How AI Search Is Recommending Credit Cards for Building Credit: Monthly Trends

Executive Summary

OpenSky leads AI recommendation coverage this month at 38.5% valid recommendation coverage, with Chime second at 35.8%, a gap of 2.7 percentage points. In July 2026 the two brands were nearly tied near the top of the category; both fell substantially in August 2026, with OpenSky holding on to the top position.

Chime registered the largest absolute decline of any tracked brand, falling from 74.3% in July 2026 to 35.8% in August 2026, a drop of 38.5 points. OpenSky fell 36.0 points over the same span, and Self dropped 20.1 points. Navy Federal Credit Union and Discover Home Loans also posted declines large enough to stand out from normal month-to-month variation. The five brands with the largest declines span both category leaders and mid-tier brands, indicating the change in measured output distribution was not confined to a single segment of the category.

The benchmark's qualified observation count fell from 534 in July 2026 to 486 in August 2026, even though the total prompt collection held steady at 800 in both months. This reflects a narrower qualification funnel in August 2026: fewer of the collected prompts produced responses that met the bar for inclusion in the public metrics.

Each monthly benchmark run begins with 800 prompt-surface observations across the defined AI/search surface universe (676 unique questions in July 2026; 701 in August 2026). All 800 prompts mentioned a tracked brand or competitor in both months. Of those, 595 were relevant and 205 were irrelevant to the category in July 2026, compared with 563 relevant and 237 irrelevant in August 2026. The public metrics use the 534 qualified observations from July 2026 and 486 from August 2026 that survive both qualification stages.

Key Findings

Benchmark Context

The report separates the raw collection universe from the qualified analysis set. Brand-level recommendation percentages are calculated within the qualified benchmark set.

The qualification funnel narrowed between the two months: total collected prompts held flat at 800, but the relevant share fell from 595 to 563 and the qualified set fell from 534 to 486. That narrower denominator is part of the context for reading the benchmark-level percentages below.

Benchmark-Level Metrics

The qualified observation count fell from 534 to 486, a change driven by the narrower qualification funnel described above rather than by a change in the total prompt collection. This shift in denominator size is part of the context for the coverage declines documented below, though it does not by itself explain why individual brand coverage rates fell.

AI Recommendation Trend

Category-wide coverage contraction: all tracked brands declined, leadership retained by OpenSky

The category recorded a broad decline in valid recommendation coverage between July 2026 and August 2026. Every tracked brand declined, with five registering drops large enough to stand out from normal month-to-month variation: Chime, Discover Home Loans, Navy Federal Credit Union, OpenSky, and Self. OpenSky retained the leadership position despite posting one of the largest declines among the top tier, while Chime's decline was the largest in absolute points.

This month's movement was not confined to one or two brands. Five of the eight tracked brands registered declines large enough to stand out from normal month-to-month variation, and the remaining three (Applied Bank, Capital One Auto Finance, First Latitude) also declined in absolute terms without crossing that threshold. The pattern reflects a broad, simultaneous shift in the measured output distribution across nearly all tracked brands rather than an isolated brand-level change.

What Changed This Month

OpenSky: Leader retains position but coverage nearly halves

OpenSky's valid recommendation coverage fell from 74.5% in July 2026 to 38.5% in August 2026, a decline of 36.0 points. The prior month's coverage put OpenSky in a near-tie with Chime at the top of the category; this month OpenSky holds a 2.7-point lead over Chime.

OpenSky's top-three recommendation rate fell from 56.5% to 28.2%, and its rank-one rate fell from 31.6% to 18.1%, indicating the decline ran across the entire recommendation ladder rather than only at the margin. Raw mention presence also fell sharply, from 80.2% to 42.6%.

Highest-priority diagnostic: Which prompt categories drove the sharpest loss of top-three placement for OpenSky, and which brands replaced it in those responses?

Chime: Largest absolute decliner

Chime fell from 74.3% valid recommendation coverage in July 2026 to 35.8% in August 2026, a decline of 38.5 points, the largest absolute drop of any tracked brand this month.

Chime's rank-one rate fell from 8.8% to 4.9%, and its top-three rate fell from 53.0% to 23.7%. Despite the decline, Chime remains the second most-recommended brand in the category. The brand's net sentiment score held near 0.9 in both months, indicating the coverage decline was not accompanied by a shift toward negative framing.

