# Unasked assumptions: when prompts omit company size, AI defaults to enterprise

> Across 457,174 recommendations on unsegmented buyer questions, Google AI Mode framed picks for enterprise buyers 3.48x more often than small businesses (23.3% vs 6.7%), while ChatGPT Search skewed 3.20x enterprise on GPT-5-6 and 1.91x on GPT-6.

HTML: https://parse.gl/research/ai-default-buyer-persona-enterprise-bias · Site index: https://parse.gl/llms.txt · Published: 2026-11-20

When a buyer asks a broad category question without stating company size or role, what kind of buyer do ChatGPT Search and Google AI Mode assume is asking, and how accurately do engines match explicit buyer segments?

We analyzed 1,369,684 recommended brand appearances classified by Parse's buyer-audience evaluator from October 1 through October 26, 2026, including 457,174 unsegmented recommendations across three engine-model tiers, 23 industry cuts, and a full 81-cell segment alignment matrix.

## Key figures

- **Enterprise-over-SMB bias on Google AI Mode when company size is omitted:** 3.48x (23.3% enterprise vs 6.7% small business across 226,553 recommendations)
- **Google AI Mode unsegmented picks framed for enterprise:** 23.3% (vs 6.7% small business, 4.6% mid-market, and 1.4% consumer)
- **Steepest industry skew: Manufacturing enterprise over SMB:** 10.53x (27.0% enterprise vs 2.6% small business across 2,495 recommendations)
- **Regulated and public-sector prompts answered with enterprise brands:** 35.3% (17,037 of 48,218 recommendations, nearly tying 36.1% exact matches)

## When prompts omit company size, AI recommends for enterprise up to 3.48x more than SMB

Broad buyer questions that leave out team size or budget do not receive neutral recommendations. Across 457,174 recommendations where the prompt stated no buyer segment, every tested engine-model group leaned heavily toward enterprise-positioned brands over small-business options.

Google AI Mode produced the widest gap across 226,553 unsegmented recommendations: 23.3% framed picks for enterprise buyers vs just 6.7% for small business (a 3.48x skew), 9.5% for developers or technical buyers, 4.6% for mid-market, and 1.4% for consumers. Meanwhile, ChatGPT Search assigned an explicit audience to roughly 62% of unsegmented recommendations (leaving 37.9% to 38.4% neutral vs Google's 50.6%).

**Assumed buyer persona on unsegmented prompts, by engine and model**

| Item | Value |
| --- | --- |
| Enterprise: OpenAI GPT-5-6 | 29.0% |
| Enterprise: OpenAI GPT-6 | 24.8% |
| Enterprise: Google AI Mode | 23.3% |
| Small business: OpenAI GPT-5-6 | 9.1% |
| Small business: OpenAI GPT-6 | 13.0% |
| Small business: Google AI Mode | 6.7% |
| Developer / technical: OpenAI GPT-5-6 | 11.4% |
| Developer / technical: OpenAI GPT-6 | 10.9% |
| Developer / technical: Google AI Mode | 9.5% |
| Mid-market: OpenAI GPT-5-6 | 6.4% |
| Mid-market: OpenAI GPT-6 | 6.2% |
| Mid-market: Google AI Mode | 4.6% |

_457,174 recommended brand appearances on prompts stating no audience segment (October 1 through October 26, 2026)._

> Unsegmented prompts do not produce segment-neutral shortlists: both engines assume enterprise buyers 2x to 3.5x more often than small businesses.

## ChatGPT GPT-6 cuts the enterprise-to-SMB ratio from 3.20x down to 1.91x

Model generation shifts how sharply ChatGPT leans upmarket. Under ChatGPT Search running GPT-5-6 (64,183 unsegmented recommendations), 29.0% of picks assumed an enterprise buyer vs 9.1% small business, a 3.20x enterprise-to-SMB ratio.

Under GPT-6 (166,438 unsegmented recommendations), enterprise share dropped to 24.8% while small-business share rose from 9.1% to 13.0%, narrowing the enterprise-to-SMB ratio to 1.91x. Neutral share stayed almost unchanged (38.4% on GPT-5-6 vs 37.9% on GPT-6), as did developer share (11.4% vs 10.9%) and mid-market share (6.4% vs 6.2%).

