Maturity Model for Microsoft 365 - Search Competency

Overview of the Concepts [tl;dr]

People search for many reasons. Any effective search strategy and supporting technology needs to reflect this and include a person-centric and organization-scoped approach to helping people find the things they need. Great search is about discovery and understanding, not the search experience itself, i.e., search is only as good as its results and the ability to assimilate them.

With modern organizations creating huge volumes of content and data every year, a search experience where users can find what they want, when they want is essential.

A good search experience benefits the organization by reducing time to find knowledge in the organization and use it to advance a process or insight. This becomes particularly powerful when users do not need to know where the content is stored. It helps reduce "re-inventing the wheel" and content duplication because the originals could not be found.

The emergence of AI-enhanced search technologies evolves search further. Search is no longer only about “finding” but also about synthesizing; search can summarize across multiple documents, extract key points, and answer in natural language. Search with AI can integrate traditional advanced Search, backend enterprise knowledge graphs, and large language models (LLMs) to deliver contextual, conversational answers rather than just lists of links and file previews. It can access disparate knowledge stores, or intentionally restricted to specific sets of content. Furthermore, it has multi-modal capabilities, allowing both input and output to use or combine text, voice, image and live video/vision.

Definition of this competency

Search is about enabling people to find the authoritative information within the organization easily using a set of keywords or search terms, through natural language conversations, or based in their activities and environment. The results may come from the Microsoft 365 platform or other systems which have been connected into the search process.

Evolution of this competency

See the Maturity Model for Microsoft 365 - Introduction for definitions of the Maturity Model levels.

People search for many reasons:

• They know something exists, but don't know where to find it
• They know something exists, but don't know how to describe it
• Search is the more efficient or rapid means of completing an activity
• They need to find if information exists or what information exists
• They need to find someone who can offer advice or skills
• They want to see what the organization has or knows

The evolution of Search starts from the basic 'index card' concept, which tells you where to find the document etc. you are looking for, epitomized by the Dewey Decimal system found in libraries. As technology developed, it became possible to search limited metadata (filename) in file repositories, then other attributes and eventually search engines were able to index contents (Semantic analysis), file properties and metadata/tags across multiple repositories. In parallel, the user experience of the search, especially for creating the query and presenting results has improved from basic or cluttered to strongly structured with previews and interaction points, plus post-search filtering or refinement. In parallel, the technologies have become aware of security and governance, reporting and feedback, content weighting and relevance (e.g. headings are more important than text), context, relevancy and 'freshness' (more recent content is likely to be more useful) and can deal with advanced content management technologies.

Search provides multiple 'experiences', often using one or more Indexes based on proprietary organizational data (semantic index), public data and 3rd party knowledge sources.

Search has evolved through the improvement of the search indexing and categorization processes using graphs, Artificial Intelligence, and Bing algorithms to build results which are personalized for each user. This enables more insightful results based on understanding of the context, where the search is performed and importance of the content.

Today we are moving into AI driven search which understands the person and provides very context specific results based on who they are, where they search from. This is supplemented by AI driven interfaces, including voice and image search. We can expect to see AI increasingly pervade search experiences combined with greatly enhanced personalization based on a wide range of context types.

This Search competency relies heavily on several other competencies including Collaboration and Information Architecture which enables more mature search capabilities within the organization. See the Maturity Model for Microsoft 365 - Introduction for definitions of the Maturity Model levels. Some characteristics should, perhaps, be addressed a little more urgently than others; we have marked these with the 'Sparkles' emoji: ✨

Level 100 - Initial

Initial level characteristics include:

100 Governance, Risk, Compliance & Security

• Out of the box search experience, the quality of results varies wildly with users often unable to find what they are looking without knowing the terms to search for.
• No formal process to curate search results or analyze search patterns.
• The danger of search finding inappropriate content is not well understood.
• Search often does not respect user privileges and access rights.

100 Technology

• Search may be restricted to File System search and a few specific applications.
• No AI based search has been deployed within the organization.
• Default out of the box configuration is in place; often with very limited capabilities (filename, title, date).
• No customizations have been made to Microsoft Search.
• Search indexes a small volume of organization content.
• There is no customized search experience to support specific business requirements.
• No enhancements are made to the search experience to aid the user.
• Configuration of authoritative data sources have not been configured to help relevance ranking.

