Generative AI: Two Truths and a Lie
Joe is a writer, engineer and technologist who specializes in translating intricate IT concepts into accessible narratives. A seasoned B2B writer, his expertise spans from SaaS and cybersecurity to machine learning and AI. As a contributor aligned with Tech Journal's ethos, Joe bridges the gap between innovation and understanding for IT pros.
Generative AI has fully captured our collective imaginations like no other innovation in recent memory. Business leaders, tech enthusiasts and more internet gurus than you can shake a stick at are chomping at the bit to harness this powerful new engine boasting near limitless potential.
“This is the fastest growing technology,” said Insight CTO David McCurdy when asked to share his perspective on the current boom in generative AI. “It’s actually delivering real value right away. So the hype cycle is real.”
82% of organization leaders agree “companies must invest in digital transformation or be left behind,” according to a recent IDC InfoBrief sponsored by Insight.
To that effort, generative AI seems like a breakthrough technology on the verge of propelling innovation and industry by incredible leaps and bounds.
But how?
This article sheds light on two truths and a lie surrounding generative AI — a tool with the promise to revolutionize the very fabric of our digital interactions.
Well aware of the vast potential that generative AI offers, many leaders are still unclear about the real-world business applications. Meanwhile, informed and eager board members and stakeholders want to know: “What’s our generative AI strategy?”
According to a recent Harris Poll, business leaders believe generative AI can create better business outcomes through greater employee productivity and customer service. 90% of industry leaders agree that generative AI can enhance a wide range of roles.
Generative AI uses machine learning and Large Language Models (LLMs) to create information — including text, images and music — that is similar to human-created content.
However, concerns around issues such as quality control, ethics and security remain stubbornly high.
51% of those surveyed in the same Harris Poll highlighted quality control and misinformation as top concerns surrounding their AI adoption goals.
The lack of a well-defined path forward has stalled many would-be pioneers eager to explore the untapped potential of generative technology.
A common thread among technology leaders who are experiencing the most success adopting generative AI is the creation of an AI policy. This serves to guide their forward momentum.
Commenting on the steps Insight took while building their in-house generative AI tool, McCurdy confirms, “One of the very first things we did is we created an internal policy for our teammates and communicated very clearly how we wanted them to interact with the tool.”
A thoughtful generative AI policy tackles top concerns and outlines how your team and AI will work together to deliver the best returns on investment. Crafting your own set of guidelines to securely and responsibly use generative AI puts your organization on a fast track to outpace competitors struggling to grasp an AI advantage.
Energized by the expanding list of real-world use cases, industry frontrunners are now looking beyond policy to the next steps in their AI journey.
We’re in the primitive stage of generative AI. Like all emerging technologies, this first iteration is — if you can believe it — feeble and dumb.
What makes this generative AI disruption even more fascinating is we’re still discovering the limits of what’s possible with generative technology. Generative AI is unique because you don’t identify a problem and then seek generative AI as the solution; you find ways to use generative AI.
No. Don’t even know where to begin.
Off the record, yes.
A little. Our organization allows it but isn’t promoting its use.
Yes. Our organization is encouraging all teammates to take advantage of it and experiment.
Unlocking the real potential of key generative tech applications:
One of the profound revelations to date is the potential for generative AI to increase the efficiency of humans bogged down by mundane or repetitive tasks. Tasks like summarizing dense materials (or long email chains), coding or even addressing customer inquiries can now be accelerated.
At the individual level, generative AI can shave seconds and minutes off a task. Compounded throughout the course of a day and across the organization, you’re now looking at considerable efficiency gains.
Insight CEO Joyce Mullen paints a vibrant picture of generative AI’s real-world reach.
“With tools like generative AI, I know we’re going to be able to eliminate a bunch of these repetitive task work hours and replace them with much higher-value work hours,” she says.
During a time when staffing is tight and employee burnout is real, an assist from generative AI can be a game-changer.
Insight CISO Jason Rader highlights generative AI’s ability to deliver return on investment with a client success story using Insight’s in-house version of ChatGPT. “They said it used to take them 30 minutes — now it takes them two. So I mean, that’s like legit ROI.”
The rise of technology has brought many opportunities. It’s also exposed a gap in talent, especially in niche IT areas. Generative AI can help IT teams by checking code, running specialized tasks or documenting complex processes — all minimizing the impact of talent shortages.
Generative AI is increasing productivity for software developers by minimizing manual input for tasks like code quality testing, debugging, documentation and onboarding new developers.
“We believe it’s a 56% increase in productivity for software developers,” says McCurdy. “We think that figure is probably low, but that is an incredible increase in productivity.”
Delving into vast datasets, pure processing power can unearth patterns and insights with unparalleled precision. The ability of generative AI to comb through data, discern patterns, and extract key insights and put them into plain language is nothing short of revolutionary — this is what Microsoft Copilot is all about.
Copilot combines GPT-4 AI with Microsoft apps and data from Microsoft Graph to rocket-boost user productivity.
You no longer need to be a data expert or know code to deliver meaningful analysis and workflows.
Story continues to page 9
Creating realistic scenarios and case studies for training purposes makes generative AI a valuable asset for improving skills faster than ever before.
Sharing a use case from an Insight warehouse team, McCurdy notes, “They used our InsightGPT solution to basically learn, ‘Hey, how could I solve this problem better?’ It taught them how to record scripts to automate this action. We showed this use case to some of our customers the other day, just in terms of the power, and you could see eyes light up!”