- **ChatGPT GPT-5-6 enterprise-to-SMB ratio:** 3.20x (38.4% neutral, 11.4% developer, 6.4% mid-market)
- **ChatGPT GPT-6 enterprise-to-SMB ratio:** 1.91x (37.9% neutral, 10.9% developer, 6.2% mid-market)
- **Google AI Mode persona-neutral share:** 50.6% (When Google names a persona, it skews 3.48x enterprise)

> GPT-6 lifted unsegmented SMB representation by 3.9 percentage points, cutting ChatGPT's upmarket tilt nearly in half relative to Google AI Mode.

## All 23 industries tilt toward enterprise, led by Manufacturing at 10.53x

Across all 23 industry groups tested on unsegmented prompts, not a single category favored small business over enterprise. The steepest imbalance appeared in Manufacturing (27.0% enterprise vs 2.6% small business, a 10.53x ratio across 2,495 recommendations), Transportation (31.3% vs 4.9%, 6.33x), Data and Analytics (23.7% vs 3.8%, 6.19x), and Hardware (24.7% vs 4.8%, 5.12x).

Even the largest B2B SaaS verticals skewed roughly 2-to-1 toward enterprise: Software (15,405 recommendations) landed at 22.3% enterprise vs 11.2% SMB (1.99x) and Information Technology (11,118 recommendations) landed at 26.0% vs 13.1% (1.99x). Only Content and Publishing (1.22x), Apps (1.23x), and Media and Entertainment (1.26x) came close to parity.

**Enterprise vs small-business bias across 23 industry groups**

| Industry | Recommendations | Enterprise (%) | SMB (%) | Ent/SMB ratio | Developer (%) | Neutral (%) |
| --- | --- | --- | --- | --- | --- | --- |
| Manufacturing | 2495 | 27 | 2.6 | 10.53 | 2.5 | 63.8 |
| Transportation | 3461 | 31.3 | 4.9 | 6.33 | 4.6 | 49.8 |
| Data and Analytics | 6163 | 23.7 | 3.8 | 6.19 | 24.9 | 39.4 |
| Hardware | 2505 | 24.7 | 4.8 | 5.12 | 17.8 | 39 |
| Education | 3431 | 27.8 | 6.4 | 4.36 | 4.3 | 50.2 |
| Agriculture and Farming | 913 | 30.8 | 7.2 | 4.26 | 3 | 43.3 |
| Artificial Intelligence | 3986 | 29.5 | 8 | 3.7 | 12.6 | 41 |
| Professional Services | 3839 | 21.7 | 6.3 | 3.45 | 8.2 | 54 |
| Collaboration | 1242 | 14.1 | 4.6 | 3.07 | 13.1 | 61.2 |
| Internet Services | 2724 | 17.6 | 7.4 | 2.37 | 8.2 | 56.1 |
| Real Estate | 1451 | 20 | 8.8 | 2.28 | 1.6 | 45.3 |
| Sales and Marketing | 6249 | 17.2 | 8.1 | 2.13 | 3.1 | 55.9 |
| Information Technology | 11118 | 26 | 13.1 | 1.99 | 14.4 | 37.6 |
| Software | 15405 | 22.3 | 11.2 | 1.99 | 14.1 | 35.5 |
| Health Care | 1926 | 24.7 | 12.7 | 1.94 | 2.2 | 45.3 |
| Design | 825 | 12.4 | 6.4 | 1.92 | 5.6 | 56.8 |
| Financial Services | 2814 | 20.8 | 11.7 | 1.79 | 6.1 | 47.2 |
| Administrative Services | 2057 | 19.8 | 11.2 | 1.77 | 11.6 | 45 |
| Advertising | 832 | 19.7 | 12.9 | 1.53 | 3.2 | 45.9 |
| Commerce and Shopping | 2446 | 21.7 | 14.9 | 1.46 | 6.2 | 41.3 |

Showing 20 of 23 rows (All 23 industries). Full CSV: /research/ai-default-buyer-persona-enterprise-bias/dataset.csv

_Unsegmented buyer prompts (October 1 through October 26, 2026). Sort by ratio, enterprise %, or recommendation volume, or open Chart view._

> Every single one of the 23 industries skewed toward enterprise on unsegmented prompts, with physical operations and analytics leading the gap.