100 User Experience

• The user experience is basic.

• End users are often unaware of search tools and have had no training. They are not familiar with using AI search tools and have not integrated them into their search regime.

• User's may not always find content that they want without knowing the right search terms.

• User confusion with the different ways of searching.

• Search is scoped to the current application; there is no global search, the search experience, presentation, and features vary widely according to the current application. Many systems have no search capability at all.

• Users use search as a last resort after asking someone browsing and other forms of discovery.

• Search requires users to know how to ask the right question, possibly with very specific syntax, query structure and case sensitivity.

• Users turn to search by default because the information architecture (navigation, site topology, taxonomy, etc.) don't assist them to find relevant content.

100 Impact

Users rarely rely on search; accessing known documents in known places (which are potentially superseded); they rely on browsing rather than search (failing to find the correct document); asking colleagues (consuming their time and attention) or creating new versions of content that already exists. Users frequently make copies of documents so that they 'know' where they are.

Most users turn to external search engines to search for information that probably exists in the organization.

In many instances Microsoft 365 usage is primarily focused on use of email. This content is unstructured, has minimal and frequently inaccurate document metadata beyond the filename and the content. User expectations are of a 'Google' search without context, scope, or organizational awareness.

At this level, the organization is using the out of box search experience which gives varied results and often leaves the user with difficulty finding the content that they are looking for. The search corpus is small with only a fraction of the organization's content being made searchable.

Users don't really trust search as they are unable to find the content that they are looking for, they find duplicate or out of date material and are not assured what they discover is authoritative. Worse, they may find the wrong content and consider it authoritative.

Some individuals are key knowledge sources, impeding their work and/or becoming 'single points of failure'.

Productivity is compromised; compliance activities are weak; organizational and colleague knowledge are poorly leveraged and there is a pervasive frustration at the inability to find things.

Level 200 - Managed

Managed level characteristics include:

200 Governance, Risk, Compliance & Security

• Some search tools respect user access rights, but inconsistencies exist, and inappropriate content may be surfaced. This is exacerbated by the power of AI based search, which is grounded in incorrect or conflicting data due to ineffective content management and governance.
• Some effort is made to promote or identify current or authoritative versions, but with limited consistency.
• Filenames are used as a substitute for metadata.
• There is no overall search governance and strategy plan.
• There is no formal process for approving or curating data sources for the semantic index.
• Users are not encouraged to use Search instead of legacy approaches.
• No role defined to administer and refine search experience.
• Risk considerations for AI use are ad hoc and reactive; often resulting in extremes like complete bans or no oversight.

200 Technology

• There are some search-based point solutions which have enhanced configuration to improve user experience.
• Some custom specific organization search results have been configured.
• AI search may be enabled in some areas, but without full semantic index configuration
• Search may be enabled within the organization, however most users bypass this and open other search engines.
• There are efforts to standardize search interfaces from system to system, however this tends to be limited to the presentation, not the format of the underlying business logic.
• Re-indexing is automated and typically occurs overnight. As such new content isn't initially findable.
• Different search syntax exists between applications.
• Limited integration of external or third‑party data sources.
• Search analytics collected but not actively used to improve AI performance.

200 User Experience

• Users have basic awareness of search, do use it for some tasks and in some systems, but rely on other methods of finding the majority of what they need. Most users are unaware of advanced search features or even the availability of search in some applications.
• Results layouts are somewhat consistent but lack refinements and high value content is not promoted to the top of the results. Result layout and features do vary between applications.
• Search tools do find content; however, this can be slow.
• Some standardization is attempted for terms, metadata, naming conventions etc. However, this is not enforced and does not apply to legacy content.
• Users frequently cannot find the content they need and fall back to other methods to confirm that they are using the correct document etc.
• ✨ Some signposting is in place, i.e. there are visual or text devices to assist the user to navigate to the correct content or ___location.
• Some users are aware of AI search but treat it as a novelty; rely mainly on keyword search.
• Minimal training on natural language queries or AI answer interpretation has been provided.
• There are inconsistent experiences across systems.

200 Impact

At this level, search usage is not ubiquitously or consistently present throughout the organization but is more popular as employees see the benefits of being able to find content. However, the search experience differs depending on where the search takes place. There may have been the migration of file content from file servers into SharePoint; it becomes possible to search across all content stored in platforms, such as Microsoft 365.

