8-minute read

Contents

What is Microsoft Copilot?

Assessing your Copilot readiness

The Copilot maturity framework

Out-of-the-box or custom solution?

Beyond implementation

Navigating your way to a “Copiloted” future

Quick summary: With the rollout of Microsoft 365 Copilot to enterprise customers, now is the time to assess your organization’s readiness and choose the best path for operationalizing generative AI.

On November 1, 2023, Microsoft 365 Copilot—the tech giant’s artificial intelligence add-on for Office workspaces—became generally available for enterprise customers. As was the case with Microsoft-funded ChatGPT just under a year ago, this application stands ready to become a game-changer on the business technology scene. So, what does that mean for organizations looking to tap into its productivity-driving benefits?

As with any new technology—particularly AI-driven applications—businesses are well-advised to avoid the temptation to “jump right in” and adopt a strategic approach that aligns with their business goals and values. In this article, I’ll share some best practices from our own Copilot readiness assessment process, insights from our maturity model, and key considerations on two options for implementing generative AI in your organization.

What is Microsoft 365 Copilot?

Built using generative AI models that were co-developed with OpenAI (creators of ChatGPT), Microsoft 365 Copilot is a virtual assistant that integrates seamlessly with Office applications such as Word, Excel, Outlook, and Teams. By combining large language models (LLMs) with data in Microsoft Graph and the Microsoft Office 365 applications, Copilot offers users a conversational interface while enabling them to automate repetitive tasks, with capabilities ranging from summarizing lengthy email threads in Outlook to creating a PowerPoint presentation from a simple prompt.

Assessing your Copilot readiness

At first glance, deciding whether to enable Microsoft 365 Copilot may seem like a no-brainer. However, the breadth of the application’s reach and the extent of its capabilities require a more measured approach, one that begins with an assessment of the organization’s readiness to implement such a powerful tool.

Logic20/20 has developed a framework for assessing Copilot readiness that calls for looking at implementation through a series of “lenses,” three of which we’ll explore here.

Lens #1: Access control configuration

To achieve its full potential as a productivity tool, Copilot requires access to all data in your Microsoft Graph, including data that you manage and govern via Microsoft Purview. In the absence of proper configuration of these ecosystems, Copilot could respond to user requests by surfacing documents that they lack authorization to access.

To avoid this “leakage,” businesses can configure their Microsoft ecosystems to increase result pertinence by controlling how much and what kind of data shows up in response to user queries. Data labeling and policy development will also be necessary for Exchange, SharePoint, OneDrive, and other locations.

Lens #2: User adoption and training

Using Copilot involves a substantial change in day-to-day work habits, and successful implementation requires far more than a couple of training sessions. This is where Logic20/20’s change management experts play a vital role in accelerating our clients’ progress towards adoption goals, developing and delivering strategic assets such as

• Communication plans
• User feedback channels
• User analytics and insights
• Continuous improvement strategies

Lens #3: Infrastructure readiness

If Copilot is to be implemented successfully and securely, the technology infrastructure must be ready to accommodate it. Achieving this readiness requires having elements in place such as

• SharePoint configuration, including document libraries and lists, permissions, and security measures
• Data cataloguing and governance
• API integration and access
• Load testing and capacity planning
• Data backup and restore procedures

Our assessment framework also encompasses a Copilot governance model that addresses ownership and accountability, establishment of a Copilot steering committee, decision-making frameworks, and continuous governance review. To help ensure the organization stays on track, we determine key performance indicators (KPIs) that encompass metrics for determining Copilot success, data-driven decision making, regular ROI assessments, and continuous adaptation based on ROI insights.

The Copilot maturity framework

We developed a maturity framework to give businesses the opportunity to assess their current status in the Copilot readiness journey. Finding out where you are on the progression provides a snapshot of where your organization is today in relation to where you want to go.