## Data and Apps assume developer buyers, while Manufacturing stays 63.8% neutral

Secondary default personas reveal where each industry shifts when it does not pick enterprise. In Data and Analytics, developer or technical framing actually beat enterprise outright at 24.9% (vs 23.7% enterprise and just 3.8% SMB). Apps (18.6%), Hardware (17.8%), Information Technology (14.4%), and Software (14.1%) also routed heavily toward technical buyers on unsegmented questions.

Meanwhile, consumer-adjacent real-estate and media prompts broke differently: Real Estate assigned 12.3% of unsegmented recommendations to consumers and Content and Publishing assigned 10.5% to consumers, compared to under 1% in Software (0.9%) and Artificial Intelligence (0.7%).

**Top industries by developer-buyer share on unsegmented prompts**

| Item | Value |
| --- | --- |
| Data and Analytics | 24.9% |
| Apps | 18.6% |
| Hardware | 17.8% |
| Information Technology | 14.4% |
| Software | 14.1% |
| Content and Publishing | 13.7% |
| Collaboration | 13.1% |
| Artificial Intelligence | 12.6% |

_Share of unsegmented recommendations targeting developer or technical buyers, October 1 through October 26, 2026._

> Technical categories split unsegmented defaults between enterprise and developers, leaving small business with 3.8% to 15.9% of placements.

## Consumer prompts match 97.7%, while public-sector buyers get enterprise picks 35.3% of the time

Even when buyers spell out their segment in the prompt, AI engines vary widely in honoring it. Across 344,368 recommendations for explicit consumer prompts, 97.7% matched consumer-positioned picks. Explicit developer prompts matched 73.5% of the time (242,934 recommendations), and explicit enterprise prompts matched 72.3% (88,614 of 122,639).

Alignment collapsed on specialized and smaller commercial segments. Mid-market prompts matched mid-market positioning 57.8% of the time (17,150 recommendations), spilling 14.9% to enterprise. Small-business prompts matched 60.4% (88,267 recommendations), spilling 20.1% to neutral and 4.3% to enterprise. Worse, regulated and public-sector prompts matched their own segment only 36.1% of the time (17,407 of 48,218), handing almost as many picks directly to enterprise brands (17,037, or 35.3%).

**Explicit prompt segment match rate vs enterprise and neutral spillover**

| Stated buyer segment | Recommendations | Exact match (%) | Enterprise (%) | SMB (%) | Neutral (%) |
| --- | --- | --- | --- | --- | --- |
| Consumer | 344368 | 97.7 | 0.1 | 0.2 | 1.2 |
| Developer / technical | 242934 | 73.5 | 13.8 | 1.4 | 10.2 |
| Enterprise | 122639 | 72.3 | 72.3 | 0.7 | 14 |
| Agency | 8869 | 61.6 | 10.3 | 4.6 | 12.1 |
| Small business | 88267 | 60.4 | 4.3 | 60.4 | 20.1 |
| Mid-market | 17150 | 57.8 | 14.9 | 6.2 | 16.4 |
| Solo / freelancer | 40043 | 43.4 | 2.7 | 7.4 | 27.4 |
| Regulated / public sector | 48218 | 36.1 | 35.3 | 3.2 | 16.1 |

_Explicitly segmented recommendations across 8 buyer segments (October 1 through October 26, 2026)._

> Enterprise positioning captures 14.9% of mid-market, 13.8% of developer, 10.3% of agency, and 35.3% of public-sector recommendations.

## Where AI sends regulated buyers and solo founders when it misses their segment

The two weakest-matched buyer prompts leak into completely opposite directions. When buyers ask for regulated or public-sector tools (48,218 recommendations), AI engines treat generic enterprise software as an almost equal substitute: 36.1% target regulated or public sector, 35.3% target enterprise, 16.1% stay neutral, and 6.3% target developers.

When solo practitioners or freelancers ask for tools (40,043 recommendations), engines drift downmarket instead: only 43.4% match solo or freelancer positioning, while 27.4% stay neutral, 11.0% spill into consumer tools (4,399 picks), and 7.4% pull from small-business picks (2,947).

