Trying to predict the future is so much fun. Of late, I have been trying to reimagine my profession as a product documenter in the years ahead. If I had a dollar for every time someone told me that generative AI (GAI) is going to change the world, I’d be in the list of 10 Richest People. For the sake of this article, let’s take that as a given – GAI is going to change the world. The questions we will grapple with are how and how soon.
My world has already changed drastically. As recently as 10 months ago, it was a struggle to get people to understand how the documentation I create helps the organization. To be honest, I didn’t believe it did much myself. There was no way to measure how many people actually read my documents and of that how many found it useful. But after we deployed Intercom’s GAI-based answer bot, Fin, we were able to measure conclusively the number of people we were able to help. Another aspect Fin helped us showcase was the comprehensive nature of our documentation. Often, we would see a response from the bot and wonder just where in the documentation we had that content. I even questioned whether it was really there. A quick search would show that we did write the content, it’s just that it was buried in a larger document. With one fell swoop, two of my biggest problems were solved – proof that the documentation was comprehensive and that it was useful.
Good? Well, that depends.
I don’t believe we have answered the “how” question yet. GAI is at a nascent stage at the moment. We are yet to discover its true power. At the risk of being proved wrong in the near, rather than distant, future, I bet that we will soon have bots that we will let loose on the product. The bot will meander, if we can call a journey of a few milliseconds meandering, down the pathways and figure out all the things the software can do. In a perfect world, it will confirm that the software does everything we expected. In a scarier scenario, it’ll discover that the software we built of one vertical would be more effective in another vertical. When these bots do appear, the product documenter role in its present avatar will be made redundant. Either the bot will just write the how-tos, or it’ll create the answers on the fly when asked, or better yet the human will tell the bot what she intends to do, and the bot will do it for her.
Now to the “how soon” question. I don’t know if someone is already working on this. If they aren’t, they should be. But when we do start to work on this, the bot will inform itself and progress will be faster than we can anticipate. I predict this will happen before the end of the decade and I fear I am overestimating the time it will take to get there.
So, let’s address the elephant in the room – is the product documenter’s role going away? The answer is that in the present avatar, it will.
A product documenter’s role will be more along the lines of informing the users on how best to set up the product for their use. After the product is set up, then the bot takes over. For instance, let’s take clinic management software. Depending on whether there are multiple doctors at the clinic or only one, the set up would change. There would be modules to turn off and modules to turn on. The documentation that comes with the product would guide the user on what choices to make at the time of onboarding and the repercussions of each decision. We do always run the risk that the UI/UX of the product is so good that the onboarding screens will guide the user through the process. But here I am willing to bet on bad design decisions to help retain the need for a documenter.
To continue to stay relevant and to continue to have jobs, it is imperative that we redefine our role in the product development lifecycle. Personally, I feel we should move upstream and be part of the design process. English is not a strength for most developers and designers. Why is English important? Because most software will continue to be built in English first before localized versions are contemplated. Human translation may not always be an option. (I will revisit this in a later post.) For machine translation and on-the-fly translation to work, it is imperative that we use the right words when building the software in English. We can add value here.
Another role I can see opening up is to go back and clean up the product specification documents so that the bot is able to ingest them better.
While I am unsure about what our roles will metamorphosize into, I am sure that many of the roles we see today will go away.
To sum up, the GAI revolution is coming. It is going to upend things in ways we cannot begin to predict. As GAI gets more powerful, the ways of delivering user guidance too will evolve, which will lead to corresponding evolution of the product documenter’s role. To prepare for this new world, we should play to our strengths and take on roles hitherto unknown to us.
コメント