Search this site

Speakers Imen Harrazi Business Development Rep Calvin Ayim Ofori Senior Atlassian Consultant

Regardless of the industry or size, we offer software that assists businesses in achieving their goals, setting new heights for growth in their sector. Case Studies Regardless of the industry or size, we offer software that assists businesses in achieving their goals, setting new heights for growth in their sector. Banking Industry An interview with a financial analyst about the benefits and uses of Jira within the banking industry. RAICO A case study about the implementation of Microsoft Business Central, a Cloud Accounting / ERP software solution. Bluewindow A case study about the implementation of Continia a document management tool that streamlines and automates the tedious process of documentation in Business Central. MTN An interview about the benefits and use of Jira within MTN in Ghana.

General Data Processing Addendum (DPA) READ MORE Atlassian End User License Agreement (EULA) READ MORE General Privacy Policy READ MORE

Most "what's new in Jira" roundups this cycle will lead with the same three words: Rovo, Claude, automation. Fine, but most of them stop at "AI can now help with X" without saying what that actually means for the work sitting in your queue today. This is the list that does, verified against Atlassian's own release notes, not the marketing copy around them. 1. Teams conversations become Jira actions (Rovo) Rovo can now read a Microsoft Teams conversation and convert what was decided into a Jira action, without anyone re-typing it. Shipped August 26. Why it matters for teams running Teams + Jira side by side: the decision and the ticket used to be two separate events, one spoken, one typed, usually by two different people at two different times. That gap is where context gets lost. Closing it isn't a convenience feature, it's the removal of a manual translation step your team has been doing by hand for years. 2. Code Context for Rovo and AI coding agents (beta) Rovo and AI coding agents can now reason across multi-repo codebases through Atlassian's Teamwork Graph, instead of being limited to whatever file is open. Shipped August 18, still in beta. Why it matters for engineering teams specifically: an AI suggestion is only as good as what it can see. A tool that reasons across one file gives file-level answers. A tool that reasons across your whole codebase gives system-level ones, which is the difference between a suggestion you have to double-check and one you don't. 3. Coding agents run natively inside Jira Automation, including Claude Claude, GitHub Copilot, and Cursor are now native action steps inside Jira Automation rules, sitting alongside every other action Jira already supports. Shipped July 31. Why it matters more than the previous two: this is the point where AI stops being something you consult and becomes something your rules trigger. Previously, using Claude inside a Jira workflow meant a separate integration or a manual handoff between systems. Now it's a step in the rule builder. That's a structural change, not a feature announcement. 4. Individual capacity planning Person-level capacity visibility is now built into Jira's planning views, not just team-aggregate velocity. Shipped July 27. Why it matters for resourcing decisions: most teams are still running capacity planning off a shared spreadsheet that's a week stale and that nobody fully trusts. A native, person-level view doesn't just save time, it removes the argument about whose numbers are right. 5. Calculated fields, natively, in Jira Budget variance, ROI, risk scores, days to due date, effort remaining, can now be calculated directly on the work item instead of exported and rebuilt elsewhere. Shipped July 27. Why it matters for reporting-heavy teams: every export is a place where a formula gets copied wrong, a filter gets dropped, or the data goes stale between the export and the meeting it's used in. Calculating in place removes that failure point entirely, and it's a bigger deal for audit and compliance-adjacent work than it looks on the release page. 6. Loom meetings auto-suggest Jira updates (Rovo) Connect a Loom recording to a Jira space, and Rovo turns the decisions and action items discussed into suggested work item updates, reviewed and applied in a few clicks. Shipped July 27. Why it matters alongside update #1: together, these two releases point at the same direction of travel. Rovo is increasingly treating conversation, wherever it happens, as a direct input into Jira. The person's job shifts from "translate the meeting into tickets" to "review what Rovo already drafted." What this list deliberately leaves out No UI polish updates, no faster create-form, no attachments redesign. Jira shipped those too this cycle, and they're worth a separate note. This list is specifically the releases that change where AI sits in the workflow, from something you ask, to something that acts. That distinction is the one most "what's new" roundups blur past. Where this fits with the rest of your Atlassian setup Updates #1, #2, #3, and #6 are only as useful as your automation rules and permission structure allow them to be. An AI action step configured against a messy rule set just automates the mess faster. If your team is evaluating any of these six, that's usually also the right moment to check whether your current Jira Automation setup is actually clean enough to hand to Claude or Rovo. Trying to work out which of this cycle's releases are worth prioritizing for your instance? Get in touch.

