In just a few short years, generative AI has gone from the futurist fringe to the corporate mainstream. More than 80% of companies say AI is now a top priority to implement, and a majority of companies in some sectors (including 63% of telecom companies and 56% of cybersecurity organizations) utilize it in some fashion.
Those numbers will only increase in the years to come. Technology entrepreneurs like Kris Duggan — whose teams have been using AI longer than most — believe we’re on the cusp of a step change in worker productivity. Organizations that effectively and efficiently implement “gen AI” stand to gain a potentially decisive advantage over the competition.
Duggan works with dozens of tech entrepreneurs and executives to deploy AI solutions in both backend and frontend settings, helping them unlock their teams’ potential and do far more with limited resources. Here’s how he advises leaders to create “AI responsive” organizations built to stay one step ahead of the market.
1. Stay Up to Date on the Latest in AI Capabilities
The first thing Duggan tells tech leaders — and anyone else who asks — is to get up to speed on the AI state of the art and remain that way as the technology advances. This is fast-moving space, after all, and even a few months of inattention could mean missing significant capability leaps.
“Leaders can’t afford to be complacent about AI,” Duggan says. “It’s well worth the time it takes to remain up to date on AI capabilities because every new or improved AI use case is an opportunity to improve your teams’ productivity and deliver better results for your customers.”
Duggan recommends devoting at least two hours each week to an “AI review.” Most leaders already set aside time to digest the latest trends in their industry, so this may not be as much of an imposition as it seems.
2. Be Bold, But Don’t Get Ahead of Yourself
The current moment demands bold action from tech leaders (and leaders in other industries) to remain ahead of the curve on AI. However, the line between boldness and rashness remains, and it’s important to remain on the correct side of it.
We’ve already seen some examples of companies backtracking on plans to fully automate customer service with chatbots, for example. Those plans seemed like a good idea initially but soon ran up against the practical limitations of AI. That won’t be the case forever; within a few years, many expect chatbots to do everything human agents can. But we’re not quite there yet.
3. Ask Whether AI Can Do a Better Job (When It Makes Sense)
Goldman Sachs recently made news (and stoked controversy) in announcing Devin, the first fully autonomous software “developer” to be deployed in a major U.S. company. The tool is made by Cognition AI, one of the leading agentic AI platforms.
“We’re going to start augmenting our workforce with Devin, which is going to be like our new employee,” Goldman Sachs CIO Marco Argenti said.
Goldman determined that generative and agentic AI had advanced to the point that sophisticated models could combine to form an agent capable of doing most of the work of a junior software developer. Devin will not replace the company’s human developers anytime soon, but it will probably help Goldman hire fewer new devs than it has in the past. Eventually, future versions of Devin could replace more senior software people.
The lesson here is a potent one: Increasingly, the answer to the question of whether AI can do a better job than a human in roles like coding and video production is “yes.”
4. Hire Employees Who Embrace (And Understand) Generative AI
As noted, AI agents like Devin won’t replace human workers all at once. However, the human workers who will be most useful in the coming transition to an agentic workforce will be those who embrace and understand artificial intelligence. Like their bosses, they stay up to speed on AI developments and implement them into their work processes.
In fact, experts like Duggan advise hiring managers to update job descriptions and interview processes to reflect this new normal. It may no longer be worth it to hire an otherwise capable individual who isn’t enthusiastic about AI’s potential and fluent in its use.
5. Test AI-Enabled Strategies First
Like any other technology, AI isn’t perfect, and its implementation doesn’t always work out as envisioned. Remember the cautionary tale of the all-agentic customer service team that wasn’t.
Make sure your team is realistic about AI’s potential. Test it in sandboxed environments before deploying it more widely, especially when the deployment will reach customers.
This is a delicate moment from a reputational standpoint; companies seen as leaning too far into automation may suffer blowback. Indeed, studies suggest that 52% of workers are worried about AI replacing their jobs, and a similar number of experts believe AI will eliminate roles currently filled by humans.
6. Focus on “Seamless” AI Integrations
For now, focus on AI deployments and integrations that feel seamless to the end user, whether they are internal teams or your organization’s customers. The great potential of AI lies in reducing friction in the user experience and boosting productivity more generally, such as in personalized prospect journeys or self-service resources. If a particular deployment can’t achieve that, it may not be worth the challenges it creates (including the aforementioned reputational issues).
Adapt for What’s Next
Recently, artificial intelligence models have begun to improve at such a clip that even those who build and test them express surprise and wonder at what’s possible — and what could be coming in the very near future.
Indeed, no one is quite sure where we go from here. Experts like Kris Duggan have been convinced of AI’s potential for some time, but they can’t predict the exact extent of the technology’s capabilities even one or two years hence, let alone five or seven or ten.
What seems certain is that AI will be able to do a great deal more in the future. Leaders in every industry have a duty to remain briefed on the technology and adapt their organizations for all possibilities. Soon, it could mean the difference between dominance and irrelevance.
