It’s been just over a week since Commsverse and I’ve finally found a bit of time to get my reflections and thoughts on the event.
Firstly, Commsverse remains my favourite event of the year. I can’t really put a finger on one single reason, it’s a combination of multiple, small compounding elements:
- The scope: It’s sufficiently broad to cover enough technology areas that brings an engaging audience, without diluting the core focus of Teams.
- Then audience: The people that join Commsverse are incredible, you’re never far from a high-quality conversation.
- The speakers: Commsverse always draws fantastic speakers. It’s just a pain when two speakers you really want to hear are presenting at the same time!
- The venue: Sometimes you need a break from the tech, you can always go and look at some fantastic cars to get away from it for 10 minutes. Much more welcoming than a sterile exhibition centre.
- The gingerbread: Every year I’ve been, the welcome bag has had a portion of Sarah Nelson’s Grasmere Gingerbread. I didn’t know gingerbread could be so good.
Who should go to Commsverse
There is something for everyone at Commsverse, business decision makers, IT administrators, IT architects, security architects, the list goes on. The structure this year was helpful too, with business, decision making sessions on day one and more technical sessions on day two.
There is always a lot of food for thought from Commsverse, you get validation on the things you are doing well, as partner this can be confirmation that certain go-to-markets are relevant, you get inspiration for adopting new and more efficient services, and you get challenged on new and developed capabilities and services.
The Highlights:
I’ve compiled a summary of each session I went to, everyone had something to take away, whether it was something we’re already doing or offering to our clients and being reinforced, something new or upcoming that we can offer our clients, or some food for thought for future go to market campaigns.
Audio Codes Keynote
Session Summary:
The keynote explored how AI has matured from disruptive hype to a pragmatic enabler of productivity in contact centres and meetings, with a focus on seamless integration into existing Microsoft Teams environments rather than wholesale replacement of human roles.
Key Highlights:
- AI as an Enhancer, not a Replacement:
The narrative has shifted from “AI will take your job” to AI being a complementary tool, as reflected in Gartner’s “Trough of Disillusionment”. AudioCodes reports measurable gains without headcount reduction. - Strategic Integration with Teams:
AI is now embedded in familiar tools such as Microsoft Teams, through Copilot, Teams Premium and third party platforms like AudioCodes Voca, enabling actionable meeting insights, timestamped transcripts, and intelligent call centre agent assistance through solutions. - Real-World Impact with Voice AI:
From Rabobank to UK supermarkets, businesses are leveraging AI transcription and automation frameworks to drive efficiencies, improve agent performance, and reduce customer service friction without compromising existing infrastructure.
A lot of people will have heard about Copilot and the benefits it can bring to Teams, however, this may not be a practical investment to all use cases. Some users will have very specific roles, where the Return on Investment (ROI) at \$30 per user per month might not be realised. Where this is the case, there are alternatives that may show a stronger ROI, enabling users to benefit from AI, regardless of their roll.
Enabling Frontline Workers
Session Summary:
This session examined how ruggedised, enterprise-grade devices and intelligent gateways are transforming frontline communications by bridging legacy infrastructure with Microsoft Teams, enabling operational efficiency and improved safety across healthcare, warehousing, and manufacturing sectors.
Key Highlights:
- Frontline Worker Enablement Through Specialised Devices:
Spectralink’s IP68-rated S-Series handsets (S33, S35, S37) are purpose-built for demanding environments, offering features such as motion-triggered alarms, hot-swappable batteries, and seamless Microsoft Teams integration—particularly valuable in safety-critical industries. - Bridging Legacy and Cloud with SIP Gateways:
A proprietary IP gateway enables organisations to retire legacy PBX and DECT systems, connecting existing infrastructure into Teams and yielding significant cost savings—Spectralink leads this space with up to 30% IP DECT market share. - Real-World Impact and Strategic Integrations:
Case studies from healthcare and warehousing demonstrate measurable improvements in safety response and communication reliability, underpinned by strategic partnerships with Microsoft and medical technology vendors to deliver a unified device experience.
Frontline workers can often be overlooked, yet over 80% of the workforce is a frontline worker. There are solutions out there that can help bridge the gap between the frontline and knowledge workers. These solutions coupled with Microsoft 365 Frontline SKUs can significantly improve engagement and communication with your frontline.
Sustainable Copilot Adoption
Session Summary:
This session provided a practical blueprint for sustainable Microsoft 365 Copilot adoption, highlighting that long-term success depends not on the technology alone, but on structured collaboration, continuous data capture, and measurable process improvement.
