How AI Search Is Recommending Language Learning Software: Monthly Trends

Executive Summary

Duolingo remains the coverage leader in Language Learning Software with valid recommendation coverage of 83.0% in September 2026, holding a 6.1-point edge over Babbel's 76.9%. Babbel holds the clear second position, up 2.8 points from August's 74.1% and up 2.3 points from the July baseline of 74.6%. The top two brands are stable and separated from the rest of the field.

The strongest upward mover this month was Lingoda, rising 2.9 points from 18.2% in August to 21.1% in September, though this remains within normal variation. Pimsleur continues a two-month upward streak, reaching 66.8% from a July baseline of 63.1%. The sharpest decliner was italki, falling 3.7 points from 49.4% in August to 45.7% in September, though the most significant story of the quarter remains Rosetta Stone's 11.9-point decline from 48.5% in July to 36.6% in September.

Against the prior month, the category was quiet: no brand exceeded the significance threshold for month-over-month movement in September. The most consequential baseline-to-current finding is Rosetta Stone's significant decline, which widened its gap with Pimsleur and Babbel in every month of the series.

Each monthly run begins with 800 prompt-surface observations across the benchmark's defined AI/search surface universe. Unique questions totaled 464 in July, 525 in August, and 518 in September. All 800 observations each month mentioned a tracked brand or competitor. Relevant prompts totaled 800 in July, 798 in August, and 792 in September; irrelevant prompts totaled 0, 2, and 8 respectively. The public metrics use 520 qualified observations in July, 522 in August, and 558 in September 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.

These qualified counts anchor every brand-level percentage below; the raw collection volume is shown for context only and is not used as a denominator anywhere in the report.

Benchmark-Level Metrics

August 2026 is the intermediate month in this series, with 522 qualified observations and a recommendation-shaped answer share of 66.9%.

AI Recommendation Trend

The top tier is stable and wide; the competitive battle is for third through sixth

September 2026 shows the same commercial structure as July: Duolingo and Babbel occupy a tier of their own, while the middle of the category continues to reorder. Duolingo's 83.0% and Babbel's 76.9% leave Pimsleur at 66.8% as the clear third-place challenger, 10.1 points behind Babbel and 21.1 points ahead of italki.

The category-level change in September came primarily from the combination of several smaller movements rather than any single significant shift. No brand exceeded the significance threshold for month-over-month movement. The most material movement in the full series is Rosetta Stone's 11.9-point baseline-to-current decline.

What Changed This Month

Rosetta Stone: Significant decliner across the quarter

Rosetta Stone's valid recommendation coverage fell 11.9 points from 48.5% in July to 36.6% in September, a significant baseline-to-current decline exceeding the threshold for normal variation. The steepest single-month movement was 12.3 points between July and August; September was essentially flat against August at 36.6% versus 36.2%.

The decline is accompanied by a drop in raw mention presence from 58.3% to 46.2%, down 12.1 points. Top-3 placement fell from 10.4% to 5.6%, and rank-one presence declined from 1.5% to 0.7%. Rosetta Stone's valid recommendation count fell from 252 in July to 189 in August, then recovered slightly to 204 in September.

The distinction to notice: Rosetta Stone's shift shows up primarily in visibility, not sentiment. When it is mentioned, sentiment remains broadly positive at 0.8 net sentiment, but it appears in far fewer responses than at the start of the quarter. The benchmark cannot distinguish platform behavior from measurement effects.

Highest-priority diagnostic: Which surfaces or prompt categories reduced their mention of Rosetta Stone across the quarter, and which evidence sources are those AI systems relying on instead?

Babbel: Stable leader gaining ground

Babbel's valid recommendation coverage rose 2.8 points from 74.1% in August to 76.9% in September, its highest level in the tracked series and up 2.3 points from the July baseline of 74.6%. This is the first month of an upward streak, though it remains within normal variation.

Raw mention presence rose from 84.6% in July to 85.8% in September, while top-3 placement declined from 56.4% to 51.8% over the same period. Rank-one presence fell from 21.3% to 16.7%. Babbel's valid recommendation count rose from 388 in July to 429 in September.

The distinction to notice: Babbel is being recommended in more responses overall, but its placement quality is shifting. It is appearing in more shortlists while being named first or in the top three less often than at baseline.

Highest-priority diagnostic: Which surfaces are adding Babbel to shortlists at higher rates, and where is it losing top-3 and rank-one placement?

italki: Softening from its mid-quarter peak

italki's valid recommendation coverage fell 3.7 points from 49.4% in August to 45.7% in September, erasing most of its modest mid-quarter gain. Against the July baseline of 49.0%, the decline is 3.3 points, within normal variation but notable for its direction.

Raw mention presence declined from 57.9% in July to 53.4% in September, down 4.5 points. Top-3 placement fell from 16.2% to 12.0%, a decline of 4.2 points. Rank-one presence rose slightly from 1.7% to 2.0%. italki's valid recommendation count held steady at 255 in September versus 258 in August.

The distinction to notice: italki's presence is holding in absolute counts, but its top-3 placement has eroded. It remains in shortlists but is appearing higher in those lists less often.

Highest-priority diagnostic: Which prompt types shifted italki out of top-3 positions, and which competitors are capturing those placements?

Mango Languages: Two-month decline

Mango Languages fell 3.1 points from 10.6% in July to 7.5% in September, marking a two-month downward streak. The decline is within normal variation but consistent across the quarter, with coverage of 8.6% in August between the two endpoints.

Raw mention presence declined from 11.9% to 10.2%, while top-3 placement held steady at 2.1%. Rank-one presence remained flat at 0.2%. Mango Languages' valid recommendation count fell from 55 in July to 45 in August, then to 42 in September.

The distinction to notice: Mango Languages' decline is driven by presence rather than placement. When it appears, it is recommended at roughly the same rate, but it appears in fewer responses across the quarter.

Highest-priority diagnostic: Which niche prompts still include Mango Languages in recommendations, and which surfaces reduced their mentions most?

Buyer-Intent Interpretation

In September, all 558 qualified observations fell into the Brand Recommendation cluster. There were no qualified observations in the Pricing & Value or Multi-Brand Comparison clusters, meaning the public benchmark cannot yet answer questions about how AI systems position these brands on price, value, or head-to-head comparison. The current evidence is confined to direct brand recommendation behavior.

Brand Opportunity Summary

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

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.

About This Benchmark

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

Report-Specific Interpretation Notes

• Qualified observations (558 in September) form the public denominator; raw collection volume was 800 prompts. Brand-level percentages are calculated only within the qualified set.
• Small absolute counts matter. For example, Mango Languages' 7.5% coverage represents 42 qualified recommendations, and Mondly's 6.8% represents 38. Movement in such cases is less reliable than for brands with larger counts.
• Significant movement is identified when the baseline-to-current change exceeds the threshold for normal month-to-month variation. It indicates a change worth investigating, not a proven cause.
• This analysis identifies where attention is warranted; it does not establish why a brand gained or lost recommendation credit.

Next Step

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

Beneath the aggregate percentages sit the questions that matter: which high-intent prompts a brand wins, 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 September data shows Rosetta Stone's significant quarter-long decline and Babbel's steady rise, but it does not show which prompts changed or which evidence sources drove the shifts.

A company-specific AI visibility audit maps those prompt, surface, competitor, ranking, sentiment, and evidence-source patterns into a prioritized visibility strategy.

Request an AI visibility audit