Highest-priority diagnostic: Which surfaces and prompt types saw the largest drop in Chime recommendations, and did the responses shift to competitors or to no-brand answers?

Self: Coverage contraction

Self's valid recommendation coverage fell from 43.6% in July 2026 to 23.5% in August 2026, a decline of 20.1 points.

Self's raw mention presence fell from 48.3% to 31.1%, and its top-three rate fell from 24.0% to 18.1%. The brand's rank-one rate moved from 3.6% to 2.7%, a modest change suggesting Self held on to some top-spot positioning even as its overall coverage contracted.

Highest-priority diagnostic: Which prompt phrasing or intent categories drove Self's presence decline, and whether the loss was to competitors or to a broader shift in how building-credit questions are being answered.

Navy Federal Credit Union and Discover Home Loans: Mid-tier declines

Navy Federal Credit Union fell from 11.8% coverage in July 2026 to 6.8% in August 2026, a decline of 5.0 points. Its valid recommendation count fell from 63 to 33, and raw mention presence declined from 20.2% to 11.1%.

Discover Home Loans fell from 4.5% to 2.1% coverage, a decline of 2.4 points. Its valid recommendation count fell from 24 to 10, and its rank-one rate fell from 0.9% to 0.0%.

Highest-priority diagnostic: Whether these mid-tier declines reflect the same category-wide pattern as the leaders, or whether specific competitor gains at their expense can be identified in the underlying prompts.

First Latitude: Dropped out of the qualified set

First Latitude had 2 valid recommendations in July 2026 (0.4% coverage) and zero in August 2026. The counts involved are small enough that this change should be read with caution rather than as a firm signal. The brand had no raw mention presence in August 2026.

Highest-priority diagnostic: Whether First Latitude's disappearance reflects the tighter qualification funnel or a genuine loss of surface presence in the August 2026 question set.

Buyer-Intent Interpretation

In August 2026, all 486 qualified observations fell into the Brand Recommendation cluster, with none in the Pricing & Value or Multi-Brand Comparison clusters. The public benchmark therefore measures which brands AI systems name when recommending a credit card for building credit, but it does not yet capture price sensitivity, fee comparison, or head-to-head preference questions. Those commercial questions require a deeper, prompt-level analysis than the public benchmark provides.

Brand Opportunity Summary

The benchmark identifies where attention is warranted; a company-level analysis is needed to explain why these patterns emerged.

Evidence Behind the Benchmark

The aggregate metrics are built from prompt-level observations (query, surface, recommendation outcome, rank, sentiment, and citations where exposed). Company-level analysis can go deeper into prompt, competitor, surface, and evidence patterns. Source presence is not automatically treated as proof of causation.

This report is part of the CiteWorks Studio AI industry market discovery research program.

• AI Industry Market Discovery Methodology
• AI Industry Market Discovery Metrics
• AI Industry Market Discovery Standards

Interpretation Notes

• Small-count movement: Several brands (Applied Bank, First Latitude) have very small valid recommendation counts. Their percentage movements, while worth tracking, rest on a small number of observations and should be read with that limitation in mind.
• Qualified denominator vs raw collection: The public metrics use 486 qualified observations in August 2026, a subset of the 800 prompts collected. The qualification funnel narrowed this month, which affects comparability with July 2026.
• Directional analysis: Month-over-month movement identifies changes worth investigating; it does not by itself establish the cause of those changes.

Next Step

The Public Benchmark Shows Where a Brand Is Winning or Losing. A Company-Level Audit Shows Why.

Beneath the aggregate percentages sit unanswered questions: which high-intent prompts are won, which competitor takes the recommendation when a brand loses, what attributes AI systems associate with each option, and which external sources shape those answers. The public benchmark shows that coverage contracted broadly in August 2026, but it does not explain why some brands held their rank-one positioning while others lost ground across the board.

A company-specific AI visibility audit maps those prompt, surface, competitor, ranking, sentiment, and evidence-source patterns into a prioritized visibility strategy. That analysis turns the benchmark's findings into a concrete action plan for reclaiming or defending AI recommendation share.

Request an AI visibility audit