**Recommendation destination shares on public-sector vs solo-freelancer prompts**

| Item | Value |
| --- | --- |
| Exact segment match: Regulated / public sector | 36.1% (17,407) |
| Exact segment match: Solo / freelancer | 43.4% (17,397) |
| Enterprise spillover: Regulated / public sector | 35.3% (17,037) |
| Enterprise spillover: Solo / freelancer | 2.7% (1,071) |
| Persona-neutral: Regulated / public sector | 16.1% (7,786) |
| Persona-neutral: Solo / freelancer | 27.4% (10,969) |
| Consumer spillover: Regulated / public sector | 0.2% (108) |
| Consumer spillover: Solo / freelancer | 11.0% (4,399) |
| Small business spillover: Regulated / public sector | 3.2% (1,551) |
| Small business spillover: Solo / freelancer | 7.4% (2,947) |

_48,218 regulated/public-sector recommendations vs 40,043 solo/freelancer recommendations (October 1 through October 26, 2026)._

> Regulated-sector prompts get absorbed by enterprise vendors (35.3%), while freelancer prompts leak into consumer tools (11.0%).

## Explore the full 81-cell buyer segment alignment matrix

Across all 1,369,684 observations spanning all nine stated prompt segments and all nine recommended brand audiences, asymmetry runs one direction: enterprise spills upward into mid-market (14.9%), developer (13.8%), agency (10.3%), and public-sector (35.3%) prompts, whereas small-business picks win just 0.7% of enterprise prompts, 1.4% of developer prompts, and 6.2% of mid-market prompts.

Use the segment selector below to inspect where recommendations land for every stated buyer segment.

**9x9 stated prompt segment vs recommended brand audience (1,369,684 observations)**

| Prompt segment | Recommended brand persona | Recommendations | Share (%) |
| --- | --- | --- | --- |
| Agency | Agency | 5467 | 61.6 |
| Agency | No segment stated | 1077 | 12.1 |
| Agency | Enterprise | 914 | 10.3 |
| Agency | Developer / technical | 490 | 5.5 |
| Agency | Small business | 410 | 4.6 |
| Agency | Mid-market | 329 | 3.7 |
| Agency | Solo / freelancer | 81 | 0.9 |
| Agency | Regulated / public sector | 63 | 0.7 |
| Agency | Consumer | 38 | 0.4 |
| Consumer | Consumer | 336460 | 97.7 |
| Consumer | No segment stated | 4122 | 1.2 |
| Consumer | Developer / technical | 1662 | 0.5 |
| Consumer | Solo / freelancer | 620 | 0.2 |
| Consumer | Small business | 532 | 0.2 |
| Consumer | Regulated / public sector | 494 | 0.1 |
| Consumer | Enterprise | 432 | 0.1 |
| Consumer | Mid-market | 26 | 0 |
| Consumer | Agency | 20 | 0 |
| Developer / technical | Developer / technical | 178545 | 73.5 |
| Developer / technical | Enterprise | 33544 | 13.8 |

Showing 20 of 81 rows (Destination breakdown). Full CSV: /research/ai-default-buyer-persona-enterprise-bias/dataset.csv

_Combined ChatGPT Search and Google AI Mode recommendations (October 1 through October 26, 2026). Pick a stated prompt segment to inspect all 9 destinations._

> Persona drift is asymmetric: enterprise brands bleed into every B2B segment, while SMB positioning stays locked inside explicit small-business prompts.

## What to do with these numbers

Audit your tracked prompts separately for unsegmented category queries and explicitly segmented buyer prompts. If you sell to small businesses, mid-market teams, or solo founders, unsegmented category rankings over-index on enterprise incumbents by 2x to 10.5x and will undercount your true visibility on qualified prompts.

At the same time, if you sell into mid-market, regulated, or developer segments, publish dedicated pages with explicit segment proof (pricing caps, team sizes, compliance frameworks, deployment footprints) so engines stop collapsing those queries into generic enterprise vendors. Compare where ChatGPT GPT-6 and Google AI Mode diverge across your exact buyer personas.