There is an increase in usage of search in general, as users find out more about the benefits of search within the organization. Users begin using search when they don't know where a specific document or item is, however, differences between search experiences confuses staff and they avoid using search in systems they are less familiar with. Colleagues remain a primary source of information or signposting to where to look. Lack of immediacy in search ensures duplicate creation remains commonplace, especially across different teams. AI search delivers occasional value but is inconsistent, with limited trust from users due to variable accuracy and incomplete coverage of relevant sources.

Productivity and compliance remain compromised; and frustration at the inability to find things persists.

Level 300 - Defined

Defined level characteristics include:

300 Governance, Risk, Compliance & Security

• ✨ User access rights are consistently applied, and processes exist to manage access.
• Basic governance policies exist for search, but AI‑driven search may not be explicitly covered.
• There is some attempt to ensure content governance in order to provide confidence in the search results, but this is not comprehensive.
• ✨ The compliance team is aware of AI summarisation risks; early review processes are in place for sensitive domains.

300 Technology

• Commonly searched keywords are configured with tailored results.
• An enterprise search exists that is connected to other file repositories and line of business applications to break down information silos and allow search across the enterprise. (This could be hybrid, Salesforce etc.). This may not be consistently available nor address all the needs of users, however.
• ✨ Search is applied consistently across services.
• ✨ 'Search verticals', which provide scopes focused on specific topics, business functions, file types and more are available, specific to the business and aim to improve precision and findability for key business functions.
• Search results are customized for key organization assets to improve findability and discovery of useful assets.
• ✨ Search is used in business applications to access large volumes of content quickly and efficiently. This gives users access to information in a way that they have not had before.
• The business is using modern search web parts to enhance the user interface and search experience.
• ✨ People are understood as information assets. Skills and expertise are captured and returned in response to search queries.
• A semantic index is configured for priority repositories; AI search is available in many core applications.
• Basic monitoring of AI search performance and adoption is available.

300 User Experience

• ✨ Users are educated on search and how to make best use of it. This includes standardised guidance on effective AI search use, including prompt examples.
• Search boxes are presented consistently and provide guidance.
• Search results are consistently laid out and provide content summaries and previews.
• Initial training is provided on interpreting AI‑generated answers and verifying against authoritative sources. There is some guidance on creating appropriate prompts, together with how to review responses for hallucinations.
• ✨ Training covers how to validate AI‑generated answers and locate source documents.
• ✨ Early adoption champions have been identified to support peers.

300 Impact

Search actively adds value to organizations, releasing staff time, improving compliance, and creating confidence that correct versions of documents, etc. are in use. Staff can locate some physical assets, skills.

At this level, Search becomes an asset to the organization. This has been recognized as an enabler that develops more efficient and effective employees. The capabilities of search are harnessed to improve the experience of businesses applications. AI search begins to measurably improve findability and productivity, though benefits are still dependent on user skill in crafting effective queries and content governance limitations erodes search confidence.

Level 400 - Predictable

Predictable level characteristics include:

400 Governance, Risk, Compliance & Security

• A role exists in the organization to manage enterprise search, to review keywords, feedback and search metrics with a view to ensuring effectiveness.
• ✨ Processes are in place to ensure staff maintain their profiles, including skills and expertise.
• ✨ The governance framework includes AI search, with defined roles for oversight and accountability. AI search governance is integrated into broader information governance, with regular audits of outputs.
• Approved content sources for grounding are documented and maintained.
• Risk registers include AI‑specific risks (e.g., hallucination, bias, sensitive data exposure) with mitigation plans.
• Compliance workflows include automated checks for regulated content in AI answers.
• Search is used to identify records and other artifacts that should be tagged.
• ✨ Centrally managed thesauri and term sets are used across search scopes that understand synonyms.
• ✨ Search is used to assist compliance processes such as subject access request and, legal eDiscovery.

