Sample ForIntel Citation Drift Atlas — Who the AI Answers Cite, and Where You're Slipping

What this sample shows

This is a public sample of a ForIntel Citation Drift Atlas deliverable, published by Foragentis to demonstrate the method. It was prepared for ClickUp (clickup.com), the project-management and work-management software company, and ClickUp is named throughout with its explicit approval — including the finding that its position is slipping behind a competitor, which is true and data-backed. The market findings — the AI-citation map, the competing review sites and vendors named in it, and the category demand read — describe a public answer surface, not ClickUp's internals, and are preserved in full.

It demonstrates what the Atlas reads: a citation-drift map of the project-management software category built from three layers — a three-engine AI-citation snapshot (which sources each engine hands buyers today), a measured 12-month organic-position trajectory for ClickUp's own pages, and a category demand universe — each carrying an explicit confidence label, with the scoped boundaries named rather than glossed. This Quarter-0 issue establishes the AI-citation baseline; true citation drift in the AI surface is what a repeat reading measures against it, and is not claimed here.

The verdict

The AI answer set has drifted away from ClickUp — and the measured slide in ClickUp's organic position is the cause, not a coincidence.

Buyers increasingly ask an AI engine, not a search box, which project-management tool to use — and the engine answers by citing live web sources. We read this on two paired axes. The citation map (Quarter-0 snapshot): across the engines that search live, ClickUp's own pages are a thin presence in the answer set — cited by one engine, absent from the other — while a small set of independent review sites hold the positions that decide the answer. The position trajectory (measured, 12 months): ClickUp's organic rank on the category head terms has slid downward on 7 of 9 head terms over the trailing year, even though ClickUp converts the fewest of its enormous keyword footprint to the first page. The two axes connect: the engines read their answers off the pages that out-rank ClickUp, so an organic slide becomes a citation slide. The play is to recapture the cited reference positions — starting on the terms where the decline is steepest.

1. When buyers ask the AI engines, ClickUp's pages are barely in the answer set. We asked the three major AI engines the questions buyers actually type — the best project-management tool, the best free option, ClickUp versus Asana, ClickUp versus monday.com — and recorded which sources each engine cited. The two engines that performed live web search returned 228 citations across some 86 distinct sources; ClickUp's own domain appears in only one of those two answer sets (cited 6 times by the engine that cites most broadly, and absent entirely from the other). One direct competitor's own pages are cited more often than ClickUp's. This is a current-state citation map, and it does not have ClickUp holding the positions a category leader should.
2. The cited positions are held by a handful of independent review sites — not by any vendor. A single comparison site, thedigitalprojectmanager.com, is cited 22 times across the two grounded engines — far more than any other source and more than any vendor's own pages. A short list of review/comparison domains holds the rest. These are the positions that decide what the engines tell a buyer, and the path to them runs through being the cited reference on those questions, not through brand spend.
3. ClickUp's organic position has measurably drifted downward for a year — and that is what feeds the AI answers. ClickUp ranks for more category keywords than either named competitor (the widest reach in the set), but it converts the fewest of them to the first page. Worse, on the category head terms its own position has slid downward on 7 of 9 head terms over the trailing 12 months — from page-1-adjacent to the 60s and 70s on terms like 'work management platform' and 'project management app.' The engines increasingly read their answers off pages that out-rank ClickUp; the organic slide is the upstream cause of the citation gap.

In one line: The AI engines are handing buyers a category answer that ClickUp's pages are slipping out of — and the measured 12-month decline in ClickUp's organic position is the mechanism. The opening is to win back the cited reference positions on the buyer questions, starting where the slide is steepest.

How to read this report. HIGH means a well-sampled, directly observed signal; MEDIUM rests on a narrower base or an effect we can see but not yet bound; UNDETERMINED means a measurement could not be settled and we say so inline rather than dress a single reading as confirmed. The AI-citation map (Section 01) is a single current-state snapshot — HIGH for the two engines that performed live web search, and a disclosed control for the one that did not. "Citation drift" in the AI surface is what a repeat engagement measures; this issue establishes the Quarter-0 baseline, not an AI-citation trend — we do not claim one the data cannot support. The organic-position drift (Section 03) is measured, over a 12-month trail. Candor about a boundary is not a hedge — it is the part that tells you what you may act on now.

01 · The citation map — where ClickUp stands in the AI answer set today

(Confidence: High.) We asked the three major AI engines the questions a buyer asks before choosing a tool — the best project-management software, the best free option, the best choice for a small or remote team, ClickUp versus Asana, ClickUp versus monday.com — and recorded every source each engine cited. (Source: a current-state Quarter-0 snapshot of the public AI answer set across three engines.) Two of the three engines performed live web search and grounded their answers in cited pages; the third answered from its training memory without searching the live web, so it cites nothing and casts no vote — a disclosed control, not a finding.

