AI will make your work more human, not less (ssr)

Written by Simen Svale

Missing Image!

I’m one of those people who are, you know, “good with computers.” But lately, I’ve been reflecting that being “good” with computers just means I deeply understand what dumb, mind-numbingly rigid mechanisms they are—and use that understanding to coerce them into doing things humans value.

Essentially, in learning to use computers as they are today, we’ve had to reprogram our human thinking into the mechanistic regimentation of our machines. We’ve done so with life-changing success, of course! I mean, people who are “good with computers” have given us everything from life-saving telemedicine solutions to… that latest Zelda game, Tears of the Kingdom. Probably the most inspired bit of entertainment in the history of entertainment.

Now, Artificial Intelligence (AI), and in particular foundation models, are shepherding us into a new era. Machines are beginning to have common sense, even something like empathy. This is about to flip the script: we’re now seeing impressive evidence of machines being able to understand us and figure out what we need to be done for a change. The breadth of ways enterprises can apply AI—practically, safely—is growing exponentially.

I say beginning because for every cherry-picked example of AI being amazing, there are thousands of moments where it misses the mark. This is still early days. AI is showing profound promise, but as we seek to apply it today, we need to see ourselves as explorers.

Trading chores for creativity

There is also the fear. Will AI take our jobs; will humans become obsolete in the workplace? While I understand where this sentiment is coming from, I have a different perspective: I believe AI will allow us to be more human. To do more of the interesting, specialized, and creative work that underlies our most fulfilled days as employees. To bring ideas to life and make the experiences we curate more compelling and valuable for our audiences. Because finally, the machines are going to understand us—all of us—even if we’re not “good with computers.”

I’m bullish on AI as I look to history. Since the 1940s, millions of jobs have been automated, optimized, and restructured into obsolescence, yet as this has happened, the employment rate has nearly doubled. Similarly, AI will make the cake bigger for everyone by increasing the value of knowledge workers’ contributions. I’m not alone in this; as we talk to companies about their ambitions for AI, we don’t hear of layoffs but of finally being able to do more with the people they have access to and can afford.

I’m even more convicted based on my personal observations. Most of the talented people I’ve crossed paths with over the last few decades, whether as customers, colleagues, or friends, only really spend about 10–20% of their time on the type of high-value, fulfilling work I described. The rest of their time is spent doing tedious, repetitive tasks (or waiting for someone else to complete their tedious, repetitive tasks). Tasks that eat up our capacity to deliver on our uniquely human potential.

We’ve been reducing these burdens for customers with the Sanity Composable Content Cloud from the start. We drive team-scaling content reuse with a platform for structured content that offers the most interoperable, customizable content back-end, and of course, our editing environment, Sanity Studio, which can be tailored to maximize the efficiency of any team. We streamline chores with these big ideas and also the many small-but-powerful capabilities we’ve built in, like intelligent image cropping and scaling so you only need to import an image once for it to render optimally anywhere.

When we remove these redundant tasks—managing separate CMSes, integrations, and content for each channel—it creates capacity for creativity. It allows innovation to flourish, which improves your team’s quality of life and lets them spark more joy and value for customers in turn. From there, your business thrives.

Now, we’re supercharging our pace of chore-slashing for our customers and in collaboration with them.

Partnering in these early days of AI

Without a doubt, companies are adopting AI as a strategic imperative: 73% of executives say their organizations have embarked on a path to intelligent automation, up 58% from 2019 (IBM). The recent Stack Overflow survey shows that 70% of developers are, or are planning to, use AI for their work. Specific to the content space, Content Science Review cites 29% of companies are using AI or machine learning within their content lifecycle (a nearly 2x increase from 2017).

It’s interesting to compare this to the pace of change within other spaces, like e-commerce. There’s really not much that e-commerce can do today that it didn’t do in 1998. Why? Perhaps because today’s senior leaders grew up with the Internet and have witnessed late adopters wither in the dust. So they’re pushing the pace for AI adoption while learning on the fly how and where incorporating AI will strike the right balance of risk and reward, of investment and return.

With this in mind, we’ve released Sanity AI Assist to give you a turnkey way to begin exploring and testing how AI can best streamline your content lifecycles. With the flexibility to build custom instructions, we’re excited to see what you decide to delegate to AI. Check out our announcement post to see a few of the use cases we’ve been tossing around.

What we bring to the table for this exploration starts where pretty much everything we do starts: with structured content. Sanity AI Assist gives large language models the ability to read your content schema and execute instructions within it, whether that’s creating content or performing workflow-accelerating (but unpublished) tasks.

Powerfully, this ability to understand structured content also allows us to build a joyfully plainspoken UI for your teams. Editors can automate complex workflows with a simple sentence one would use to guide a capable colleague.

We care deeply about making Sanity Studio an excellent experience for content teams. When I first started playing with Sanity AI Assist, I was taken aback by how simple it was to automate even rather complex tasks. It felt as though I was doing programming work with just a couple of sentences.

And this is the unlock: there are so many burdensome tasks your content teams do that probably could be solved with some creative programming, but it’s just not practical to invest developer time to solve these chores. Now, you (or your editors) can simply tell Sanity AI Assist what repetitive task you’d like it to take care of and free up bandwidth for higher-order work. Once an instruction is created, it can be shared and used by anyone in your workspace, which scales efficiencies.

So, if you could delegate work your content team performs to a current-day work version of the Jetson’s Rosie, what tasks would you pass along? Your chores, of course. The work you get little satisfaction, growth, or pride from doing but that you need to get done well (without micromanaging). We’ve teased a few of our first examples of Sanity AI Assist use cases in our launch post, but the opportunities for high-return, immediate applications are vast. What will you take off your content teams’ plates?

Years of more human work to come

Every week, we’re witnessing moonshots become everyday AI. At one point, I “discovered” that GPT can guide itself: I asked it to plan out its own work, then execute that plan, evaluate the output, and revise until it deemed the work complete. Then I went on a trip with my family for 5 days. When I came home, AutoGPT had blown up—essentially the solution I was about to build.

At Sanity, we don’t deign to know precisely what the landscape will look like 6 months from now, but we’re keenly aware of the challenges our customers face. We’re confident in our roadmap, which is both concrete and responsive enough to use emergent technology as it helps us deliver the right value.

Expect us to expand Sanity’s AI capabilities to solve problems like creating an informed proposal for content modeling (and remodeling); understanding your corpus of content to guide the highest value content to create or evolve; and begin acting as a producer, orchestrating work in a manner personalized for every team member in the content creation value chain.

We’re humbled to exist in this era of long-imagined but unachieved artificial intelligence. We see it as a privilege to co-define this future of “thinking machines” to bring more humanity to our work.