To ensure that failed projects are not delivered due to overestimating the importance of AI, focus on principles and keep expectations of experimentation alive.
译自 What Generative AI Means for Product Strategy and How to Evaluate It 。
Looking back over the last 10 years, we can see a bunch of technologies that initially received huge hype but ultimately didn't become "mainstream". Whether it's being eventually abandoned for failing to cross the chasm, or barely surviving with lagging adoption, you probably know exactly what I'm talking about.
However, it seems that every time a new hype cycle begins, there are people in companies who start crying out, claiming that if they don't follow the wave, their own products or services will be left behind and their organizations' lives will hang in the balance.
Years on, for example, can you imagine how a complete failure your product would be if you threw yourself into virtual reality (VR) experiences, released non-fungible NFTs (non-fungible tokens) for trivial things, or even fully integrated Bitcoin into your checkout experience?
While it may seem easy to put generative AI in these overhyped tech piles, whether because you think it's annoying, distracting, or even anxious, it's foolish to do so.
Pandora's box has been opened. Anyone with an hour, Google, and some curiosity can see that this new generative AI technology is:
- true
- It is accelerating
- was adopted quickly
- Suitable for most digital interaction scenarios
Technological change, constant principle
In this sense, generative AI is much like tablets, smartphones, phones, and PCs before it. Our underlying technology is constantly changing, while the basic principles of users and their problems remain the same.
Innovation comes from discovering user problems that users themselves think are important and then solving them in a faster, cheaper, and easier way. The car solves the core problem from A to B better than the horse-drawn carriage, but the problem itself remains the same.
This for generative AI means a paradigm shift in the problem paradigm that can and/or should be solved. What was previously too hard or too expensive to consider at all, can now be done in minutes through the API.
Suddenly, a pain point that has historically been too difficult to solve needs your attention.
Dealing with stakeholders
Although generative AI is real and here as mentioned earlier, both executives and junior employees in your own organization are likely to shout and either claim that you need immediate adoption everywhere or ignore it because "it's a distraction."
As with all new product initiatives, be careful. Reacting impulsively to these comments can cause you to miss the core strategic processes that we at VMware Tanzu Labs believe all product teams should undertake.
To avoid starting useless projects due to overestimating the importance of AI, focus on principles and maintain positive expectations because you are still experimenting. After all, your product may still hit the market prematurely, or some unforeseen circumstances may cause failure.
Here's a theoretical example:
We can be sure that there are a lot of voices within the blog CMS platform calling for them to add a one-click language model to the blogger's blog so that any blog reader can "chat" with the blog.
But does it make sense? Building such a feature will certainly require considerable time and effort, but what might happen after release? Our bet: silence.
Is the main purpose of users reading personal blogs to get specific technical answers? Probably not. Users read the blogs of people they admire, yes, to learn, but also to understand the author's point of view, to entertain them with the collation of thoughts and opinions, and to learn about recent developments. Users may not know how to prompt the AI in the chat interface, nor may they necessarily care.
Now, will a tech blog about the details of automotive mechanics create value from providing a chat interface? This is quite possible, as we think that most people who visit this kind of blog are looking for specific answers to specific questions that can be asked via a chat interface.
If you roll out the former instead of the latter because you feel pressured to do something AI-related, then you're ignoring the basic principles and pushing out something useless. Delivering what is requested will only calm the voice temporarily, and then your roadmap will be questioned again because you overreact to it.
AI strategic models
The first step in reviewing or building a strategic model of an AI product is completely non-negotiable. You need to know your users and understand their problems. If you make a decision intuitively without conducting field visits or user interviews within a few quarters, you think you know, it will lead to failure. Be honest with yourself and keep going.
Once you have this knowledge, you need to explore and document what your client's work is going to do (JTBD). We don't necessarily advocate that you fully adopt JTBD in your legacy organization, but questioning and documenting why users choose you and what they want to accomplish is fundamental here.
Based on outlining what you know customers want you to do for them, you need to evaluate how well your product or organization is doing the job or task, and what measures are being adopted.
The final step is to understand whether AI can improve this performance or make the problem or work completely irrelevant.
An example of the final step is: Can you use AI to better deliver a single sign-on experience? Or can you eliminate the need for SSO altogether by continuously authorizing access to the application via the user's webcam (imagine always-on Face ID)?
For many organizations, if AI makes a problem or job completely irrelevant, they won't seek to be a replacement, as it could be a politically sensitive issue — and if it could, it would be a Kodak moment. Now is the time to avoid disaster and ask this question if you want to continue innovating.
If generative AI seems to enhance the way your product or organization gets work done for users, the Product Owner needs to ask, "Can AI impact this in a significant way?" The key factor in importance is understanding the measures by which users judge the success of your product.
Fine-grained assessment
So far, we've been vaguely talking about the product or product size you're evaluating your strategy for.
That's because everything from product managers to CEOs needs to develop AI strategies for what they oversee, whether it's a customer application or a whole host of departments. Many of the products that PM is responsible for will be completely replaced by a higher-level replacement departmental strategy.
Again, while this may be an uncomfortable topic, the market is evolving, and as practitioners, we have a responsibility to consider the impact of AI.
The assessment level might look like this:
firm
- Why do you exist?
- What are the core outcomes you help your clients achieve?
department
- Why does this sector exist?
- How does this department interact with the user/brand experience?
team
- What is this team doing?
- For which user?
- To achieve what results?
products
- What can this product enable users to achieve?
- Are the issues addressed by this product caused by other parts of the user/brand journey that we control?
We believe the result of this exercise is that the organizational part that engages the user journey is well suited for internally generated AI capabilities, products, and experiences, while the organizational part that is not relevant to the user will be most severely disrupted by third-party AI tools in the future.
To dive into the noise around generative AI and take advantage of what's available to you today, join our upcoming webinar, "Generative AI 101: The Realities of Generative AI and What Business Leaders Need to Know."