Certainly, to nobody’s surprise, artificial intelligence (AI) dominated last week’s Enterprise Connect 2024 event, from presentations on the keynote stage to demos on the expo floor. Since a year ago, most vendors have shifted from speaking of the art of the possible, particularly with generative AI, to today’s deliverables (along with plenty more to come).
Along with product availability has come increasing recognition of AI’s value for work purposes. We know this from conversations at Enterprise Connect, as well as from our recently published AI for Business Success 2024-25 global research study with nearly 700 companies. In this study, 82.3% of participants say they’ve become more accepting of AI’s value for their workplaces. More telling yet is that most companies are already using or planning to adopt AI for customer and employee interactions this year.
The study, conducted in December 2023 and January 2024, shows rapid adoption of generative AI for both employee and customer interactions, with more than 40% actively using today vs. the 27% of companies using generative products in June of last year. The top uses today are generating content for customers and for marketing purposes, with reviewing open-ended comments from customer feedback and generating content for agents to use during conversations being the top wish list functions.
Of course, none of this is to say that companies are throwing caution to the wind on AI. Rather, most companies are applying AI in a few areas for employee and customer interactions or are running pilots. (Worth noting, however, is that this study’s research success group, as determined by improvements across a variety of business metrics, is twice as likely to use AI extensively for both types of interactions than all other companies.)
Particularly for generative AI, the quality of the data set and data privacy remain top concerns. And most companies aren’t comfortable letting generative AI tools do their thing without some babysitting. Less than 15% of companies say generated content requires no vetting prior to using it, whereas 81.2% say either all (47%) or some (34.2%) content must have human oversight before using.
Companies not only are requiring human oversight when generative AI is using external data, but also when it is pulling content from their own knowledge base—and even when that knowledge base is managed and known to be accurate. Nearly 49% of companies still feel human oversight is mandatory each time content is created using data from their knowledge base, with 37.3% saying it’s necessary for some content.
Among other factors most companies are considering as they plan for the use of generative AI are service-level agreements (SLA) from their vendors. Among the nearly 62% of companies that either have or expect an SLA for generative AI, upwards of 71% are looking for guarantees on speed of response and response accuracy. Other desired SLA components are uptime (49.1%), reimbursement of business loss (45.5%), and latency (44.5%).
And while 62% of companies have or expect an SLA for generative AI, 42.4% say they won’t work with a vendor on AI without one. SLA is the fourth must-have for companies before working with a vendor on AI, behind analytics tools that show accuracy of AI responses, experience training models, and integration with multiple data sources.
As vendors flesh out their AI roadmaps and build out their AI portfolios, enterprise organizations are expecting them to do so with rigor. As do we. (Watch for Metrigy’s assessments of AI-related Enterprise Connect product announcements for employee and customer engagements, to come soon!)
