Although the concept of using artificial intelligence across your organization can seem like a good idea, it’s one thing for the C-Suite to decide to roll it out and another to actually get all your staff on board.
Adoption is a considerable barrier to AI success in the Salesforce ecosystem, and it can become even more problematic when training non-technical employees. It can be difficult, but certainly not impossible – as Bryan O’Neill, the CTO of FormAssembly, shared with SF Ben.
Why Is This Important?
At this year’s Dreamforce, one of the most dominant discussions was AI adoption. Specifically, how to get started when it feels like the ecosystem and industry around you are rushing ahead.
AI exhaustion directly feeds into low AI adoption, creating a significant bottleneck when trying to roll out new AI tools, features, and processes. AI exhaustion can be exacerbated for employees who don’t come from technical backgrounds, struggle to understand the technology, or feel like they’re not getting enough training. That’s why it’s so important that they’re supported before, during, and after any AI implementations.
During this year’s conference, I sat down with Bryan O’Neill, the CTO of FormAssembly, to see how companies can best support their non-technical staff in the AI era. His advice for businesses hesitant to take the leap was clear: “Just try it – that’s number one.
“Don’t be the person who’s scared to try,” he told me. “It’ll be hard to screw up, so you will learn a lot more by just trying it, which is kind of like anything, right? Learning hasn’t changed.”
1. Find Your Enthusiasts and Give Them a Platform
You’re inevitably going to find two kinds of people when it comes to implementing AI in your business: the enthusiasts and the doubters. For Bryan, ensuring that the enthusiasts have a platform to share their enthusiasm is key, as it could lead to a change of heart or a boost in confidence for those less sure.
“There are pockets of people across any organization that are always very enthusiastic about these things,” he said. “What you want to do is encourage those people to really find their voice and speak up.”
Leading the AI front in your organization can’t be a one-person job – relying on the employees who believe in the technology can be immensely helpful.
2. Create a Space to Share AI Wins
Building off the back of that tip, Bryan also detailed how it was important not only to have a space for the enthusiasts but for everyone to get involved too, and that a big part of that was trying to make AI as inclusive as possible.
“We did a very simple, low-tech thing using Slack,” he explained. “We created a channel called AI Wins, and we encouraged everybody who had achieved something using AI to post it and share their experience with the rest of the company.”
This way, employees who interact with any AI in any way can feel included and more comfortable with getting on board – not just those who code with it or build with it.
“We incorporated an ‘AI Wins’ section into the weekly meeting where people come forward and talk about how they’re using Gong or using Claude, not just engineering, but all those other teams. People see that it’s safe to do these things. It’s not just the remit of your engineering teams. It’s available to everybody to try.”
3. Recognize That Domain Experts Have Some of the Best Ideas
Another reason why it’s important to get non-technical staff involved is that it’s within these specific teams that you’re likely to find some of the best AI ideas. You don’t have to be an engineer or builder to suggest practical ideas for AI, and utilizing domain knowledge should be a priority.
“We actually have development going on outside of our engineering team as well, with non-engineers building capabilities that they are the subject matter experts for,” Bryan explained.
“Great ideas can come from anywhere. If you’re working with people that really understand your customer base and the challenges your customers face, they’re going to have great ideas just the same as anyone in your engineering or product organization.”
It can also be helpful to pair up non-technical people with technical people to brainstorm, train, and implement AI, so that the business experts become “extensions of [the] engineering teams.”
4. Be Honest About The Risks
“You’ve got to recognize that there’s risk involved,” Bryan said. “It’s almost too easy [not to].”
There are lots of risks when it comes to implementing AI. Data leaks, hallucinations, agent sprawl, and lacklustre or inaccurate outputs are just some of the ones that everyone – non-technical staff and beyond – should be keenly aware of.
These risks can hold a lot of people back, especially non-technical staff. Sugarcoating the dangers of AI will only push these employees further away; being upfront about the risks and how the company is managing them is needed before even thinking about starting an implementation.
“If you don’t have the engineering mindset, you’re not necessarily thinking about performance, scalability, and stability. We try to provide guardrails around work that is generated by non-engineers so that they get very early feedback.”
5. Involve Every Department, Not Just Tech
Last but not least, we’ve covered getting the teams beyond engineering involved in AI, but it shouldn’t stop there. The teams in your organization not situated firmly within tech, like marketing, sales, or content, should also be made to feel involved.
“The thing that makes [successful implementation] faster is the fact that we’re embracing the thoughts, viewpoints, and needs of people in sales, in marketing, in accounting – not just product and engineering,” Bryan explained.
Final Thoughts
Perhaps my favorite takeaway from my conversation was his candid comment about the fear around AI: “Don’t be the person who rejects the idea that the world is changing. Figure out how you can be part of that change.”
There are numerous reasons to be scared, skeptical, or put off by AI. I’ve written in a lot of detail about many of them. You shouldn’t ignore your feelings around them, but you also shouldn’t close the door on AI entirely without trying to approach it ethically, safely, and collaboratively. Hopefully these tips from Bryan make it a little easier to do that.