74% of leaders say ineffective communication is slowing their teams down. Not lack of talent. Not lack of tools. Communication breakdowns. The irony? Most organizations already have plenty of software. Jira for tracking. Confluence for documentation. Slack for chat. Loom for video. The problem is that these tools don't talk to each other. Information gets trapped in silos. Context gets lost between apps. And people spend more time hunting for answers than actually doing the work. That's exactly the problem Atlassian's Teamwork Collection was built to solve. Launched at Team '25 and now fully available for Premium and Enterprise customers, it's more than a bundle of apps. It's a unified system of work that connects Jira, Confluence, Loom, and Rovo AI agents through a shared data layer called the Teamwork Graph. By the end of this guide, you'll know exactly what's included in the Teamwork Collection, how to unlock it for your organization, what it costs, and whether it's the right fit for your team. We'll also cover what's new in 2026 and how to avoid common implementation pitfalls. As an Atlassian Gold Solution Partner, we've guided organizations across Europe and Africa through this exact journey. Let's break down what you need to know. Breaking Down the Core Components Let's look at each component in the Teamwork Collection and how they work together. Jira: The System of Record for Work Jira has evolved far beyond its software development roots. With the consolidation of Jira Work Management, it now serves as a unified project management platform for every team in your organization. Business teams get access to templates for campaign management, HR onboarding, financial planning, and more. Technical teams keep their agile boards, sprints, and DevOps integrations. Everyone works in the same system, which means leadership gets unified visibility across all workstreams without chasing down updates from different departments. The integration with Confluence means documentation lives alongside the work it describes. Link a spec doc to an epic, and anyone viewing the epic can see the latest version without switching apps. Confluence: Your Team's Knowledge Hub Confluence acts as the central nervous system for team knowledge. Real-time collaborative editing means multiple people can work on the same document simultaneously without version control nightmares. The AI-powered writing assistance helps teams draft faster, summarize long documents, and maintain consistent tone across content. But the real power comes from how Confluence connects to the rest of the ecosystem. Jira issues can be embedded directly into pages and stay synced automatically. Loom videos can be embedded with transcripts and action items extracted by AI. For organizations struggling with "where did we document that?" Confluence becomes the single source of truth that actually sticks. Loom: Async Communication Made Simple Loom brings asynchronous video messaging into the workflow. Instead of scheduling yet another meeting for a quick update, record a 2-minute video and share it with a link. The AI features are what make it practical at scale. Loom automatically generates summaries, extracts action items, and creates transcripts. The Meeting Assistant can join live calls, capture everything discussed, and publish a summary directly to Confluence with linked resources. This directly addresses one of the biggest productivity drains in modern work: meeting overload. One in ten workers loses a full day each week to meetings. Loom helps teams communicate with the richness of video without the scheduling overhead. Rovo: AI That Knows Your Business Rovo is where the Teamwork Collection gets genuinely interesting. It's an AI platform with three core capabilities: Rovo Search connects to over 50 third-party applications and delivers personalized results based on your role, projects, and work history. Instead of wading through irrelevant documents, you get curated answers that connect you to the right people and information. Rovo Chat provides conversational AI that maintains context across dialogue sessions. Ask a follow-up question, and it remembers what you were discussing. The Deep Research mode can gather comprehensive information and generate first drafts of requirements documents or project summaries. Rovo Agents are the headline feature. These are autonomous AI teammates that can take action in your Atlassian tools. They can create Jira issues, update Confluence pages, generate diagrams, and execute multi-step workflows. Crucially, they respect your permission settings (they can only access what you can access). Out-of-the-box agents include: Customer Insights Agent: Transforms raw feedback into actionable insights using a structured approach Brand Voice Crafter: Ensures content aligns with your brand guidelines Brainstorm Facilitator: Generates idea cards in Confluence whiteboards (coming soon) Diagram Creator: Turns discussions into visual diagrams (coming soon) Workflow Builder: Creates and optimizes Jira workflows using natural language (coming soon) Meeting Insights Reporter: Summarizes decisions