Level 1: Awareness

At this level, the organization has a limited understanding of Copilot, and minimal infrastructure adjustments are made, mainly to accommodate basic usage. There is limited awareness of data governance implications, with minimal data usage, and basic training programs are introduced to create awareness and basic competence. The business uses Copilot primarily for individual tasks or small projects with limited datasets.

Level 2: Basic preparation

Here the infrastructure configurations begin to align with Copilot requirements and initial data governance and compliance considerations are addressed, with policies starting to take shape. Basic training programs for users are launched to provide fundamental skills. Preliminary integration and testing with existing systems begin, with data integration primarily confined to specific departments. Expansion beyond departmental silos also begins, with data from different units integrated for some use cases.

Level 3: Intermediate implementation

In organizations at Level 3, infrastructure is well-configured to support Copilot, with increased processing power and storage. Data governance policies and compliance measures are more structured and actively monitored, and comprehensive end-user training programs are in place, tailored to different roles. Integration with existing systems is in progress, with data sharing across various lines of business. Basic security measures and identity management are in place to ensure responsible usage.

Level 4: Advanced optimization

At this stage, infrastructure is fully optimized for Copilot capabilities, with a focus on scalability. Robust data governance and compliance practices are adopted and consistently enforced. Advanced training programs are offered for specialized user roles and deepening skillsets, while integration with existing systems is completed and fine-tuned, supporting more advanced use cases. Security measures become comprehensive, including advanced threat detection, and identity management is streamlined. A clear roadmap is established for scaling up data usage, focusing on security and compliance.

Level 5: Mature innovation

When the business achieves full maturity, infrastructure is being continuously adapted to accommodate emerging Copilot features, with a focus on agility and innovation. Data governance and compliance are consistently refined to ensure ethical and responsible data usage. Continuous advanced training and support mechanisms are in place to foster innovation. Deep integration with existing systems and custom developments drive data-driven innovation and automation across the organization. State-of-the-art security measures, including AI-driven threat detection and advanced identity management, are upheld to protect sensitive data. The organization is using Copilot for data-driven innovation and automation at an enterprise scale, with a focus on responsible and ethical AI practices.

Out-of-the-box or custom solution?

As Microsoft continues to expand the Copilot rollout, many businesses will be attracted by the apparent ease of implementing the application out of the box. However, as we explored above, a structured approach, supported by a readiness assessment and testing, is required for this kind of implementation to be successful.

Another option available to organizations is to use generative AI to develop custom solutions on smaller and more closely controlled sets of documents. This knowledge discovery solution can extract insights and answer questions from pre-loaded documents. With this enterprise-grade approach, lines of business can unlock a wealth of knowledge and ensure data remains confidential at the same time.

This approach opens up new horizons for enhancing operational efficiency and customer service. Beyond its versatile applications in employee handbooks, benefit guides, and project documentation, this cutting-edge technology finds an ideal home in the realm of customer support and FAQs. Chatbots trained on existing support documents and frequently asked questions become a reliable and verifiable source of information, providing swift and accurate responses to customer queries with access to the actual document providing the answer.

Because custom solutions enable businesses to hand-select documents that are safe to make available to all users, a security layer is usually sufficient for restricting access. Additional advantages to this approach include accelerated speed to innovate, greater flexibility, and more localized control for the line of business.

Beyond implementation

No matter where the organization finds itself on the maturity scale, laying the necessary groundwork and viewing the initiative through the appropriate lenses can help you take full advantage of everything the application has to offer … but the journey doesn’t end there.

Microsoft will continue to release new Copilot features, so the infrastructure must be ready to adapt at any given time. Data governance and compliance must be consistently updated and refined, and continuous advanced training and support mechanisms must be in place.

Navigating your way to a “Copiloted” future

As is the case with all new technologies, implementing Copilot successfully—and securely—requires a strategic approach. Assessing the business’ maturity level at the outset can pave the way for a targeted roadmap, and deciding between out-of-the box and custom applications can save considerable time and stress down the road. By following the practices we explored here, businesses can ensure that the benefits of Copilot—or any AI technology—will continue for years to come.