Key Highlights:
- From Conversations to Context: Structured Collaboration is Foundational
Copilot thrives when teams transition from fragmented chats and emails to structured Teams channels, ensuring it can access meaningful context and contribute value. “Copilot does not know more than the user.” - Idea to Insight Workflow with Microsoft 365 Tools
A streamlined process, from voice memos and transcriptions to OneNote, Word, and PowerPoint, enables rapid content generation, reducing cognitive load and promoting regular collaboration, especially when enhanced with Copilot templates. - Data-Driven Adoption with Viva Insights and ADKAR
Real-time dashboards and change models like ADKAR ensure Copilot’s impact is measurable, helping teams track focus time, meeting reductions, and overall digital well-being while embedding adoption as a perpetual process rather than a one-off project.
The ADKAR (Awareness, Desire, Knowledge, Ability, and Reinforcement) model is an excellent way of thinking about Copilot adoption. Copilot adoption is not a single project that lasts for a couple of months, and it’s done. It will be a continuous evolving engagement that lasts for as long as Copilot!
MVP Panel Discussion
Session Summary:
This expert panel unpacked the interplay between Microsoft 365 Copilot, AI tools, and data governance, highlighting the need for strong compliance controls, structured adoption strategies, and business-aligned leadership to harness AI’s full potential securely and effectively. This session was a really enlightening, mainly getting the assurance that the challenges we experience internally and with our clients is experienced by pretty much everyone else out there too!
Key Highlights:
- Data Governance as a Prerequisite for AI Success:
With “75% of input being rubbish,” the panel stressed cleaning up unstructured data and applying tools like Purview, Defender for Cloud Apps, and SharePoint Advanced Management to enforce sensible access, prevent data leakage, and ensure Copilot works within governed boundaries. - Adoption Must Be Led by Business, Not Just IT:
True value emerges when organisations treat Copilot adoption as part of business transformation—not merely IT enablement—supported by champions, measurable KPIs via Viva Insights, and a full-time role just below senior leadership to sustain digital change. - Balancing Practical Use with Strategic Innovation:
Focus should begin with simple, high-value use cases—such as diary automation or meeting insights—while piloting Copilot capabilities in sandboxed scenarios to build confidence, validate ROI, and reengineer processes rather than layering AI onto inefficient workflows.
21 Questions to ask your integrated telephony provider
Session Summary:
This session highlighted Gamma’s approach to Microsoft Teams Voice enablement, emphasising AI integration, hybrid deployment strategies, and business continuity, all underpinned by resilient infrastructure, strong Microsoft alignment, and a single-provider service model. Although this was a vendor sponsored session, the principles apply to every voice provider.
Key Highlights:
- Resilient Teams Voice with Integrated AI and Continuity Tools:
Gamma delivers Teams Voice via Operator Connect with 100% uptime since 2021, combining AI enhancements like Co-pilot call summaries with tools such as Citron Call Manager for seamless failover and business continuity planning. - Strategic Deployment and Legacy Transition Planning:
Emphasis was placed on mapping end-user cases before deployment, assessing legacy device requirements, and supporting clients through porting, migration, and skill uplift—tailored to varying in-house capabilities.
Compliance Call Recording
This session positioned recording technology not merely as a compliance necessity but as foundational infrastructure for AI readiness, highlighting how structured data ingestion enables advanced analytics, legal defensibility, and operational excellence.
Key Highlights:
- Recording as Strategic Infrastructure for AI and Compliance:
Capturing every interaction—calls, meetings, and transcripts—is critical not just for regulatory frameworks like MiFID and GDPR, but also for training AI models and surfacing insights, likened aptly to an “oil rig” mining the new oil: data. - Security-First, Customer-Owned Architecture:
The platform enables organisations to retain control over their data via bring-your-own-storage models, role-based access, full audit logging, and support for private endpoints—ensuring sovereignty, compliance, and operational trust.
We often work with clients who don’t believe call recording is important, or they believe the native recording capabilities in Teams (convenience call recording) is sufficient, when it’s not. There are lots of options available to meet your compliance and training needs.
Understanding Copilot and Oversharing Risks
Session Summary:
This strategic session explored how Copilot amplifies existing data governance flaws, presenting both risk and opportunity, and offered a prescriptive, tool-backed approach using SharePoint Advanced Management and Microsoft Purview to mitigate oversharing and secure AI enablement.
Key Highlights:
- Copilot Surfaces What You Permit—Not What You Intended:
Copilot respects user permissions but exposes latent oversharing risks in platforms like SharePoint and OneDrive; issues such as public sites, broken inheritance, and lack of sensitivity labelling must be addressed to avoid accidental data exposure. - Phased Mitigation via SAM and Purview Tools:
A structured “Oversharing Blueprint” recommends beginning with a pilot ring-fence, then scaling governance using tools like SAM Permission State Reports, Purview DSPM for AI, DLP policies, and Restricted SharePoint Search, enabling secure Copilot adoption without paralysing innovation. - Sensitivity Labels as the Cornerstone of Safe AI Interaction:
Purview labels—when applied to both files and containers—define what data Copilot can summarise, block, or ignore, with DLP for Copilot enabling fine-grained control over summarisation of highly confidential content, ensuring protection scales with adoption.
