In O’Reilly’s latest study of AI adoption in the enterprise, “lack of skilled people and difficultly in hiring” topped the list of the greatest barriers that organizations face in 2021. This won’t come as news to many IT leaders—indeed, they might be surprised it hasn’t happened sooner.
Industry analysts and consultants have been talking for years about the demand for tech skills outstripping supply, with Korn Ferry notably estimating the world will be short 4.3 million tech, media and telecommunications workers by 2030.
Has the pandemic widened the gap?
Almost certainly. Artificial intelligence, in particular, has proved key to solving many of the business problems surfaced by COVID-19. Right now, AI is helping brands to extend, personalize, and secure their digital services, to support remote working, even to combat the wave of fraud unleashed by social and economic disruption.
But there’s another reason LinkedIn is being flooded with vacancies for tech professionals: the pandemic has caused many organizations to reconsider the business value of agility. And developing your agility doesn’t just mean embracing intelligent, flexible, digital solutions. It means having the skills and tools to adapt them on the fly without always having to wait on an external vendor.
All this begs the question: how can enterprises harness AI effectively when hiring relevant talent is so hard, and being reliant on a vendor’s timelines and responsiveness is, for many, an increasingly unattractive prospect?
Carve a path between “DIY” and “Don’t DIY”
Very few organizations could afford to build the machine learning infrastructure that key applications of enterprise AI depend on, for example, providing conversational customer experiences through sophisticated virtual assistants.
But thanks to a perfect storm of technological innovations and trends—Cloud, APIs, microservices, open-source, CI/ CD methodologies—few organizations need to. Instead, it has become possible for brands to tap into advanced AI technology and build their own experiences around it.
The “DIY” path can look incredibly tempting for organizations looking to own their AI initiatives and create new solutions at pace. But as we’ve seen, while the technology may be readily accessible, the skills aren’t. For example, to create a successful conversational self-service experience, you need AI specialists, speech scientists, and experts in conversational design. You need the talent that’s still so hard to come by.
One increasingly popular strategy is to seek a middle path between innovating with AI on your own and asking the third party to do it for you.
Choose your AI vendor not just on the quality of their technology—which is still absolutely key—but also on their ability to be a trusted, collaborative partner.
Look to them to provide development and design tools with best practices baked in—whether that’s through drag and drop interfaces, low or no code, or pre-built models that immediately set your team on the right track. Make sure they can provide tutorials, training and support forums. And, perhaps most importantly of all, make sure they can parachute in the specialist skills you’ll struggle to develop or hire.
Democratize the Software Development Lifecycle (SDLC)
It takes a village to raise a successful, AI-driven solution. And that can make it very slow work. From developers to test engineers, from infrastructure architects to business stakeholders—if everyone works in their own silo, the progress will be glacial, and the results will be uneven and underwhelming.
Ideally, everyone who touches the SDLC should be working on the same platform, each with tools appropriate to their role. This way, you’ll accelerate time to market and make sure your AI-driven innovations are inextricably tied to your business’ needs.
Make the most of the time and resources you do have
Another shortcut to successful AI-based services? Understanding the resources you already have and finding ways to avoid duplicating work.
Let’s say a business is looking to develop a sophisticated virtual assistant (VA) based on conversational AI to provide services to customers visiting its website. It can get a head start by analyzing the questions customers already ask on its voice channels and using this insight to shape and accelerate the development of its new VA’s capabilities and Natural Language Understanding model.
The business in our example should also think carefully about whether it will want to replicate that conversational experience later on—for example, to offer it to customers who get in touch through Facebook Messenger rather than through its site. If so, building the experience on an omnichannel platform that allows it to be tweaked, tailored, and redeployed could save a huge amount of rework in the months to come.
What to Ask AI Vendors, and Why
As we’ve seen, rapidly building agile, AI-driven technology solutions in the midst of a skills shortage is absolutely possible. You can reduce the level of AI expertise your own IT professionals need by giving them tools designed to make following best practices easy. You can take a hybrid-DIY approach, bringing in specialist skills when you need them. You can avoid silos in the SDLC and save duplicated effort.
With that in mind, here are a few key questions businesses should ask when evaluating potential partners:
- What skills will we need to make our project a success? If we can’t hire them, can you provide them?
- Do you provide a unified platform that our teams can use at every stage of the SDLC? What tools do you provide? Are those tools low or no code?
- What training/tutorials will we be able to access? What support?
- Can we decide where the applications we create are hosted, or are we tied to your own cloud?
- How do you work with other organizations? What results are they seeing?
To put these strategies into action, organizations need the support of technology vendors who can not only provide enterprise-grade AI and tools, but expert guidance and scarce but essential skills.
[To share your insights with us, please write to email@example.com]