from multiple Loom recordings (coming soon) You can also build custom agents using Rovo Studio, a no-code toolkit that lets you define agent behavior with natural language instructions. The Teamwork Graph: Why It Matters We've mentioned the Teamwork Graph several times. Let's clarify why it's so important. Most AI tools suffer from the same limitation: they don't understand your business. They can process text, generate content, and answer questions, but they lack context about how your organization actually works. A generic AI assistant doesn't know that Sarah in Marketing is working on the Q2 campaign, or that the "Phoenix Project" refers to your infrastructure migration, not the mythical bird. The Teamwork Graph solves this by creating a unified data model of your organization. It maps relationships between: People and teams Projects and goals Pages, documents, and decisions Work items and their dependencies This graph structure (based on graph theory mathematics) means the system understands connections, not just content. When you ask Rovo about "the Phoenix Project," it knows which Jira issues, Confluence pages, and Loom videos are relevant to you specifically. As Matt Chester, Senior System Engineer at Reddit, put it: "By bringing together all of this scattered information, people will know where to find what they need. A lot of times they don't even have to switch tools to find it." The practical impact is less time hunting, more time doing. Context appears where and when you need it, not buried three clicks deep in a different app. Teamwork Collection Pricing and Plans Let's talk numbers. The Teamwork Collection is available in three tiers: Plan Price What's Included Best For Standard Coming 2026 Jira, Confluence, Loom, Rovo (limited features) Small teams starting out Premium $21.50/user/month ($215/user/year) Everything in Standard + advanced features, 24/7 support, analytics Growing teams Enterprise Custom pricing Everything in Premium + Atlassian Guard Standard, unlimited instances, enterprise security Large organizations Source: Atlassian Teamwork Collection Pricing Key pricing considerations: Premium and Enterprise are available now. Standard is expected to launch sometime in 2026, though Atlassian hasn't announced a specific date. The collection offers up to 40% savings compared to purchasing Jira, Confluence, and Loom separately. License pooling across tools prevents the common problem of overlapping seats (where the same person needs a license for multiple products). Atlassian Guard Standard, which includes SSO and SCIM provisioning, is included at no additional cost with Enterprise plans. When does it save money vs. cost more? If you need Enterprise Jira, Enterprise Confluence, and Loom Enterprise, the Teamwork Collection Enterprise will save you approximately 25% compared to buying separately. However, if you need Enterprise Jira and Confluence but don't use Loom, you might pay 18-20% more than buying just those two products individually. The value proposition depends on which tools your organization actually uses. Source: Atlassian Licensing FAQ What's New in 2026 The Teamwork Collection continues to evolve. Here's what's new or coming soon: January 2026 Feature Focus: Standups in Jira: A new way to run daily standups directly from your Jira Software board Jira Seasonal Releases: Atlassian is moving to a predictable seasonal release cycle for major features Rovo Skills for Admins (EAP): Early access program for Rovo capabilities that handle repetitive admin tasks Enhanced Automation: New Branch with Conditions and Add Web Link actions for Atlassian Automation How to Unlock Teamwork Collection Ready to get started? Here's the practical path to activation. Prerequisites: First, you must be on Atlassian Cloud. The Teamwork Collection is not available for Data Center deployments. If you're still on Data Center, you'll need to migrate to Cloud first. Second, you'll need Premium or Enterprise plans. Standard isn't available yet, so there's no entry-level tier for small teams at the moment. Third, you'll need organization admin access to make licensing changes. Step-by-Step Activation: Assess your current state. Audit your existing Atlassian licenses and identify what you're already paying for. Calculate whether the bundled pricing saves you money. Choose your tier. Premium works for most growing teams. Enterprise makes sense if you need Atlassian Guard, unlimited instances, or have complex security requirements. Contact Atlassian or a Solution Partner. You can purchase directly through Atlassian, but partners often provide additional implementation support and can help optimize your licensing structure. Plan your rollout. Don't flip the switch for everyone on day one. Start with a pilot team, establish best practices, then expand. Configure Rovo agents. The out-of-the-box agents need some setup to work with your specific workflows. Plan time for configuration and testing. Train your teams. The tools are only valuable if people actually use them. Invest in training, especially for the AI features that require some learning to use effectively. Common Pitfalls