The two grounded engines returned 228 citations across 86 distinct sources (one engine cited 138 sources across 69 domains; the other, 90 across 28). Inside that answer set, ClickUp's own pages are a thin presence: the broader-citing engine cited clickup.com six times; the other did not cite the domain at all. By contrast one direct competitor's own pages were cited eight times in one engine's answer set — more than ClickUp's. The engines clearly know the category; the issue is that they are not consistently being pointed back to ClickUp.

Figure — Who answers from live sources, and who doesn't (Quarter-0 AI-citation snapshot; cited sources by engine). Across the two engines that search live, ClickUp's pages appear in one answer set (cited 6 times) and are absent from the other; one direct competitor is cited eight times. The ungrounded engine casts no citation vote and is held as a disclosed control.

One caution we hold honestly: this is a single snapshot. True AI-citation drift — the movement of these positions quarter over quarter — is what a repeat read measures against this baseline. We are not presenting a citation trend the data does not yet contain. What this snapshot establishes is the starting line, and it is not a leading one.

02 · The positions to win — a few review sites decide the answer

(Confidence: High.) When you look at which sources the engines cite, the answer set is concentrated — and it is held by independent comparison and review sites, not by the vendors themselves. (Source: the cited-source composition across the two grounded engines, Quarter-0.) One site, thedigitalprojectmanager.com, is cited 22 times across the two grounded engines, far ahead of every other source and ahead of any vendor's own pages. A short tail of review/comparison domains — project-management.com, cloudwards.net, zapier.com and a handful of others — holds most of the rest. These are the citation positions that decide what the engine tells a buyer.

Figure — The cited positions are held by review sites, not vendors (cited sources across the two grounded engines, Quarter-0). One comparison site, thedigitalprojectmanager.com, dominates the citations at 22 — cited more than any vendor's own pages. Among vendor domains, ClickUp is cited level with peers and behind one direct competitor.

The implication is direct and it is good news as much as bad: the route into the AI answer set is not a brand-awareness problem, it is a cited-reference problem. The engines reward the page that most clearly and credibly answers the buyer's exact question. Winning the comparison and 'best-of' questions — with content the engines will cite, and with placement on the review sites that already hold the positions — is what moves ClickUp's pages from the tail of the answer set toward the front of it.

03 · The drift, measured — ClickUp's organic position has slid for a year

(Confidence: High.) This is the temporal axis of the brief — and unlike the citation snapshot, it is a measured 12-month trajectory. (Source: a 12-month organic-position trail of ClickUp's own pages on the category head terms.) We tracked ClickUp's best organic position on the category head terms across the trailing year. The direction is consistent and it is the wrong one: on 7 of the 9 head terms with enough history to measure, the position got worse over the year. On 'work management platform' ClickUp fell from the mid-teens to the low 70s; on 'project management app,' from 12th to the high 60s; on 'task management tool,' from the low 30s to the low 70s. A single bad month is noise; a year of decline across most of the head terms is a trend.

Figure — Best organic position, month by month, on three representative head terms (lower line = worse). Across all nine measurable head terms, position worsened on seven. This is the engagement's one measured time-series.

The two axes of this brief connect here. ClickUp holds the widest keyword footprint in the category — it ranks for more terms than either named competitor — but it converts the fewest of them to the first page, and that conversion is eroding. The AI engines read their grounded answers off the pages that rank highest; as ClickUp's pages slide down the results, the engines have fewer reasons to cite it and more reasons to cite the review sites and competitors that now sit above it. The organic drift is not a separate problem from the citation gap — it is its upstream cause.

04 · The demand to defend — a large, real market, with a few traps

(Confidence: Medium.) The good news underneath the drift: this is a genuinely large market. (Source: a buyer-side keyword demand universe for the category plus a head-term SERP-feature capture.) The buyer-side keyword universe runs to nearly 1,900 terms carrying measurable search demand, with head terms like 'project management tools' (about 22,000 monthly searches) and 'task management software' (about 18,000) anchoring it. The category is not shrinking; the search audience for what ClickUp sells is real and broad. The slide in Section 03 is a position problem, not a demand problem.

Figure — A large, real demand surface with two read-honest cautions (buyer-side keyword universe, head-term volumes and SERP features). Head terms carry real, broad volume, but single-month spikes inflate the apparent trend, and AI Overview answer boxes appear on more than half the head-term result pages — so the citation map of Section 01 is not a side issue.