It’s important to remember that Microsoft 365 Copilot has access to what the user has, no more, and no less. Note. The user might not be aware they have access to certain data, and users might not be aware of how widely they’ve shared their data!
Surfacing external data in Copilot
Session Summary:
This session demonstrated how Copilot Connectors extend Microsoft 365 Copilot’s reach by securely integrating external data sources such as Azure DevOps and Salesforce, enabling semantic search and agent-driven intelligence within a unified Microsoft interface.
Key Highlights:
- Copilot Connectors Unlock External Insights in Microsoft 365:
Over 100 pre-built or custom connectors can ingest content from platforms like ServiceNow, GitHub, and Zoom, allowing Copilot to return richer, citation-based responses using semantic indexing across internal and external data. - Secure by Design with Granular ACL Mapping and Encryption:
External data respects original access controls, supports tenant-region residency, and integrates with Microsoft Purview policies, with optional staging mode, modern authentication, and daily incremental crawls for sensitive, high-trust deployments. - Agent Builder Brings Targeted Intelligence to Users:
Lightweight agents, such as a DevOps task retriever, can be rapidly deployed through a guided interface—surfacing contextual answers directly within Word, Excel, and Teams—offering a practical gateway to Copilot Studio’s broader potential.
Not everyone’s data resides in the Microsoft eco-system, and this isn’t a limit for Copilot. Third party platforms, custom line of business apps and on-premises data can all be surfaced in Copilot.
Copilot analytics, With Viva Insights
Session Summary:
This advanced workshop explored how Copilot analytics, deeply integrated with Viva Insights, can be used to measure adoption, identify team-level trends, and ensure data governance—provided organisations implement robust data upload processes and assign the correct administrative roles.
Key Highlights:
- Copilot Insights Begin with Proper Licensing and Role Assignment:
Full-featured dashboards require at least 50 licensed users and proper delegation of roles such as Global Administrator and Insights Analyst; without these, granular reporting (e.g. departmental trends) and privacy features like differential masking may be unavailable. - Organisational Data Drives Targeted Analytics, But Must Be Handled Carefully:
Uploading custom attributes via Viva Insights (not the Admin Centre) ensures powerful filtering without privacy breaches—misconfigured CSVs and incorrect encoding are common stumbling blocks that must be avoided. - Privacy and Analytics Go Hand in Hand:
With safeguards like minimum group size thresholds and distribution masking, the platform ensures data remains anonymous while allowing actionable insights through Power BI templates, skills tracking, and usage heatmaps that can inform change management strategy.
Making sure you have a positive ROI in your Copilot investment is crucial. At \$30 per user per month, you need to ensure you see improvements in productivity and processes improve. The tools are available for you to measure this.
Landis Day Two Keynote
Session Summary:
The day two keynote outlined Landis’s strategic position as a Microsoft Teams-native contact centre vendor, spotlighting its dual-engine architecture and AI-powered capabilities, all built to enhance agent efficiency, simplify operations, and align with Microsoft’s evolving Unify model for contact centres.
Key Highlights:
- Dual-Engine Flexibility with Extend and Unify Models:
Landis enables organisations to adopt both the mature Extend model and Microsoft’s forward-looking Unify integration, allowing hybrid queue handling, modularity, and future-proof deployment based on operational readiness. - AI-Enhanced Agent and Supervisor Experience:
From natural language IVRs and real-time coaching to automatic call summarisation and CRM updates, AI augments every stage of the contact flow, reducing manual effort and enabling richer insights for supervisors. - Clear Differentiation from QSAP and Dynamics:
Unlike Microsoft’s own offerings—limited in functionality or heavy in implementation overhead, Landis provides rapid time to value, full Teams-native integration, and enterprise-grade features including skills-based routing, WFM, QM, and comprehensive reporting.
Using Copilot with SBC Logs
Session Summary:
This session showcased a practical demo of using Microsoft Copilot, Azure Event Hub, and AI-powered assistants to automate log analysis, reduce noise, and deliver context-rich, human-readable alerts to support teams via Microsoft Teams.