400 Technology

• Search usage is analyzed and used to improve search results.
• Contextual search is embedded in line of business systems.
• Most systems and workplace tools provide consistent access to the enterprise search.
• ✨ Information is stored in such a way as to enhance findability.
• ✨ Search extends beyond files and information to locations, physical assets, relationships and more.
• Content discovery emerges as a business tool that exposes content to users who might not have known about it, by displaying information related to the search items.
• Prospective search is used to display content without the need to enter a search term, such as commonly viewed news, article documents; context drives the relevancy of this.
• Predictive search begins suggesting matches to search terms as the user enters the search query.
• Advanced queries can be created using a defined query language or natural language via AI.
• Frequency of content indexing is appropriate to the periodicity ("freshness") of change of different repositories and business processes.
• ✨ AI search is fully integrated across all relevant systems, with consistent semantic index coverage.
• Automated ingestion pipelines keep indexed content fresh and security‑trimmed.
• Search analytics is actively used to tune AI relevance, verify grounding and identify content gaps.
• Integration with related knowledge and collaboration tools provides cross‑platform knowledge surfacing.

400 User Experience

• Users are skilled at discovery of content, information, and skills across the organization.
• ✨ Search results provide previews across most content types; offer interaction (such as email, save for later), and display useful information dependent on content.
• Users can access search from most applications and elements of their digital workspace, including mobile.
• Common queries can be saved by users and notifications relating to new results are possible.
• Recommended content and common bookmarks to standard or 'best bet' results are proactively published by the organization.
• Users are frequently unaware that search is used to retrieve information within their workspace.
• ✨ External and public ___domain content is included in content activities, such as finding appropriate images.
• ✨ AI search is embedded in daily workflows; users confident in crafting effective prompts.
• ✨ Feedback mechanisms are in place for users to flag inaccurate or incomplete AI answers, and peers help with alternative prompt design.
• ✨ Personalisation and AI memory are used to improve relevance and engagement.
• ✨ AI search adoption metrics are tracked and reported to leadership.

400 Impact

This level sees Search being managed throughout the organization. Processes are in place to add new content, search verticals and search result layouts and Microsoft Search configuration.

Search is a key business information tool that enables most processes. It is widely seen as the most effective means of discovering, retrieving, and confirming business information, for identifying skills and expertise across the business and integrating knowledge from multiple systems. Most staff update profiles and participate in appropriate tagging

AI search consistently accelerates access to accurate, context‑relevant information, reducing time spent locating and validating content across the organisation. Search results can be relied on; the current versions are reliably returned; inappropriate or incorrect content is rare.

Level 500 - Optimizing

Optimizing level characteristics include:

Search is part of everyday life for an employee at the organization. New innovative ways of exposing content are investigated. Search metrics are used to analyze user behavior and understand gaps in the information that is being returned.

500 Governance, Risk, Compliance & Security

• The organization seeks to continually enhance all aspects of the search process and experience; extending scopes, optimizing results and linking related information based on continual feedback
• ✨ Advanced analytics are used to understand search usage and this provides further insights into activities across the company
• ✨ Search is actively used to identify potential threats to information governance, security and legal obligations.
• ✨ Automated tagging and other metadata are in use
• Context (staff profiles, locations etc.) is integrated with source processes such as Joiners-Leavers, in order to maintain high quality of data.
• AI search governance is subject to continuous improvement loop, using analytics and feedback to refine policies, with proactive compliance monitoring of AI outputs in high‑risk areas.
• AI risk posture reviewed regularly, with adjustments based on emerging regulations and technology changes.
• Governance extends to prompt libraries, ensuring consistent, compliant, and bias‑aware AI interactions.

500 Technology

• The search corpus is broadened with search being available across bespoke and line of business systems.
• The search corpus is used to enhance knowledge management tools.
• Opportunities to enhance search are looked for to ensure data is surfaced to improve productivity based on effective analysis.
• ✨ Effective search is ubiquitous and uniformly available across desktop, mobile and other experiences.
• External resources are included in the search scope.
• Staff profile updates are monitored and automated to ensure accuracy and completeness.
• Search is used to discover and auto-tag content.
• ✨ Users are highly skilled at finding information using tools and new staff are trained in the tools for their role.
• ✨ Automated classifiers are used to add tags to all content types, including image, audio and video, in order to ensure it is discoverable.
• SEO approaches are applied to content.
• ✨ AI search is extended to all relevant enterprise platforms, including custom applications, providing focused retrieval and integration of knowledge. This supports multi‑modal inputs (voice, image, text) and outputs (summaries, charts, action items).
• Predictive and proactive search experiences are deployed, including autonomous agents, surfacing relevant content before queries are made and identifying content issues.
• Continuous optimisation of semantic index and connectors is based on usage patterns and business priorities.

