcuts available, even if those lead us to the wrong answer.”
Itamar Gilad, On GenML, Artifacts, and Product Management
A helpful mental model
From my perspective, I find it helpful to divide your area of focus into one of two areas:
Are you trying to rapidly sift through tons of data and understand things that have happened in the past?
Or are you trying to design a new and different future?
1. Understanding the Past
By definition, any data we can analyze comes only from the past.
ChatGPT, for all its promise, was trained on data sets that end roughly around 2021.
So if we’re trying to look backwards and understand the history of something for research – closely aligned to the Kraftful.com Product Research model – that rapidly sifts through and summarizes “app store reviews, support tickets, call transcripts” – and provides actionable insights in natural language conversation would seem to be an effective use of the technology.
2. Designing the Future
But when it comes to creating something that hasn’t yet existed– your company’s strategy, the associated OKRs to measure its effectiveness, and the Product Management to make it real, I would be reluctant to use the tools.
Why?
1. Designing Strategy as a Team Sport
Designing Strategy benefits from having a diverse, cross-functional group committed to the process.
Together, they step through the 7 Steps of the Strategy Process Map.
The goal of working through it isn’t just to get to the final result – the Strategy.
It’s the learning and shared experience that happen when the group is collaboratively designing, discussing, testing, diverging, reviewing, converging, and ultimately making a set of choices.
2. OKRs Created Leadership Together with the Cross-Functional Team Doing the Work
Once you have a clear set of Strategic Choices, setting metrics that let you know you’re on track via OKRs is the next step.
I understand how Generative AI can give you a “rough draft” to start with.
My concern is OKRs are best done when teams, and teams of teams, meet with their leadership to discuss and craft a set of measures aligned to their Strategy.
Leadership assists with drafting the aspirational Objective that the team provides input to.
After mapping the User Journey, and identifying Client Behavior Change Outcomes to influence along that journey, I would suggest that it’s the team that proposes the Key Results.
To delegate this to Generative AI would seem to miss the opportunity to have some valuable conversations between leaders and their teams that can lead to shared understanding of where to focus for the coming quarter.
Product Management: Seeing the Best Decisions Get Made
Product Management involves a wide array of skills and competencies that differ depending on which expert you ask.
Some involve tons of research, sifting through and grokking tons of data. This is exactly the problem that Kraftful.com’s Product Research platform seems to serve brilliantly well, as it “synthesizes user feedback from app store reviews, support tickets, call transcripts” and returns insights in plain language.
But it’s been said that “product managers aren't the ones who make the best decisions, they're the ones that see that the best decisions get made.”
And I’m not sure we’re quite there yet, asking ChatGPT to make decisions for something that hasn’t yet been tried.
Translating Strategy into Software
One big part of Product involves creating artifacts teams use to guide their Design and Engineering work to produce working, tested software.
Teams come to rely on Product Requirements documents (PRDs) or Epic Business Cases, or User Stories so they can take action.
In some teams and organizations, they “delegate” these tasks to Business Analysts, who go off by themselves and churn through and write all of the necessary documentation, then hand them off to the team for “execution.”
Siloing this work with one person, and unquestioningly taking their output to mindlessly code against seems to return us to the Waterfall value of big, up-front planning.
And even the founder of Waterfall, Winston Royce, felt made the approach was “risky and invites failure.”
Design artifacts collaboratively
The best teams create these artifacts as a shared, cross-functional, collaborative experience.
(Seeing a pattern here?)
If we do have to slice it down into a smaller group, at minimum, we look to involve “The Product Trio” at the very minimum – the shared perspective of the Product, UX, and Tech Lead/Architect roles.
All three of those perspectives are crucial to lean in and design the artifacts that the team will be using as a guide for forward action.
But I think you still need to learn how to prompt
Being able to prompt Generative AI, like being able to cook a meal, engage a friend in conversation, negotiate, draw a picture, use a computer, get the most out of a phone or a tablet, or even run a Google search, will be just one more necessary measure of effectiveness in today’s world.
But what do you think?
TL;dr
Totally OK to outsource drudgery to Generative AI tools.
I would prefer to take on the hard work of both System-2 thinking, as well as the promise of designing the future with a small, diverse, collaborative, cross-functional team.
That’s it for this round!
Join me next time as we continue to go Upstream and use Strategy to make you more effective everywhere.