Two cautions keep the demand read honest. First, a handful of the head terms show eye-catching single-month spikes — 'task management software' jumped to roughly 135,000 in one month against a typical ~4,400, and 'project management tools' to ~74,000 against a typical ~14,800 — so the 'is the category exploding?' question should be read as steady and large, not as a sudden surge to chase. Second, the live SERPs for these terms increasingly carry an AI Overview answer box — present on more than half the head-term result pages we captured — which means the citation map in Section 01 is not a side issue: for a growing share of these searches, the AI answer is the result. Defending this demand now means winning the cited-reference positions, not only the blue links.

05 · What this means for you — the recapture plan, in priority order

1. Win the cited-reference positions on the buyer questions. The engines cite a small set of comparison and 'best-of' pages. Build (and earn placement on) the content that most credibly answers the exact buyer questions — the comparisons, the free-tier and small-team guides — so ClickUp's pages move from the tail of the answer set toward the front. This is a cited-reference problem, not a brand-spend one.
2. Stop the organic slide where it is steepest first. ClickUp's position has drifted down for a year on most head terms. Triage by where the slide is sharpest and the term is most valuable — 'work management platform,' 'project management app,' 'task management tool' — and rebuild those pages to out-rank the review sites and competitors that overtook you.
3. Convert your footprint, don't just widen it. ClickUp already ranks for more category terms than any competitor; the gap is conversion to page one. Prioritize depth on the terms you nearly rank for over chasing new ones — a far cheaper path to the top positions the engines actually read.
4. Treat the AI Overview as a ranking surface, not a curiosity. More than half the head-term result pages now show an AI answer box. Optimize the high-intent pages to be the cited source in those boxes, because for a growing share of searches that box is the answer the buyer sees.
5. Re-measure on a cadence — that is what turns this snapshot into drift. The single largest open question is which way the AI citations move next. This issue sets the Quarter-0 baseline; a repeat read each quarter converts it into a measured citation-drift trend you can manage against (see the scope note below).

Scope, confidence & what a deeper engagement adds

This issue reads the public category surface across the signals the analysis produced reliably. Below is exactly what each finding rests on — and exactly where the boundary sits. These are scoped diligence boundaries, named with the work that closes each one, not failures.

• The AI-citation read is a single snapshot, not a measured trend. Section 01 establishes where ClickUp stands in the AI answer set today. True citation drift — how those positions move — requires a second reading against this baseline. The recurring engagement re-probes the same buyer-question set each quarter and reports the measured movement; this one-shot delivers Quarter 0 plus the measured organic-position trajectory that points the way the AI answers are heading.
• Three engines, two grounded. Two of the three AI engines performed live web search and ground the citation map; the third answered from training memory without searching, so it casts no citation vote — a disclosed control, not a gap in the finding. A deeper engagement re-runs the third engine with live search enabled so all three contribute to the baseline.
• Per-page content coverage. A full read of which buyer questions each cited page answers best rests on parsing those pages; several of the cited review pages sit behind cookie walls or counter-signal review measures and did not parse cleanly this issue. We therefore did not infer coverage from a partial set. A deeper engagement fetches the full cited corpus and turns 'which questions are under-served' into a ranked content target list.
• Demand durability. The market is large and the head terms carry real volume (Section 04), but a formal 12-month demand-durability verdict — durable growth versus seasonal or news-cycle spikes — could not be settled cleanly this issue and is carried as Undetermined rather than asserted. A deeper pass adds an outlier-resistant trend test across the head terms.
• Attribution methodology. Source-to-claim attributions in this brief are analyst-asserted from the cited corpus — the standard methodology note — rather than machine-pinned line by line. A deeper engagement binds each attribution to a specific cited source row.

This is a Quarter-0 citation map paired with a measured organic-position trajectory. To commission the recurring engagement — quarterly re-probes that turn this baseline into a measured AI-citation drift trend, all three engines grounded, the full cited-page coverage map, and a ranked recapture plan for the positions you are losing — reach the ForIntel desk directly at forintel@foragentis.com or scope it on the ForIntel order page.

This is a public sample of a ForIntel Citation Drift Atlas deliverable, published by Foragentis to demonstrate the method. ClickUp is named with explicit approval — including the finding that its position is slipping behind a competitor, which is true and data-backed; the AI-citation, competitive and demand findings describe a public answer surface and are preserved in full. The AI-citation map is a single current-state Quarter-0 snapshot, HIGH for the two grounded engines and a disclosed control for the third; the organic-position drift is HIGH and measured over a 12-month trail; the demand read is MEDIUM and a formal demand-durability verdict is carried as Undetermined by design rather than presented as confirmed.