Chatbots, those AI-infused assistants that can be used by enterprises for any number of time-saving and productivity-increasing tasks, are some of the lowest hanging fruit in any organization's quest to adopt generative artificial intelligence (genAI).

Beginning with OpenAI's ChatGPT in November 2022, chatbots have been increasingly adopted to automate various business processes and assist employees in performing mundane tasks.

Chatbots can be integrated through APIs with enterprise resource planning (ERP) and customer relationship management (CRM) backend systems to utilize proprietary data and automate data-intensive tasks such as updating customer records.

Enterprise chatbots can augment software code creation, summarize documents, help develop marketing and advertising campaigns, and automate helpdesk tasks and customer interactions. While not without error, the large language models (LLMs) behind chatbots are designed to learn and improve over time.

Research firm IDC refers to all genAI-enabled conversational AI applications that act as assistants as "copilots," though Microsoft copyrighted the term with a capital 'C' (Copilot).

IDC

Copilot is Microsoft's version of OpenAI's ChatGPT, an genAI chatbot that uses the same LLM (GPT-4) and can access the internet through Microsoft's Bing.

"IDC expects that copilots [lower-case 'c'] will proliferate over the next several years and will eventually become the standard interface for most software products, like web interfaces are today," the firm said in a new report.

Microsoft, a majority investor and partner with OpenAI, has integrated ChatGPT into most of its enterprise products under the name Copilot. And, to date, Microsoft has dominated the enterprise chatbot market through that product.

"Several vendors are preparing copilots that aren't mentioned here or were just released, including such vendors as Slack, Salesforce, Oracle, SAP, Cisco, and others," said David Schubmehl, IDC's research vice president for AI and automation. "We expect this to be a very dynamic market going forward."

Microsoft has in recent month released a wave of copilots for Microsoft 365, Microsoft Dynamics and even Microsoft Teams. This month, IDC released a snapshot survey of users from North America, Europe and Asia where AI assistants are leading in adoption.

When asked which genAI applications or applications providers they're working with or plan to work with today and in two years, most chose something other than Microsoft 365 Copilot, the company's most popular enterprise chatbot.

At first glance, the IDC survey of more than 600 multinational users appears to put  Amazon and its CodeWhisperer AI software code generator well ahead of the competition, which includes a second-place Google Gemini for Workspaces tool.

About one in six (16.4%) of survey respondents chose CodeWhisperer as the chatbot they're using today and 20% said they expect to be using it in two years. Google's Generative AI for Workspaces had 14.1% of users today, though that number dropped to 11.1% for expected use two years by 2026.

That's where the domination over Microsoft ends, however.

The next four copilot products chosen by respondents were owned by Microsoft, either directly or through a subsidiary, such as online developer repository Github.

Microsoft's 365 Copilot was named by 9.1% of respondents - a figure that changed only slightly in two years to 9.3% of users. GitHub's Copilot was used 10.2% of the time by respondents, but in two years that number would plummet to 3.4%.

Microsoft Dynamics 365 Copilot is used by 7.5% of users today, rising to 12.6% in two years, while Microsoft Security Copilot was chosen by 7.5% of respondents for use today and 7.2% for use in two years.

IDC's survey took place in October 2023, but the firm said the numbers would likely not have changed significantly since that time, and the current analysis of them remains accurate.

Taking all of Microsoft's Copilot products together - including GitHub's variant - the company's genAI tools were in use by 34.3% of respondents, with 32.5% expecting to be using the same products two years down the road.

Still, the top use for copilot-style tools remains for augmenting software code production, according to Schubmehl.

"The coding copilots are being used to generate code, analyze code and even produce documentation," Schubmehl said. "The other copilots are offering advice and recommendations as users work with enterprise applications."