to Avoid: Assuming it's just a license change. The Teamwork Collection represents a different way of working. Without change management, you'll have new tools but old habits. Underutilizing Rovo. Many organizations enable the collection but never configure the AI agents. That's like buying a smartphone and only using it for calls. Ignoring the migration path. Data Center customers need a Cloud migration strategy before they can access the Teamwork Collection. Don't underestimate the complexity. Real Results: What Teams Are Achieving The productivity claims around AI tools often feel vague. Here are specific, quantified outcomes from organizations using the Teamwork Collection: 105 minutes saved daily per person through human-AI collaboration (Source: Atlassian AI Collaboration Report) 50% reduction in time spent searching for information after implementing unified search 80% of routine development work auto-completed using the PR Generator agent 2,000+ questions answered by the Onboarding Buddy agent in its first month at one organization Business impact extends beyond individual productivity: Projects delivered up to 3x faster in some organizations 31% increase in project success rates on average Up to $20 million in cost savings for a 2,500-person organization through tool consolidation Who benefits most? CIOs see consolidated tools and improved ROI CTOs get faster alignment and shorter delivery cycles Team leads spend less time in status meetings Individual contributors waste less time context-switching between apps Getting Started with Teamwork Collection Let's recap the key points: The Teamwork Collection unifies Jira, Confluence, Loom, and Rovo into a single system powered by the Teamwork Graph. It's available now for Premium and Enterprise customers, with documented productivity gains that justify the investment for most organizations. Your next steps: Evaluate your current tool stack and identify redundancies Calculate potential savings using Atlassian's bundled pricing Reach out to Atlassian or a certified partner for a detailed assessment Plan a pilot rollout with one team before expanding organization-wide How On Point helps: As an Atlassian Gold Solution Partner with 11+ years of ERP and collaboration experience, we guide organizations through Teamwork Collection implementation from start to finish. We don't just handle the technical setup. We provide the hand-holding support your team needs to actually adopt new ways of working, from licensing analysis to change management to ongoing training. Whether you're migrating from Data Center, optimizing existing Cloud licenses, or rolling out AI agents for the first time, we can help you unlock the full potential of the Teamwork Collection. Contact us for a personalized assessment of how the Teamwork Collection could work for your organization. Frequently Asked Questions How do I unlock Atlassian Teamwork Collection features in my existing Cloud instance? You'll need to upgrade to a Teamwork Collection Premium or Enterprise license. Contact Atlassian or an Atlassian Solution Partner to transition your existing licenses. The process involves a licensing change rather than a technical migration, but we recommend planning for change management to help teams adopt the new capabilities. What is the best way to unlock Atlassian Teamwork Collection value for a small team? Currently, the Teamwork Collection requires Premium or Enterprise plans, which may be more than a small team needs. Consider whether your team would benefit from consolidating existing tools first. When the Standard tier launches in 2026, it will provide a more accessible entry point for smaller organizations. Can I unlock Atlassian Teamwork Collection on Data Center? No. The Teamwork Collection is only available on Atlassian Cloud. Data Center customers must migrate to Cloud first. This migration involves moving your data, reconfiguring apps, and retraining users, so plan accordingly. Which Rovo agents should I configure first to unlock Atlassian Teamwork Collection benefits? Start with the Customer Insights Agent if you handle support tickets or feedback, or the Brand Voice Crafter if content consistency is a priority. These available-now agents deliver immediate value while you wait for upcoming agents like Brainstorm Facilitator and Workflow Builder. How long does it take to unlock Atlassian Teamwork Collection productivity gains? Most organizations see initial productivity improvements within the first month, particularly from Rovo Search reducing time spent hunting for information. Full adoption of AI agents typically takes 2-3 months as teams learn to incorporate autonomous teammates into their workflows. Plan for a 90-day pilot before measuring ROI. Do I need technical expertise to unlock Atlassian Teamwork Collection capabilities? Basic setup can be handled by any organization admin. However, configuring Rovo agents for your specific workflows and integrating with third-party tools benefits from experience. This is where an Atlassian Solution Partner can accelerate your implementation and help you avoid common configuration mistakes.

