Microsoft Teams BOT

How to use AI and BOTs in Microsoft Teams to connect employees to knowledge

BOTs introduction

This blog explores the options in using Microsoft Teams BOTs and AI technology to support employees to connect to information and knowledge resources. The most common BOTs (roBOTs) are chat BOTs that facilitate FAQs / Q&A and some user experiences are a little underwhelming. There are several reasons for this, and none relate to the technology, more on how the process and user experience has been considered, or not! Such as:

How do you ensure the Q&A knowledge base is kept up-to-date?

How do you ensure the user experience is contextual and relevant?

If the answer returned does not address the user’s question, where is the escalation point?

BOTs are best utilised as part of a wider service offering that includes a human element, ensuring the user feels they are receiving a personalised experience. For instance, if the BOT can answer a percentage of user questions, not only does this introduce immediate efficiency and productivity gains, but it then means the remaining questions provides the opportunity to deliver a personal experience when needed, and this reflects on the overall BOT experience, as well as surfacing valuable insight into enhancing the service. 

“Use of bots to handle employee questions reduces the number of IT and HR support tickets by 10% to 15%"

Challenges in connecting employees to knowledge, where BOTs can support

There are many business use-cases where a chatbot can be extended to deliver user, management and organisational value. Let’s look at a few challenges where BOTs can help provide a resolution.

How to connect employees to experts when they don’t know who the subject matter experts are.

How to connect employees to knowledge when they don’t know where the information is.

How to build a Q&A knowledge base organically – in the flow of work, because we all know that regardless of best intentions, they are not kept up-to-date.

How to harness knowledge gleaned from 1-2-1 engagements for the benefit of all, 1-2-many.

How to extract tacit knowledge into a central knowledge store to ensure it stays within the business when employees, who have the knowledge, leave.

How to alleviate repetitive information requests and make individuals, departments and processes more efficient.

How to support new / updated information needs.

How to demonstrate that your organisation is adopting modern tech.

Business use-case where BOTs can add value

Let’s translate this into business process use-cases, where a BOT can support and add value.

·         Onboarding – if you are inducting several people into the business and into specific roles, there will be a host of common questions for all new recruits 

·         Digital Transformation – the “how do i…” – i.e. on the Microsoft 365 Apps such as Teams, SharePoint etc.

 ·         Policies and Processes Understanding – the “what is the…”

 ·         Function Support – HR / H&S / Compliance / etc. How many emails / calls do these functions get per day, where the information provided could be delivered through a BOT? The functions could become experts where if the BOT cannot answer a question, by answering an escalated question through the BOT (rather than on email or on the phone) this then automatically gets added to the Q&A knowledge base – 1-2-1 engagements (which are happening anyway) benefiting 1-2-many

 ·         Product / Service Launch – supporting the communication and education

 ·         Programme / Project Insight – supporting collaboration

 ·         System Help, FAQ’s

 ·         Tacit Knowledge Extraction – extract contextual knowledge from subject matter experts into a central knowledge store, building continuity, supporting sustainability and protecting organisational IP

Including a BOT in your Managed Services proposition

If you are an organisation who delivers services that incorporates any level of knowledge transfer, BOTs can also support your service provision to clients, especially where you are delivering upskill and training as part of this. BOTs can help address several common service provision challenges…

Connect employees to expertsOnce the training programmes are completed, which in the main deliver skills into relatively small percentage of employees for an organisation, how do you connect the people who have been upskilled (the experts) to employees so they are able to provide support?

Provide a 24/7 managed serviceHow do you provide 24/7 access to your support services, FAQ’s, Help guides, data sheets etc.?

Efficient knowledge accessHow do all employees (and not just the trained ones) access the content provided through your training programmes?

Effective information updatesHow do you keep the organisation, experts and employees up-to-date on evolving subject matter?

Client retentionCommercially, how do you support your client retention model after the training programmes have been completed, where organisations have not signed-up for any additional services?

Additional and reoccurring revenue streamHow can you generate a reoccurring revenue stream from the knowledge your training programmes provide?

Company / service profileHow do you raise your profile to a wider audience that has not commissioned your services but who utilise Microsoft Teams for their day-to-day work collaboration portal?

Added valueAre you looking to add value to existing programmes and / or raise your profile in using modern tech as part of your service delivery?

New channel to marketMicrosoft Teams has 4m users from round 1m organisations, where deploying Apps into Teams, to support services in the flow of work, is a key objective – how can you connect to this audience? 