Key Highlights:
- Automated Log Ingestion and Filtering Pipeline:
Syslog events from an SVC gateway are streamed into Azure Event Hub, filtered through Stream Analytics, and written to SQL, with Azure Functions retrieving and transforming logs into alert summaries using templated prompts to avoid LLM inconsistency. - Intelligent Notifications and Duplicate Prevention:
Alerts, enriched with group-level context and potential fixes, are sent to Teams via webhook; a text comparison function avoids alert spam by detecting duplicate events, optimising signal-to-noise ratio for first-line responders. - Copilot Agent Integration with Internal Documentation:
A 365 Agent pulls from a restricted knowledge base indexed in Azure AI Search (e.g., AudioCodes manuals), surfacing relevant answers and resolutions alongside alerts—empowering junior staff and enabling iterative improvement of support workflows.
This was an excellent demonstration of how logs from Session Border Controllers (SBC) can be integrated with generative AI agents to provide real time, human readable information to first line support engineers. Although this focused on SBCs, this can be concept can be applied to a lot of use cases.
Copilot Daily Workflow
Session Summary:
This demo showcased how Microsoft Copilot streamlines daily workflows—from email and document drafting to data analytics and custom agent creation—by integrating deeply across Microsoft 365, enabling efficient, personalised, and collaborative user experiences.
Key Highlights:
- Personalised Email and Workflow Automation:
Copilot adapts tone by recipient type, analyses recent communications to refine reply style, and surfaces missed emails, meetings, and tasks—allowing users to begin the day with structured context and clear priorities. - From Brainstorm to Boardroom in One Loop:
Users can voice-capture ideas into Word, structure content via Copilot, validate citations, and seamlessly export to PowerPoint—transforming a “brain splurge” into a complete white paper and deck within one interface. - Data Analytics and Agent-Driven Reporting:
Analyst agents process internal and external data (e.g., sales, Kaggle), generate visuals and trend reports, and custom agents automate repetitive reporting tasks in preferred formats—bridging insight generation with stakeholder-ready presentation.
Embedding Copilot into your everyday work is important to realising the value. This was an excellent refresher for me on how you can build it into every part of your day to make yourself more productive, informed, saving time and enabling creativity.
Optimising Network Performance for Microsoft Teams Audio/Video
Session Summary:
This highly detailed session examined Microsoft Teams network performance from both infrastructure and endpoint perspectives, stressing the need for UDP enablement, DNS optimisation, and cross-team collaboration to deliver consistent audio and video quality across hybrid working environments.
Key Highlights:
- UDP and DNS Are Critical for Teams Media:
Blocking UDP or routing DNS queries through centralised infrastructure often leads to degraded performance; split-tunnel VPNs, local DNS breakout, and firewall policies that reject (not drop) traffic are essential for optimal latency and call quality. - Teams Media Is Not Just About Bandwidth:
Real-time communication hinges on jitter, packet loss, and latency—not throughput; traditional network tools (e.g., SNMP) are insufficient, requiring endpoint telemetry, TURN tests, and Teams-specific analytics for accurate diagnostics. - Cross-functional Alignment Drives Success:
Bridging the operational gap between “networking” and “Windows” teams is vital; shared understanding of real-time media protocols (e.g., Interactive Connectivity Establishment (ICE)/Session Traversal Utilities (STUN)/Traversal Using Relays around NAT (TURN)), proxy behaviour, and Microsoft routing logic enables proactive troubleshooting and resilient deployment strategies.
Most issues that arise with a Teams call or meeting are nothing to do with Microsoft, it’s down to either your network, there’s a lot that impact the performance of your Teams call, from DNS, to VPNs, or using uncertified peripherals, like unsupported headsets.
Risk management for generative AI
Session Summary:
This session explored the risk landscape of generative AI tools in the enterprise, illustrating how Microsoft security and compliance platforms—namely Defender for Cloud Apps and Purview Data Security Management for AI—can help organisations detect, assess, and govern both sanctioned and shadow AI usage.
Key Highlights:
- Shadow AI is the New Shadow IT:
With 78% of employees bringing their own AI tools to work, visibility is essential; Defender for Cloud Apps surfaces unsanctioned Gen AI usage, enabling policy enforcement, app risk scoring, and synchronised blocking with Defender for Endpoint. - Purview Data Security Management for AI (DS AI) Enables Granular Oversight and Data Loss Prevention (DLP) Enforcement:
When integrated with managed endpoints and browser extensions, Purview captures user prompts, data flows, and sensitive content exposure across 450+ Gen AI tools—powering insider risk detection, auto-tagging of new tools, and policy refinement. - Balance Productivity and Protection through Conditional Access and Labelling:
Conditional access, custom sensitive info types, and audit trails support flexible governance strategies—allowing organisations to protect intellectual property while building the case for responsible AI enablement aligned to business risk.
Purview and Defender for Cloud Apps has come on leaps and bounds over the last couple of years. The latest features provide a treasure trove of capabilities to identify the use of unsanctioned AI tools, and what data is being shared with them. This can be used to both harden safe guards in your environment, or to support the business case to equip your teams with sanctioned AI tools.