Most JSM knowledge base setups fail because teams don't understand three things: They think they need to buy a Confluence license for every customer who reads an article (they don't, if configured correctly). They write articles the way humans scan them (in prose walls), which breaks Atlassian's AI engines trying to parse the content for Rovo-powered deflection. They build the knowledge base as a static library instead of a living system where agents turn solved tickets into articles in one click. The setup itself takes 30 minutes. Getting it right — so it survives your licensing audit, works with your AI layer, and scales as your team grows — requires understanding the licensing paradox, the AI layer, and the Knowledge-Centered Service loop. This guide covers all three. The Confluence Licensing: How to Give Portal Access Without Buying Seats The Problem Every IT Director Hits You set up a knowledge base. You link a Confluence space. You publish 10 articles. Then your admin counts how many customer portal users just accessed the KB and panics: "Do we need 500 Confluence licenses now?" The answer is no, but only if you understand the distinction between how Confluence content is accessed versus how Confluence is licensed. Confluence licensing charges for "product access", the ability to log into Confluence and write/edit pages. Reading an article through JSM's customer portal doesn't trigger a license, as long as that Confluence space is configured correctly. Reading an article inside Confluence itself (not via JSM) does trigger a seat cost. This distinction is where most setups go sideways. The Permission Blueprint: Who Can Read What, Without Licensing Bloat Use this table as your configuration reference. Different user roles, different permission requirements, zero hidden licensing bombs: User Role JSM Project Access Confluence Space Permission Atlassian License Cost Notes Service Desk Agent Service Desk Team Can View, Edit, Publish Paid: JSM Agent + Confluence License Agents need licenses to write/edit articles. These are your $8–10/month team seats. Internal Employee (Contractor/Staff) Service Desk Customer Can View Only (via Portal) Free ($0) They read KB through JSM portal. No Confluence login needed. No seat cost. Configured via portal-specific permissions. External Client (B2B Customer) Service Desk Customer Anonymous Portal Access Free ($0) They never see Confluence. They read KB through JSM public/client portal. Requires "Allow anonymous access" toggled on. No account, no seat. Partner/Reseller (Portal) Service Desk Customer (Organization-scoped) View Only (Portal) Free ($0) Same as external, but scoped to their organization. Still no Confluence access, no seat cost. The key insight: Confluence product access is separate from content visibility. You can have 5,000 portal users reading Confluence articles without a single additional license, as long as they're reading through JSM's portal interface and the Confluence space is set to either "All portal users" or "Anonymous" under visibility. Step-by-Step: Configuring Portal-Only Access (No License Required) In your JSM project settings: Go to Knowledge base → Select the linked Confluence space Under Visibility, choose: "All logged-in portal users" — internal employees and contractors see articles in the portal without needing Confluence seats "Anonymous (public web)" — external customers on your public portal see articles without any account at all Under Confluence Space Settings (in Confluence itself): Ensure the space is set to Restricted in Confluence's native permissions Restrict who can edit (agents only) via the Confluence space admin settings Do NOT set the space to "Public" or "Open" (that would require licensing for different access models) Why this works: JSM's portal renders the Confluence articles inside JSM's interface. The reader never logs into Confluence. They access zero Confluence "product features," so Atlassian doesn't charge a license seat. The content is served, the reader never touches Confluence, the license bill stays flat. Audit your current setup: If your team is paying for more Confluence licenses than you have people editing articles, you probably have permissions misconfigured. Pull your Atlassian billing report and your Confluence product access list — they should match (agents) plus a small buffer (admins). If they don't, you're overpaying. Is your Knowledge Base Set up So Atlassian Rovo Can Actually Parse Them Why