Where to start in deploying a Microsoft Teams BOT

So, how do you go about getting a Microsoft 365 BOT that works within Teams, where you can align to your service provision and / or specific employee support requirements?

Microsoft provide a number of Azure technologies that support the provision of BOTs in Microsoft 365. The core ones are:

·         QnA Maker Service

The QnA Maker Service provides the complete knowledge base authoring experience. You can import documents, in their current form, to your knowledge base. These documents (such as a FAQ, product manual, spreadsheet, or web page) are converted into question-and-answer pairs.

·         LUIS (Language Understanding and Intelligence Service)

LUIS is a conversational AI service that applies custom machine-learning intelligence to a user’s conversational, natural language text to predict overall meaning, and pull out relevant, detailed information (the intent). 

Because people ask the same question in different ways, languages and spelling, LUIS interprets the intent of a question to search content resources for an applicable answer. LUIS trains itself, so the more questions ask the more intelligent it gets.

·         Azure BOT Service

The Azure Bot Service provides an integrated environment that is purpose-built for BOT development.

There are other services that a BOT can utilise, but the principle is that Microsoft provide all the services you need to support basic template BOTs and also deeper level customised BOTs that have additional features and / or very specific use-cases. 

Microsoft 365 Teams BOTs deployment strategies

There are a number of different BOT options available to you in Microsoft Teams.

·         Teams Application Templates

There are template BOTs available to be deployed within Microsoft Teams. They tend to be base Q&A BOTs that rely on population of a Q&A knowledge base (facilitated by the QnA Maker Service in Azure). They provide a start point and can add value if you already have Q&A content or can link to provided Q&A content. 

They tend to have a single-use case and require separate deployments to facilitate multiple knowledge bases – i.e. a M365 Upskill BOT, a HR BOT, a H&S BOT, a Support BOT etc. 

The templates require enterprise level M365 licensing, and the underlying services may (please assess each template t&c’s) attract costs – mainly on a consumption basis (i.e. the more questions asked the more it costs), although relatively speaking the costs are minimal and if you get the use-case right then the cost will be insignificant compared to the value realised.

These BOTs normally require deployment by PowerShell scripts so will need to involve IT.

They are deployed within your own tenancy and there is no warranty or support (outside of the open-source forums). 

They are open-source Apps that conform to recommended best practices for security and infrastructure, and all community-submitted changes are reviewed to ensure continued conformity.

 

Click here for more information and examples on Microsoft Teams App Templates

 

·         Power Virtual Agents (PVAs)

You can build your own chat BOTs using PVAs, which is part of the Power Platform. It is SaaS (Software as a Service), offering wizard driven BOT creation with low code/no code. Intelligent Virtual Agents and Bots | Microsoft Power Virtual Agents 

PVAs licensing is consumption/session based – i.e. cost per engagement with the BOT (Pricing Plans | Microsoft Power Virtual Agents) 

Building BOTs using PVAs supports specific and scalable / complex use-cases, but still uses the underlying Azure BOT framework services – for an example case study see: Microsoft Customer Story-PwC simplifies data retrieval with intelligent bots built on Power Virtual Agents 

Microsoft provide a toolkit to start to build PVA BOTs if you wish to develop inhouse – GitHub – OfficeDev/Microsoft-teams-apps-adopt-bot

 

·         3rd Party Apps – available to download into Teams from the Microsoft Marketplace

Vendors can develop BOT Apps that are published in the Microsoft Teams App store, which can be downloaded directly into Teams. The process to get an App published in the store requires validation and verification that the BOT adheres to prescribed standards and utilises AAD (Azure Active Directory) authentication. This is to provide organisations with confidence that Apps have been constructed correctly and are supported by a Microsoft Partner. 

Some have very specific use-cases (i.e. a BOT that links to a specific ERP / MRP / HR system), where the vendor wishes to provide access to the wider application via a Teams user interface. Others have generic use-cases with pre-built additional functionality (than the out-of-the-box templates) and come with warranty and development roadmaps to ensure they stay relevant and up-to-date. 

Vendors who publish Apps in the marketplace are encouraged to offer free trials.

tiLly {today I Learned} is an example of this kind of BOT
It Connects users to Q&A, Content Search and Experts across multiple services from a single BOT

Download Here:

·        Custom Apps

Custom Apps are similar to 3rd Party but they have not been approved for publishing within the marketplace. This does not mean that they are of a lesser quality or do not have a valid use-case, just that they are not available to download from the marketplace and must be deployed / uploaded manually into Teams. 

Obviously, because they have not been through the Microsoft approval process extra diligence is required to ensure the BOT is of the required quality and assurance policies are in place.

