In the early days of unified communications, I spent a lot of time writing and talking about the need for organizations to focus on end user adoption, usage, and training. I presented and moderated sessions on this topic at Enterprise Connect for several years, and participated in countless webinars to help organizations develop user adoption strategies. In 2020, fellow No Jitter contributor Kevin Kieller and I conducted a BCStrategies study (sponsored by Vonage and Zoom) surveying end users to identify the role of user training and adoption in driving usage of unified communications and collaboration (UCC) tools.
What goes around, comes around ... With the rapid rise of AI technologies, notably generative AI tools including copilots, assistants, and companions, it’s time to revisit the important role that user adoption and training plays for organizations deploying these technologies.
Adoption Doesn’t Just Happen
When I ask most vendors what they’re doing to help customers with end user adoption and training for their AI solutions, particularly for generative AI, as well as copilots and assistants, most reply that their tools are intuitive and don’t really require any training, and that adoption happens naturally. The same was said about UCC technologies over the years, but our BCStrategies 2020 survey proved otherwise. For example, our survey found that:
- Two-thirds of end users who received adequate initial training found that the UCC solutions made them more productive, while ALL of those who did not receive adequate initial training felt that UCC capabilities did not make them more productive.
- Almost three-quarters of end users felt that more UCC training would help them be more effective.
- Organizations that began internally marketing and promoting the new UCC solution several weeks prior to deployment were more likely to have their user adoption expectations met compared with those that didn’t internally promote the system before deployment.
We noted that in order to get the full benefits and return on investment in UCC systems and services, organizations must invest the time, money, and resources in user adoption programs supported by on-going training programs. The same is true for AI/generative AI tools and technologies.
Damnit Jim, I’m a Doctor, Not a Prompt Engineer
Prompts are key for generative AI, as this is how people give instructions and
speak with an AI, generally using natural language such as writing an instruction. Prompt engineering is used to tailor or refine the prompts in order to get the AI to generate the optimal or desired results. However, most users are not prompt engineers trained in how to properly phrase the prompt inputs to get the best output.
While watching several Enterprise Connect keynotes highlighting generative AI copilots and assistants, I kept thinking to myself, “This is really cool, but how are end users supposed to know what to ask the copilot to do? Or how to do it?” We saw demos of the copilots/assistants not just summarizing interactions -- which is becoming table stakes -- but creating charts and graphs, building a project plan, drafting and refining emails, and more. These generative-AI based assistants can be used by knowledge workers to prepare for a meeting, or by sales managers to research competitors or respond to an RFP, or by contact center managers to identify key metrics and trends – the list goes on. That is, they can be used once the user knows what directions to give the copilot.
Some of the vendors offer prompt templates to help identify what you can ask the copilot/assistant to do, but if these tools are as powerful as the vendors claim, then more is needed to help end users identify what to ask and how to ask it. Perhaps even more important, most end users don’t know how to ensure personal and corporate privacy and security when using generative AI and need training and assistance to prevent private information from being shared.
To get the most out of their AI investments, organizations need to help end users understand what the AI can and can’t do, what it should and shouldn’t do, and what they will have to do get the desired results and output. While simple functions like summarizing a meeting, identifying meeting task assignments, or drafting a chat response may not require user training, the more complex functions certainly require some basic training on how to create effective prompts, as well as security precautions that must be considered.
Microsoft Setting the Example
User adoption and training strategies are clearly needed, and the company taking the lead in this area is Microsoft. Microsoft recently introduced a Copilot Success Kit, which focuses on user enablement (preparing the organization and employees for the AI transformation journey) and technical readiness (addressing technical deployment and optimization including governance, security, compliance, and management). The kit is extremely comprehensive and includes readiness assessment tools, insights on how to identify the right scenarios and use cases, prompts to get started with, internal marketing materials to engage users, and much more. Below is the executive summary for how to bring users onboard with Copilot.
I encourage all vendors with AI offerings – and especially those with generative AI offerings -- to focus on end user usage, adoption, and training to help customers get the most out of these tools. This will help ensure that organizations receive the full benefits and return on investment in their AI solutions, while helping to promote security, privacy, and compliance.
This post is written on behalf of BCStrategies, an industry resource for enterprises, vendors, system integrators, and anyone interested in the growing business communications arena. A supplier of objective information on business communications, BCStrategies is supported by an alliance of leading communication industry advisors, analysts, and consultants who have worked in the various segments of the dynamic business communications market.