Your Article Library Fails the AI Test In 2026, Atlassian Intelligence and Rovo use Retrieval-Augmented Generation (RAG) to read your Confluence space. When a customer types a question in the portal or via Slack, Rovo searches the KB, finds relevant articles, and — critically — extracts the exact answer from the article text. If your articles are prose walls, Rovo can't find the answer efficiently. It either hallucinates a guess or surfaces a vague article title without the specific workaround, forcing the customer to click through and read manually. That kills deflection before the AI even tries. The solution is a structured article that an RAG engine can parse. Writing for the AI Layer: The Formatting Rules Rule 1: One Answer Per H2, Maximum 400 Words Per Section Structure articles so that each major problem gets exactly one H2 header with one complete answer beneath it. If you have three different "How do I reset X?" problems, give each its own H2 and self-contained answer block. Bad (AI struggles here): How to reset passwords and manage access Employees often need help with password resets. You can do this in several ways depending on the system. First, you might try the password reset portal, or you could submit a ticket, or sometimes you can use the command line... [continues for 8 paragraphs] Good (AI parses this cleanly): How to reset your password through the self-service portal 1. Go to [portal URL] 2. Click "Forgot Password" 3. Enter your email 4. Check email for reset link (3-minute expiry) 5. Create new password (12+ chars, mixed case required) **If you don't receive the email:** Check spam folder. If still missing, [submit support request]. Rovo reads the second version and immediately knows: Portal reset in 5 steps, email requirement, what to do if it fails. The first version leaves Rovo guessing. Rule 2: Use Semantic Hierarchy (H1 → H2 → H3, Not Just Bold Text) RAG engines read heading structure, not visual formatting. If you have a "Prerequisites" section, make it an H3 under a clear H2 parent. Don't bold the word "Prerequisites" and hope the AI figures it out. Rule 3: Label Every Article With Semantic Tags Confluence labels are the signal Rovo uses to restrict auto-search recommendations to the right request type. Use consistent label naming: request-type:password-reset request-type:access-provisioning severity:p1-critical product:business-central audience:internal or audience:external When a customer submits a request with type "Password Reset," Rovo only searches articles tagged request-type:password-reset. Without labels, it searches everything and brings fluff. Rule 4: Lead With the Action, Not the Explanation Your first sentence should be the fix, not the problem description. Not this: "Many users have trouble with authentication systems because of token expiry. This can happen when your session times out..." This: "To log back in after a timeout, use [URL] and select 'Resend verification code.'" Rovo pulls the first 1-2 sentences. Make them the answer. Rule 5: Use Bullet Lists for Steps, Not Prose Paragraphs Numbered lists with short, scannable steps are 10x better for RAG parsing than procedural paragraphs. Rovo can extract "Step 3: Click 'Approve'" from a list instantly. It struggles to extract the same instruction buried in paragraph prose. Audit Your Existing KB for AI-Readiness Pull your top 10 highest-traffic articles on your knowledge base setup. For each, ask: Does it have clear H2 headers for each distinct problem? Is the first sentence the action/fix, not background? Are steps numbered lists, not prose? Is it tagged with request-type:* labels? Does it stay under 400 words per section? If 3+ articles fail these checks, your AI deflection is probably under-performing. Rovo can work with what you have, but it won't excel until content is restructured. Using Resolved Ticket to Create Live Article in One Click The Problem: Articles Written in a Vacuum Most knowledge bases grow like this: Manager decides "we need a KB" Team writes 10 articles from guesses Customers search for answers not in the articles Tickets pile up Nobody updates the KB because the agent who solved ticket #4,827 never documented it The Knowledge-Centered Service (KCS) model flips this: the agent solving a complex ticket is the expert who should write the KB article, not someone months later trying to reverse-engineer what happened. In JSM, this loop can be almost automatic. Setting Up the Agent-Driven