This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them. Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 8860726.
Generative AI Has Arrived – Here’s How to Assess Where It Fits In Your Enterprise
Generative AI has captured technology media headlines since the launch of ChatGPT nearly a year ago. The appeal of GenAI is that it can replace mundane tasks and improve productivity and efficiency; the concern is that this technology could be so disruptive, it replaces human employees entirely.
Even as we all sift through the benefits and threats that GenAI presents, communications and collaboration vendors have rushed to bring GenAI capabilities to their platforms, in some cases announcing previews before they were fully available. In the last several months the GenAI space has accelerated. Here are some of the products that have been announced or released:
- Google’s release of Duet AI in late August
- Zoom’s release of Zoom AI Companion earlier this month, available for free to all Zoom paid-plan customers
- Slack announcing Slack AI allowing developers to easily bring GenAI into workflows
- Microsoft’s announcement that its Copilot GenAI assistant will be generally available in November
These recent announcements follow releases, previews, and roadmaps from just about every other collaboration provider from Cisco to Dialpad, Notion to RingCentral, and more. GenAI capabilities range from assistants for content creation; auto-transcription, summarization, and action item identification in meetings and phone calls; and real-time and after-call coaching and feedback. Tomorrow, capabilities will continue to improve and expand.
As GenAI moves from preview to reality, IT leaders are faced with the challenge of understanding the value, as well as assessing the potential risk, and determining when and how to deploy GenAI capabilities to their workforce.
Key concerns these leaders should consider include:
Zoom shook the market when it announced that its AI Companion is available for free (to paid-plan customers), especially since competitors including Microsoft and Google had previously announced that their GenAI add-ons require additional cost ($30 per-user, per-month, in both cases). Even for free services, deploying GenAI is likely to require end user training and ongoing support as individuals learn how to use the tools. For companies evaluating paid add-ons, they must weigh the potential productivity benefits versus the new tools’ cost and identify personas where GenAI features deliver the most bang for the buck. Examples could include creation of sales and marketing campaigns and content, elimination of manual transcription (e.g. in healthcare settings), automated identification of next steps for customer service and sales calls, and much much more. Both vendors and customers must work together to identify specific use cases that can benefit from GenAI to justify licensing and additional operational costs.
GenAI models are trained on large language models to predict next steps, such as words that follow other words. Today, models such as ChatGPT train on publicly available information on the web. And while that’s extremely useful in enabling natural language search and creation, it doesn’t allow for companies to use GenAI for specific business cases that require analysis of their own data. In response, most vendors are offering support for customer-isolated LLMs that ensure that no customer data is shared in the public domain or with other customers. Companies wishing to leverage such models must ensure that their security, compliance, and governance capabilities are sufficient to analyze risk, ensure retention, and conduct proper audits.
Digging a bit further, GenAI add-ons create their own content, which for regulated organizations means classification, archiving, and eDiscovery support based on appropriate requirements. Compliance managers must take a proactive approach in identifying how their organizations are adopting GenAI to avoid risk of non-compliance.
(Editor’s Note: No Jitter’s Martha Buyer has more on the compliance and privacy risks presented by the data being employed by GenAI products.)
User Perception and Awareness
Metrigy recently released the results of a study of over 500 end-users of their perceptions toward Generative AI in the workplace. Most (85.1%) were familiar with GenAI and almost 15% were already regular users. Utilization skewed more toward younger workers and those with a college education.
Startlingly, just 12.9% of end-users surveyed said they fully trusted GenAI and almost 31% said they did not trust it at all. Trust factors include false results from hallucinations or poisoning of LLMs. Data here also varied by age, with 43% of those 45 or older not trusting GenAI versus just 18% of those under the age of 45 who said they did not trust it. When asked what would get those who do not trust GenAI to change their minds, the largest response was “human oversight,” though almost 32% said they would never trust it (just 21.9% said “Government regulation”).
You can’t say “no”
The reality of GenAI is that it does offer significant potential workplace benefits. While some companies have attempted to ban it, bans are likely to lead to employees going around IT and using it themselves through personal devices. This creates additional risk for organizations. As we’ve seen in recent years, companies that don’t have a way for employees to use approved messaging apps with customers often find that employees are turning toward public / consumer apps, putting organizations at risk of large fines for non-compliance. The same is likely to happen with GenAI for those companies who ban its use. Therefore, again, companies must take a proactive approach toward implementing GenAI in a manner consistent with security and compliance needs.
Islands of GenAI
With just about every vendor introducing GenAI capabilities, companies are likely to deploy multiple GenAI engines, potentially limiting the overall ability of GenAI to improve productivity. For example, a company using Microsoft 365 for content creation and management, and say Webex for calling and meetings, will find that they have two separate GenAI tools, each trained on separate data. This could lead companies to converge separate UC&C platforms but hopefully it also leads collaboration vendors to establish APIs that allow GenAI tools to share data with one another.
The Bottom Line:
Generative AI is quickly moving from concept to reality. It offers significant potential to improve productivity, but it also creates security, governance, and compliance risks, as well as the need to evaluate its benefits to determine ROI. And it requires continual user education to ensure that employees are comfortable with the technology and can obtain benefit from it. Before incorporating it into the existing technology environment, IT professionals should examine all the possible factors that GenAI affects.
About Metrigy: Metrigy is an innovative research and advisory firm focusing on the rapidly changing areas of workplace collaboration, digital workplace, digital transformation, customer experience and employee experience—along with several related technologies. Metrigy delivers strategic guidance and informative content, backed by primary research metrics and analysis, for technology providers and enterprise organizations.