500 User Experience

• ✨ Custom Search Results are created to augment key information in the search results to support improved discovery and findability. These are monitored and a process exists for updating search scopes, presentation, filters etc. as the business needs evolve.
• ✨ Search is ubiquitous; users can access search consistently from all applications and locations within their digital workspace, including mobile and voice.
• Users can proactively provide feedback on search results, to drive improvements.
• ✨ AI is used to enhance search based on deeper knowledge of the user context and business activity.
• Search experiences are embedded in business processes and in many cases, users aren't even aware that search is supporting their work.
• Users routinely use AI search for synthesis, decision support, and proactive insight discovery.
• ✨ Prompt libraries and role‑specific templates are widely adopted to accelerate common tasks.
• AI search experience is continuously refined based on behavioural analytics and feedback.
• ✨ Culture shift: search is seen as a collaborative partner, not just a retrieval tool.

500 Impact

Search technologies are considered critical business systems, carefully managed with designed-in resilience. They are a key tool for ensuring compliance; it also underpins staff and process effectiveness.

Staff are committed to the content processes that maintain search; at the same time search is highly automated and 'invisible' delivering insights and finding knowledge without user input. Search itself provides management key insights into the health, activities, and productivity of the business.

AI search proactively surfaces insights and connections before they are requested, driving continuous knowledge discovery and informed decision‑making at scale.

Scenarios

• A project manager looking for similar projects to the one that they are just about to start and then needing to recruit an appropriately skilled team.
• A salesperson searches for similar proposals to use when creating a new proposal.
• A junior member of staff needing to find the company tax number.
• Staff searching for internal and market news relating to an insight or innovation they are considering. An engineer researching a solution to a manufacturing failure, who needs to collate procedure, machine manuals, line SPC data and actions alongside reports of similar events outside the company.
• The legal team finds contracts which will expire soon and can work on renewals where appropriate.

Cost & benefit

The following benefits can be achieved as the Search Competency increases in maturity:

• Reduced user frustration when trying to find content they know is available.
• Reduced time wastage finding information or the right person (commonly upwards of 20 minutes per person per day at level 100
• Increased innovation.
• Increased awareness of useful information and knowledge. Improvement in employee engagement.
• Increased sharing of knowledge and best practice.
• Decentralized management of content but with centralized consumption.

A modern Search experience is part of a modern digital workspace which can attract the right workforce.

Benefits are found in sharing stories, knowledge and understanding but are difficult to quantify and measure.

Resources to learn more

• Classic vs Modern Microsoft Search
• Building Custom Microsoft Search Connectors
• Search bookmarks
• Creating custom search results pages in SharePoint Online

Conclusion

Organizations that implement a successful Search strategy will see direct impacts to the bottom line. Employees being able to "discover" information which leads to innovation within the organization, reduced costs and time efficiencies can have a huge impact on worker productivity. The cost benefit in users not duplicating work and finding corporate knowledge are difficult to quantity but exist.

Search enhances the other competencies and is a great way to begin reaping rewards from the Microsoft 365 platform.

Organizations should capture success stories to provide examples of the benefits to ensure that this essential service gets the attention and resources that is required for it to be a successful resource.

Common Microsoft 365 Search technologies

• Microsoft Search (using Microsoft Graph)
  • Office search
  • Microsoft 365 Copilot
  • Copilot Studio
  • SharePoint Modern Search
  • Modern Search Web Parts
  • Search Connectors
  • Semantic Index for Copilot
  • Microsoft Graph Connectors
  • Prometheus orchestration layer
  • Azure OpenAI Service
  • Search Analytics
• SharePoint Search
• Power BI Q&A
• Managed Metadata/Term Stores
• eDiscovery and audit (Microsoft Purview)

Resources

Principal authors:

• Simon Doy, MVP
• Simon Hudson, MVP

Contributing authors:

• Marc D Anderson, MVP
• Emily Mancini, MVP, UXMC
• Sadie Van Buren

The MM4M365 core team has evolved over time and these are the people who have been a part of it.

Core team

• Marc D Anderson, MVP
• Simon Hudson, MVP
• Simon Doy, MVP
• Sharon Weaver
• Galen Keene
• Pia Langenkrans
• Mats Warnolf

Emeritus

• Sadalit (Sadie) Van Buren
• Emily Mancini, MVP, UXMC