Article Creation Workflow Step 1: Enable the "Create Article" Button in Agent View In your JSM project settings: Go to Knowledge base → Create article from issue Toggle this ON Set the article workflow to include an approval state (Draft → Under Review → Published) Now when an agent resolves a ticket, they see a "Create article" button directly in the issue view. One click, and a draft article based on the ticket description + resolution appears in your Confluence space. Step 2: Create a Template for Draft Articles Don't let agents start blank. Create a Confluence template that pre-fills the structure: Title: [Autofilled from ticket summary] ## Problem [Autofilled from issue description] ## Solution [Space for agent to fill in] ## Related Issues [Autofilled links to similar tickets] --- *Drafted from ticket [TICKET-ID] on [date]. Awaiting approval.* When an agent creates an article from a ticket, they get this template and only have to fill in the "Solution" section. That's low friction. Step 3: Assign Approval to a "KB Owner" Role Don't let every draft auto-publish. Create a service project role (e.g., "KB Content Owner") and route all drafted articles through that person for 24-hour review before publishing. This prevents: Articles with confidential customer data Typos or incomplete solutions going live Duplicate articles (the KB owner can spot that during review) The approval step takes 2 minutes per article but saves dozens of bad articles from reaching customers. Step 4: Measure KB Creation Volume Alongside Deflection Add this to your monthly reports: Articles created from tickets: Should trend up as agents engage Articles approved vs. rejected: If rejection rate >10%, your template or process is broken Time from ticket resolution to article publication: Target <3 days (draft creation + review) Once you have 5+ agents contributing articles monthly, your KB scales without the manager writing everything. A Real KCS Loop in Action Tuesday 10 AM: Customer submits ticket "Business Central currency conversion failing for USD→EUR" Tuesday 2 PM: Agent debugs, finds root cause (decimal place configuration in chart of accounts) Tuesday 2:15 PM: Agent clicks "Create article" → template pre-fills → Agent types 200-word solution Tuesday 4 PM: KB owner reviews draft, adds label request-type:business-central + severity:p2, approves Wednesday 9 AM: Article goes live tagged and auto-search-enabled Thursday: Next customer with the same issue searches the portal, finds the article, self-serves. Ticket deflected. That's the loop. The knowledge base grows because agents who solved the problem documented it, not because a manager guessed what customers need. Putting It Together This is a licensing defense, an AI optimization, and a living knowledge system: Layer 1 (Licensing): Portal-only access with correct permissions = zero hidden Confluence license costs. Your bill stays flat even as customer portal traffic grows. Layer 2 (AI): Structured articles with semantic headers, labels, and short paragraphs = Atlassian Rovo can parse them accurately. Deflection jumps from 20% to 40%+ when AI can extract specific answers. Layer 3 (Growth): KCS loop where agents turn resolved tickets into articles = your KB grows from your actual ticket patterns, not from assumptions. It scales without hiring a dedicated content person. Teams that execute all three simultaneously see: Ticket deflection increase 30–50% in the first 6 months Confluence licensing costs decrease or stay flat despite 3x portal traffic growth First-time deflection rates on new article topics (agent knows the exact answer because they just solved it) Teams that skip one usually stall. Skip licensing clarity and you overpay. Skip AI formatting and deflection caps at 15–20%. Skip KCS and your KB becomes stale. How to configure JSM knowledge base access without Confluence license costs: User Type Can Access KB Via Confluence License Needed Configuration Service Desk Agent Confluence editor + JSM agent view Yes ($8–10/mo) Grant Confluence space "Edit" permission Internal Staff / Contractor JSM Portal (web or mobile app) No Set Confluence space to "All logged-in portal users" visibility External Customer / B2B Client JSM Public Portal No Set Confluence space to "Anonymous (public)" visibility Partner / Reseller Org-scoped JSM Portal No Create organization in JSM, assign portal scope, leave Confluence permissions restricted Bottom line: Confluence licenses cover editing. Reading