Considerations when deciding which BOT solution to utilise

The following may provide some direction when reviewing which Microsoft Teams AI and BOTs technology to deploy in order to support your employee connection to information and knowledge resources.

Cost

All BOTs have an underlying cost that consists of Azure Services consumption and / or M365 license costs. Ensure you understand the scalable cost of utilising BOT technologies to create a sound business and ROI case.

Complexity

Template BOT Apps remove the complexity from a creation and deployment perspective, but because they are Open Source and deployed within your own tenancy, you still need to understand the underlying services utilised to assess cost, scalability and support requirements.

PVA and Custom created BOTs are complex by nature. Although Microsoft are positioning that BOTs can be created without code, and they can, you still need a technical aptitude to build and understand the underlying architecture – for the same reasons as the Template BOT Apps.

3rd Party Apps are the simplest, as they are created, provisioned and supported by a Microsoft Partner. They tend to be packaged to remove complexities – both from a technical and cost perspective.

Use-case

Understand your clear use-case and what value (financial or otherwise) you wish to achieve and how you will measure this.

Some BOTs have single use-cases and others can provide a wider offering, so consider both the business and user angles on this. Part of the value of BOTs is to provide users with a connection point (to knowledge, experts, services etc.) and you don’t want to complicate this by asking them to work out which BOT they should engage with.

From the business-case perspective, where are the quick wins to be had in utilising BOTs – where’s the best fit to deliver efficiencies, productivity gains and provide enhanced service levels.

Adoption and Change Management (ACM)

ACM is key, whether you are including a BOT as part of your service delivery to clients or looking to deploy within your organisation. At an organisational level it may be obvious to engage the BOT to ask questions and connect to experts, as this scales 1-2-1 engagements for the benefit of 1-2-many and organically builds central knowledge stores, but to an individual they just see their need and may still feel that calling or emailing the subject matter expert (if they know who they are) is easier and feels more personal.

Content

Consideration of what content you wish to connect to – is it internal and / or external content; what is the structure / quality of the content; does it have the right level of governance and security; does it have the right level of taxonomy; do you have the structured update / review processes in place, etc.

Content management is a whole separate subject, but if you are incorporating search (in addition to standard Q&A) whilst BOTs can interpret question intent and link to applicable content, it will only surface what content you commissioned it to work with and it will only be able to find the right content that your taxonomy and security policies allow.

Other considerations on content are whether you wish to access multiple Q&A resources and multiple asset search resources. Some BOTs only have single use offerings, so careful consideration of use-cases and scalability are required at the outset.

Managed Services

Some BOTs, such as the tiLly {today I Learned} BOT, have the capability to connect to multiple resources, both internal and managed service. The BOT ranks the responses from all resources and presents the most relevant to the user. As such, consideration should be given to potential managed service requirements where you have a specific or specialised knowledge service need that dictates external commissioning – either to provide this specialism and / or provide a continual update feed.

A relevant example of this is Microsoft 365 itself, the amount of content is both broad and deep and it changes on a monthly basis. If you are providing a M365 Q&A and learning asset BOT to support the upskilling in the use of M365 Apps (such as Teams, SharePoint, Planner, Power BI, Viva etc.), you need to commit to keeping the content up-to-date. There are managed service providers that do this for M365.

Analytics

Ensure your chosen BOT delivers analytics in order to provide the required level of insight to measure the impact. Consider usage analytics – number of BOT engagements, expert support analytics – number of escalated questions, most prominent experts, highest rated experts, time to respond, and L&D analytics – common subject matter engagements.

Using services such as Azure Insights will allow you to capture insight from BOT engagements and Power BI will facilitate the creation of dashboards and reports to slice the data by organisational structure (as defined in your AAD) – i.e. company, country, region, department etc.

Support

With the template-based BOTs you will need to rely on the open-source community for any support. With internally built BOTs, any support will come from your internal build team and with 3rd party this in provided by the provider (check the t&c’s).

A roadmap for enhancements is another consideration, so you can ensure that you either have the internal vision and resources to keep the BOT up-to-date as technology and use-cases evolve, or that the providers roadmap is aligned to your own existing and future needs.

 

Microsoft Teams BOTs

Share :

Twitter
Facebook
LinkedIn

Newsletter

Subscribe to our fortnightly newsletter with stories from our latest adventures and the best travel tips

tiLly

More News

contact us

We only store the collected data for as long as is necessary to process your request and provide any requested services and / or requirements. For further information please refer to our privacy notice.