Crowdsourcing allows organizations to capitalize on the creativity of both experts and regular users. The nature of generative technology means one use case can lead to another, and so on, creating a cascading effect on new discoveries.
“You’ve got to give the technology out to folks and then see where it takes them,” says McCurdy. “We’ve seen organizations putting a generative AI tool in the hands of their teammates to discover how they can use it in their day-to-day jobs. People are finding their own uses — things that nobody would’ve thought of in a boardroom or in a conference room.”
What’s so exciting is the community-driven nature of these discoveries. Sharing crowdsourced examples creates enthusiasm and engagement across the organization. Crowdsourcing solutions also works to accelerate learning and the adoption of generative AI.
“This is going to help people do things that weren’t possible without this kind of technology. Even, what, six months ago, 12 months ago,” points out McCurdy. “This is going to be the power of acceleration that we’re going to see in terms of solving issues.”
It may be tempting for leaders to think, “We'll just wait and see — we can always catch up later.”
But McCurdy warns there’s tangible consequence for not keeping up with this trend.
“Your job won’t be replaced by generative AI — it will be replaced by someone who uses generative AI,” says McCurdy.
The same is true for businesses — and IT decision-makers know that. 49% say the ability to keep up with technological innovation compared to competitors is one of the greatest threats to their organizations over the next 12 months.
The rapid pace of emerging competitive challenges sends leaders a clear signal: evolve with the times or be left behind.
It’s important to dispel the widespread illusion that AI’s adoption means the end of human input and creativity. Rather than replacing us, AI is positioned to become our most potent ally. Software organizations, such as Microsoft, are leaning into the concept of a generative AI “copilot” — confirming this is designed to be a collaborative technology.
However, this flourishing partnership will require broad-based education to reach its full potential. Prompt engineering has quickly emerged as a key skill set to improve the quality of business outputs.
The upside is that anyone can be an exceptional prompt engineer — no coding skills required. Insight launched an internal training series to help all employees, from architects to accountants, learn to write effective prompts. (See our handy cheat sheet on page 13-14).
Insight workplace specialist Roland Leggat predicts a bright future for human and AI partnerships.
“I firmly believe that the role of AI in our future will look vastly different from anything we can currently envision,” says Leggat. “I also believe that if we learn to embrace AI, it can be a great enabler for work-life balance.”
Generative AI thrives when coupled with human creativity. Where AI can process vast data and propose patterns, humans provide context, ethical considerations and nuanced interpretations.
It’s not competition; it’s collaboration.
These four key takeaways will help you — and your teams — get started on your generative AI journey.
1. Create a generative AI policy.81% of organizations are developing or have developed a generative AI policy. An effective policy will provide guidelines for data usage, ethical considerations, security protocols, and boundaries between AI creation and human intervention.
2. Prioritize your data quality.The bedrock of generative AI success is clean, organized data. Accuracy and effectiveness rest on the quality of the data the AI model trained on.
3. Highlight human oversight in AI-driven projects.The value of human oversight in AI-driven initiatives can’t be emphasized enough. For all its capacity to analyze and learn from information, AI lacks moral intuition and judgment.
4. Let people experiment.As noted previously, sometimes the best use cases for generative AI comes from unexpected places. With the proper security and guidance in place, encourage employees to play, test and share their successes. When compounded over time, even small assists from generative AI — like help summarizing a long email chain — can result in huge gains in efficiency, productivity and worker satisfaction.
Maximizing your results with generative AI requires a balance of quality data, robust policy, informed human oversight and experimentation.
For business leaders and IT decision-makers looking to harness the power of this technology, continuous education is a necessity. And working with a knowledgeable and experienced partner to guide you offers more than a competitive advantage. The risks of generative AI are real — but so are the returns.
Unlock the potential of generative AI. Insight can take your generative AI strategy further, faster.
Learn more
Get better output by structuring your prompts:RTF (Role, Task, Format)CTF (Context, Task, Format)RASCEF (Role, Action, Steps, Context, Examples, Format)
Megaprompts: Write one large message that contains all the structure points (RTF, CTF, RASCEF)
Prompt chaining: Break down the task into smaller steps. (Ask for a blog post idea, then an outline, then the content.)
In the realm of generative AI prompt craft, “context” is king. One of the most effective assets for providing context for tools like Chat GPT are “roles.” Roles help inform the AI model about how to respond to a request based on the knowledge and experience associated with a profession or job title.
Try the TKO formula:Title: (profession or job title)Knowledge: (specific area of expertise or knowledge)Objective: (objective or goal)
Data analyst: “You are a data analyst with expertise in business intelligence tools. You provide insights in a logical and detailed manner, aiming to help the user understand the importance of data-driven decision-making.”
Risk management expert: “You are a risk management expert with a background in assessing and mitigating potential threats. You provide advice in a proactive manner, aiming to help businesses safeguard their assets and reputation.”
Supply chain manager: “You are a supply chain management expert with a focus on lean methodologies. You provide advice in a systematic manner, aiming to help businesses optimize their supply chain processes and reduce waste.”
The key is to provide enough context in the role description for ChatGPT-like tools to stay on topic and adapt its output to suit your needs. Providing specific details improves the quality of the generated response.