through JSM's portal interface doesn't trigger a license, regardless of how many portal users you have. FAQ: Licensing, AI, and KCS Questions Q: We use Atlassian Intelligence (Rovo). Do we need to do anything special to our knowledge base? A: Rovo works out of the box once your Confluence space is linked, but it performs much better if your articles follow the AI formatting rules above (semantic headers, labels, short paragraphs, action-first). Run an audit of your top 20 articles against those rules; if <50% comply, spend a week reformatting. You'll see deflection jump. Q: Can we auto-generate articles from tickets using AI? A: Atlassian Intelligence can draft articles from ticket text, but don't let them auto-publish. They often capture extraneous details or miss the core solution. Use the KCS approval state so a human can review before live publication. Q: Our Confluence space is marked "Public." Are we overpaying for licenses? A: Possibly. Pull your Confluence product access list. If it includes everyone in your portal user base, yes — you're paying a license per portal user. Change the space to "Restricted" and rely on JSM's portal permissions instead. You'll drop license count immediately. Q: How do we know if articles are AI-friendly? A: Rovo includes a "Search suggestions" report showing which searches returned no results or low-confidence matches. Those are your gaps. Pull that report quarterly; articles that don't appear in successful searches probably need reformatting (better headers, clearer labels, shorter sections). Q: Should we write separate articles for humans vs. AI? A: No. One well-structured article works for both. The AI-friendly formatting (semantic headers, labels, action-first) is also more scannable for humans. You're not optimizing for robots at the expense of people. Where This Fits in Your ITSM Strategy A knowledge base is the first deflection layer in a mature ITSM practice. Once KB deflection is working (20–40% of requests), that frees your team to focus on the recurring incidents your KB can't catch because nobody's fixed the root cause. That's where incident management and problem management enter the picture, every problem record should generate a KB article that prevents the next round of tickets. For teams building JSM from scratch, this also connects to your ITSM implementation roadmap — knowledge base maturity is often the first measurable quick win before you tackle deeper process work. Further Reading Atlassian Support: The difference between internal and external knowledge bases — Official guidance on how visibility permissions avoid license creep Atlassian Support: Set up your knowledge base with Confluence — Step-by-step technical setup Knowledge Base management in JSM best practices — Operational best practices Full suite of ITSM metrics including KB utilization Setting up a JSM knowledge base properly means more than clicking "create space." It means understanding your licensing model, formatting for AI, and building a system where agents keep it alive. Get in touch if you want to audit your current setup against these three layers, licensing spend, AI readiness, and KCS automation.

Start off your program by giving some basic information about what participants can expect to learn. Introduce the topic and provide a basic outline of what's to come using videos, images, and text.

Break down your content into manageable steps. Use videos, images, and text to explain your main points, and set clear expectations and goals. You can also include files and audio as additional resources. Consider adding a questionnaire to ask participants what they hope to take away from the experience, or a quiz to test their understanding before getting started.

1. Chart of accounts 2. Account payables 3. Account receivables 4. Bank accounts 5. Bank reconciliation 6. 1 company 7. VAT setup Additional :Customized company document templates 1. Sales & Purchases invoices 2. Credit memos 3. Statements - 8 hours of training. - 3 months post-implementation support Total cost = $9,000

- Contacts - leads and Opportunities - marketing campaigns - Sales quotes - Approvals - analytics * 5 hours training * 3 months post-implementation support Cost : $7,000

* Employee data management ( contract, qualifications etc.) * Leave management development required * Payroll management - development required $9,000 NB : Payroll & leave management pricing might go up based on company's requirements