Organisations are viewing AI as the cure-all for their business problems—so, when AI results disappoint, or when AI utilisation numbers don’t seem to warrant the investment, the blame is placed squarely on the technology. More often than not, however, the error lies with humans.
While some businesses have invested rigour into leveraging AI to solve specific problems, far too many organisations have simply provided AI capability to their user base and allowed them to ideate how to best employ it. While early adopters in the workforce find the latter approach highly appealing, most business users are unwilling to invest the time and energy to experiment on their workflows with AI. As a result, after having a go at using the capability, they simply pile AI in the dustbin of infrequently used tech.
Deloitte, in its 2024 Year End AI report, noted that nearly one in five businesses surveyed cited their top three AI adoption blockers being the difficulty identifying use cases, the lack of an adoption strategy, and cultural resistance from employees.
Fortunately, by tweaking their approach and adopting a few best practices for how to interact with AI, business professionals can better deliver the positive outcomes they seek from this innovative technology.
A quick caveat
When discussing AI adoption, it pays to qualify the scenario. The world readily accepts that AI performs well at very specific tasks such as transcribing and summarising video meetings, automatically providing short replies to email threads, or analysing facial features to provide eyewear recommendations. However, adoption inconsistencies tend to emerge when organizations deploy the popular user-prompted desktop AI assistants. This is my focus.
What’s your problem?
One of the biggest mistakes enterprises make is assuming that AI is going to be a genie who automatically makes “anything” happen — that’s not how it works.
The first step in the process is to home in on business problems. Sounds obvious, but many organisations—in their rush to get their hands on AI and get it deployed throughout the organisation—skip this crucial step. Pause for a moment to ask: What business problems should we try to solve with AI? Are we trying to increase revenue? Reduce costs? Improve customer service? Reduce labour?
If you don’t know what you’re trying to achieve with the new technology you’ve purchased, that’s not AI’s fault—that’s a shortcoming on the human side. AI can’t figure that out for you, so you need to do that work on the front end.
Most users don’t want a blank canvas
As with any other technology, AI doesn’t operate in a vacuum. Organisations have to think about their people, the context in which those people work, and—most importantly—how those people can actually use a tool like AI in the context of how they work.
So, once an organisation has homed in on a particular business problem, they should break down the workflows associated with that business problem into their component parts and ask: Where can I use AI to assist my typical end user?
AI is great at tackling tasks that are time-consuming in nature, such as comparing a proposed contract to a collection of similar contracts a business has executed to surface the key differences and risks. A task like this might take a lawyer hours or even days to complete, but it could be accomplished in a matter of seconds with AI, and it would account for massive savings in resource time. In this case the technology does the heavy lifting, but value was created when the heavy lifting was applied to a measurable problem.
Another example is deriving insight from data collections. An operational manager in a large enterprise has several data sources that they regularly need to run reports against, synthesise the data, and then derive and present insights to upper management. Frequently, this task is a manual endeavour, and it can take days to produce a quality output, depending on the size of the data stores.
Fortunately, this is a narrowly defined task where AI can shine. AI can quickly scan all the data repositories to serve up insights—including some that the human might not have even come up with on their own.
For tasks like this, the human will still be in the loop to review what the AI has come up with, but the AI will drastically reduce the amount of time it takes to get a specific piece of the overall workflow done. That counts as a win.
The upshot? When businesses isolate specific problems and leverage AI as part of the solution, it tends to generate good outcomes. What does not tend to generate good outcomes is giving a business user access to AI and telling them, “Okay, have a go at it! Improve the way you work.” In other words, a blank canvas approach is a recipe for disappointment. This is a situation where it pays to be narrow-minded.
Lend the “average user” a hand
As they get started on their AI journey, organisations would also do well to remember that unless they take a thoughtful approach, AI will likely follow the same technology adoption curve within their organisation that every other new tech tends to: There will be passionate early adopters, but the vast majority will use it a couple of times and then drop it in the metaphorical dustbin.
For example, your average business user just doesn’t have time in their day to experiment with different prompts and different phrasings to try to get just the right output from generative AI.
As a result, they may not give the AI enough information or context to generate meaningful results, and they’ll throw up their hands in frustration and say: “You know what? I see the potential in this technology, but I’m not sure this is worth my time right now.”
It’s up to the organisation to lend the average business user a hand and “connect the dots” for them. An easily accessible prompt library or one-click prompt “buttons” that perform a single, well-defined task can make interacting with AI much more approachable and fruitful.
Again, a wide open “play around with it and see what happens” approach is not the way to go here—a little guidance goes a long way.
A foundation for success
Ultimately, the key to success with AI lies not in blaming the technology for its shortcomings, but in empowering humans to leverage its potential with purpose and precision. Organisations that nip human error in the bud will be avoiding many of the pitfalls that can hamper successful collaboration between end users and AI, helping to unlock a future of innovation and efficiency.
