Last verified: August 25, 2026
What You Can Now Do
Context Memo now displays related memos directly on each problem node in the Problem Graph board, so coverage is visible on the same surface as demand. Before this, matching a published memo to the buyer problem it answered meant holding two views in your head: the graph on one side, the memo library on the other. Gaps were inferred. Now every node carries the memos matched to it, and a node with nothing attached reads as unmet demand. The decision this changes is the one your team makes weekly: which memo to write next, and which topic you have already covered well enough to leave alone.
Where It Is in Context Memo
Open the Problem Graph board for the brand you're auditing. The board itself is unchanged in layout: problem nodes clustered by topic, connected by relationship edges. What's new is the side panel. Click a node and it expands to show the memos matched to that specific problem statement, ranked by relevance, with the match strength visible per memo.
How to Use It
- Open the Problem Graph for the brand you're auditing. The board loads with problem nodes clustered by topic and connected by relationship edges.
- Click any problem node. The side panel expands to show matched memos, ranked by relevance to that problem statement.
- Read the match strength on each result. A strong match means a memo resolves that problem directly; weaker matches mean the topic is mentioned without being answered.
- Scan the cluster around the node you're inspecting. Compare the memo count on the node you care about against its neighbors, so you can see whether a gap is isolated or the whole cluster is thin.
- Flag nodes that carry high buyer intent and either zero matched memos or only weak matches. That list is your production queue, in priority order.
- Return to the board after publishing. The new memo appears on the node it answers, which confirms the gap closed rather than moved.
Why We Built It
Gap analysis is only useful when it's specific. "We're underexposed on integration questions" doesn't tell a marketing team what to do on Monday. "The node How does X integrate with HubSpot? has zero matched memos while three adjacent nodes each have two" does. That was the gap in our own product: teams told us they could see the taxonomy and they could see the library, but they couldn't see the overlap without exporting both and reconciling by hand. The cost was duplicated effort on topics already well covered, and silence on nodes where buyers were actively asking.
The second effect is quality control, and it's the one most teams underestimate. When a node shows five loosely related memos and no direct match, the topic is being touched but not answered. AI models cite the source that most cleanly resolves the prompt, not the one that mentions the subject in passing. Five near-misses on a single node is often more informative than zero, because it tells you the intent exists and your coverage is diffuse. That's a rewrite decision, not a net-new memo decision, and the board now makes the difference legible.
What It Does Not Do Yet
A few limits worth knowing before you build a workflow on top of this:
- Empty nodes don't queue themselves. Seeing a gap on the board doesn't start a memo. You still pick the node and begin the memo yourself from the memo generation flow.
- Match strength is a signal, not a verdict. Relevance is scored against the problem statement, so a memo can rank as a strong match and still answer the question shallowly. On any node you plan to defend, read the top match before you conclude the node is covered.
- The board reflects what you've published. Coverage is computed from memos in your library, so a node will stay empty until the memo answering it is live. Newly published work shows up on the node it answers, not on the nodes it merely touches.
- This is a coverage view, not a citation view. The node tells you whether you have content addressing the problem. Whether models are citing that content is still read in citation tracking and share of voice.
One practical note on sequencing: the board is most useful for brands that already have a body of published content and some domain authority behind it. If your library is close to empty, nearly every node will read as a gap, which is accurate but not informative. Build the foundational coverage first, then use the board to find what's missing.