How Global Partners built one product layer over 700+ Jira epics – with airfocus

Global Partners is a Fortune 500 company and industry-leading integrated owner, supplier, and operator of liquid energy terminals, fueling locations, and guest-focused retail experiences, operating one of the largest terminal networks in the US, from Maine to Florida and into the Gulf states.

The company runs one of the largest integrated liquid energy terminal networks in the US, and its Data, Analytics, and Insights (DAI) organization sits at the center of that operation, turning enterprise data into decisions across BI, data science, and executive reporting. But as the team scaled, so did the coordination tax: hundreds of epics, multiple teams, and no shared layer above the delivery work to hold it all together. What DAI needed wasn't another list to maintain. It needed a product intelligence platform that could sit above the noise, and eventually, one that an AI could reason over directly.

"airfocus fits as the product source of truth in that architecture: it's where products, initiatives, and roadmap intent live," says Nick Pranovich, Technical Product Manager at Global Partners.

airfocus lets you build whatever hierarchy your organization needs, and that is a genuine superpower of the product.

Nick Pranovich

Technical Product Manager – Global Partners

Scale, fragmentation, and no product layer above the delivery work

"Our team's biggest challenges were scale and fragmentation. We had hundreds of epics in flight in Jira, multiple teams with their own priorities, and a need for a coherent picture of what the roadmap was, what it would deliver, and when," says Nick Pranovich, Technical Product and Program Manager at Global Partners. 

1. No shared definition of a product: "The delivery layer (Jira) was healthy, but there was no product layer above it: no shared definition of what a 'product' or 'initiative' was, and no consistent way to answer 'what's the status of X and what business value does it carry?' without someone manually assembling it."

2. Tools that worked, then didn't scale: "We tried several approaches over time: Smartsheet, spreadsheets, and even a custom Jira issue type acting as an 'umbrella' parent level above epics for product-level tracking. Each of these worked for a while, and each became unsupportable with every scale-up wave."

3. Capacity planning outgrowing Excel: "The spreadsheet setup was fairly sophisticated: we built scripts, dynamic sorting, and a couple of what-if scenarios. But capacity planning was very hard to do well in Excel."

There wasn't a single breaking point. "Less a single moment than an accumulation. The spreadsheet approach worked, but we wanted things it could not offer at scale: asynchronous parallel updates by team members, real collaboration, and native linkage with Jira."

A product operating model built for AI, with airfocus as the foundational source of truth

1. AI transformation – airfocus as the product layer behind the AI

"AI is central to how we work. In product operations, the highest-leverage AI use case we've found is connecting our tools through MCP (airfocus, Jira, and Confluence) and letting an AI assistant (Claude, in our case) work across all three," says Nick.

• airfocus as the source of truth: "airfocus fits as the product source of truth in that architecture: it's where products, initiatives, and roadmap intent live, and having it accessible via MCP means the AI can reason across strategy (airfocus) and execution (Jira + Confluence) simultaneously. That combination unlocked new opportunities."

• Hierarchy flexibility: "airfocus lets you build whatever hierarchy your organization needs, and that is a genuine superpower of the product. We needed team-level workspaces where each team owns its backlog, rolling up into a strategic portfolio view, and airfocus lets us model exactly that."

• A gate-passing Jira integration: "This was one of the gate-passing factors during our POC; without a solid bridge to the execution layer, no product tool would have worked for us."

• From planning tool to live data layer: "Flexible custom fields shared across workspaces (Target Delivery Quarter, Condition, Priority, Business Unit, and so on) let us standardize planning language across teams without forcing them into identical processes. And important: the MCP server and API openness, which turned airfocus from a planning tool into a live data layer for our whole product operating model."

2. Connecting hundreds of epics across seven teams

"The biggest problem at our scale is connecting work pieces. With 700+ epics across projects and seven teams, different leaders naturally use different semantics for their epics and products, and the related work suggestions are very useful for bridging that: items that are described differently but belong together get connected in airfocus," says Nick.

3. A partnership that matched the size of the bet

"The customer success relationship has been genuinely good: responsive, and receptive to product feedback. When we've raised gaps, they've been taken seriously and routed to Product. For a tool that sits at the center of our operating model, that partnership mattered as much as the feature list."

A single product taxonomy, hours back every week, and an MCP layer the team relies on daily

• Executive reporting from live data: Global Partners’ roadmap status report joins ~700 Jira epics to airfocus products across four Jira projects, “with a need to eventually report on completion percentages, condition tags, status notes, and owner filters, this is now produced via the MCP and API integration."

• The spreadsheet layer retired: "Maintaining the Excel setup was more time-consuming and it did not scale. Today we save at least a few hours per week, and just as importantly, roadmap maintenance is now distributed across the team instead of bottlenecked on one person."

• Faster to find and connect work: "It's easy to search and find the required item across the whole portfolio. When linking two items, the dropdown is semantically sorted so the most relevant items are always at the top. This is amazing in day-to-day use."

• One taxonomy across strategy and delivery: Nick's team organized ~700 Jira epics into product-aligned clusters, each mapped to airfocus products, with a Jira custom field linking every epic back to its airfocus product.

• An MCP server the team leans on daily: "The MCP server wasn't why we chose airfocus; we would have selected it either way. But once it was added, we gave it heavy usage very quickly and now rely on it," says Nick. Today the team uses it for high-level roadmap status reporting, portfolio health reviews, generating planning-cycle objective statements, and enriching product descriptions by synthesizing linked Jira epics and Confluence PRDs, with "every write gated on human approval."

• Judgment support, not just speed: Asked whether airfocus AI helps teams build the right products, not just work faster, Nick's answer was direct: "Yes. The deeper value for us is judgment support. Before an executive review, we know the picture we're presenting reflects live data. Deciding better is what builds the right products, and for that positioning to stand, the product layer has to be structured, connected, and machine-readable."