Introducing Copilot Chat: Get a head start on new capabilities

Are you intrigued by artificial intelligence and autonomous agents but unsure where to start? With Microsoft’s recently announced Microsoft 365 (M365) Copilot Chat, you may be able to get both of the above in a simple, single package. This new offering gives organizations the ability to build out AI-powered agents in an easy-to-use format.

Accessible through either the Copilot app or your Teams environment, Copilot Chat promises to bring agentic capabilities to users’ M365 workflows.

Here are the basics of Copilot Chat and what you need to know to get a head start on these new agentic AI features.

What are agents, and how does Copilot Chat work?

Traditional chat bots are known for their simple knowledge-based response and directory capabilities – you ask them a question, and they give a pre-written response or a link to where the solution to your inquiry may be found.

While AI agents can still give these traditional responses, Copilot Chat agents are, unlike their predecessors, task-based. They chat with users, then perform automated actions with data pulled from text files, SharePoint, and external websites. Your users and IT teams can customize and create agents for individual use cases or personas, making them a fast and versatile productivity tool for knowledge workers.

Copilot Chat agents live within Copilot Studio and are secured with the Enterprise Data Protection (EDP) and management controls available in your M365 subscription.

Practical use cases for Copilot Chat

Here are just a few examples of ways your users and teams can take advantage of Copilot Chat:

Call centers: Copilot Chat helps frontline service advisors reduce time to resolution (TTR) by taking simultaneous actions across a wide range of systems and inquiry-relevant information.

Sales professionals: Copilot Chat can help your sales leaders more effectively prepare for client calls by quickly reviewing information on past client interactions in your connected CRM system.

Helpdesk: With Copilot Chat, helpdesk employees can quickly resolve technical issues by accessing troubleshooting guides, knowledge base articles, and previous support tickets with minimal effort. When a user reports a problem, Copilot Chat can use natural language to search the system for similar issues and their solutions. This method reduces the time spent diagnosing and fixing problems.

Considerations for implementation

If you’re interested in deploying Copilot Chat for your users, consider the following to ensure you start at the right scale, for the right price:

1. Data estate

Copilot Chat’s actions are limited by the data available to them. When a user logs in through Entra ID, they’ll have the capability to create agents. However, if you haven’t organized or prepared your data to handle certain tasks, the agents may not perform to their full potential.

Properly housing and tagging your data will enable your agents to work more effectively and provide your users’ desired outputs.

2. Pricing model

Users with M365 Copilot Chat licenses can access their AI agents under a consumption-based pricing model. Simple answers pulled from the web are free while pre-written answers, generated responses, and automated tasks are metered.

These metered interactions with your agents will either:

1. Pull from an allotted message count assigned to an environment.
2. Operate on a pay-as-you-go basis at a cost of $0.01 per message.

You can choose either option through Azure. Give careful consideration to which model works best for your organization before deploying Copilot Chat to your teams.

It’s important to note that users with M365 Copilot licenses do not pull from metered message counts. Use of Copilot Chat is included with their licenses without any extra consumption costs.

You can manage your costs and better understand your AI agent consumption through the Power Platform Admin Center.

3. Usage rates

Because of Copilot Chat’s consumption model, it’s important to continually monitor its usage. If certain teams are high Copilot Chat users, consider if other Copilot licensing options might better fit their more intensive needs.

How SHI can help you get a head start

Deploying generative AI is never as simple as flipping a switch, no matter the scale. Data governance, licensing options, and end-user training all make the difference between a fully realized suite of AI agents and a lackluster chatbot that can’t perform to its true potential.

Our experienced Microsoft and AI experts can break down your options and prepare an AI deployment that meets your goals. Leveraging our proven Copilot Enablement model, we can:

• Guide you and your team in determining which Copilot offering best suits your needs.
• Help you understand the true cost of implementation and maximize your return on investment.
• Ensure your data is ready to fully leverage the agents you and your organization create.
• Establish a governance plan.
• Support training and adoption to get the most out of Copilot Chat.

To learn more about Copilot Chat or get started with your own Copilot Enablement program, contact our experts today.

As Presales Infrastructure Solutions Engineer within SHI’s Advanced Solutions Group (ASG), Zack Shelton has more than a decade of experience in industries including healthcare, manufacturing, financial services, and technology. His background has equipped him with a deep proficiency in Robotic Process Automation (RPA), AI-driven applications, and more.