**Buyer-persona positioning actions by target segment**

| Target buyer | Observed AI default | Positioning action |
| --- | --- | --- |
| 1. Small business (SMB) | Only 6.7% on unsegmented Google AI Mode vs 23.3% enterprise (60.4% exact match when stated) | Track explicit SMB-qualified prompts separately and publish transparent entry-tier pricing and quick setup pages |
| 2. Mid-market & agency | 14.9% of mid-market and 10.3% of agency prompts leak to enterprise brands | Publish mid-market and multi-client agency comparison pages contrasting total cost and complexity vs enterprise suites |
| 3. Regulated / public sector | 35.3% of recommendations leak to enterprise brands (nearly tying 36.1% exact match) | Pair category terms directly with specific compliance attestations and public-sector case studies |
| 4. Enterprise | 72.3% exact match plus 13.8% to 35.3% spillover across developer and public-sector prompts | Defend unsegmented category prompts on both ChatGPT GPT-6 (24.8%) and Google AI Mode (23.3%) |

_Grounded in 1,369,684 segment-aligned recommendations (October 1 through October 26, 2026)._

> Never evaluate SMB, mid-market, or regulated AI visibility from unsegmented prompts alone; monitor explicit segment prompts and publish segment-specific evidence.

Related: [Compare competitors by buyer segment in Parse](https://parse.gl/use-cases/compare-competitors) · [Inspect persona splits directly via Parse MCP](https://parse.gl/mcp)

## Study scope and counting rules

The full persona-alignment matrix covers 1,369,684 recommended brand appearances scored from October 1 through October 26, 2026 under one uniform Parse evaluator version across ChatGPT Search (gpt-5-6 and gpt-6) and Google AI Mode. Each observation pairs the prompt's stated audience label with the recommended brand's target audience label assigned inside that answer.

The headline unsegmented engine comparison covers 457,174 recommendations where the prompt stated no buyer segment across engine-model buckets with at least 500 recommendations (64,183 ChatGPT Search gpt-5-6; 166,438 ChatGPT Search gpt-6; 226,553 Google AI Mode). The 9x9 matrix totals 457,196 unsegmented recommendations because it retains the 22 unsegmented recommendations from minor model strings, and covers 122,639 enterprise-prompt observations across the nine core audience labels.

Industry enterprise-to-SMB ratios cover all 23 industry groups present in unsegmented prompts with sufficient volume. Segment classifications measure how AI engines frame each brand in context rather than independent verification of a product's pricing tiers or customer roster.

## Method

We analyzed 1,369,684 recommended brand appearances paired with prompt-level and brand-level buyer audience classifications across ChatGPT Search (gpt-5-6 and gpt-6) and Google AI Mode from October 1 through October 26, 2026 under one uniform Parse evaluator version, including 457,174 unsegmented recommendations in 500+ observation engine groups, and 23 industry groups.

## Data

- [Full study dataset (CSV)](https://parse.gl/research/ai-default-buyer-persona-enterprise-bias/dataset.csv): Unsegmented engine skew, industry enterprise-to-SMB ratios, and explicit buyer-segment alignment.

## For agents

- [Study dataset (CSV)](https://parse.gl/research/ai-default-buyer-persona-enterprise-bias/dataset.csv)
- Any brand's shares and rivals, with 95% ranges: `https://agent.parse.gl/v1/tools/check_brand?subject=%3Cbrand-domain%3E`
- A brand's shares by topic, with 95% ranges: `https://agent.parse.gl/v1/tools/query_metrics?args=%7B%22measures%22%3A%5B%22share_recommended%22%2C%22share_named%22%5D%2C%22subject%22%3A%22%3Cbrd_id%3E%22%2C%22by%22%3A%5B%22topic%22%5D%7D`
- Search the index: `https://parse.gl/api/public/search?q=<words>`

Reuse: Figures and tables on this page are free to republish with attribution: a link to https://parse.gl/research/ai-default-buyer-persona-enterprise-bias beside them. A standard followed link (dofollow) is preferred.

- No account: Public figures and up to 5 examples per list, about 120 brands a day per address.
- Free account: $5 in credit, your brand's first check (20 buyer questions) and a free weekly check.
- After your first top-up: Exports and deeper search results. The first top-up starts at $5.
- Live checks: $0.01 per question per engine ($0.05 on Gemini and Claude).
- Full public answers and pages: 25 free per day, then $0.005 each.
- Public answers older than 90 days: $0.001 each the first time your workspace counts or lists one, then yours.
- An agent can save work before signup: `POST https://agent.parse.gl/v1/workspaces`
- Connect an assistant: `https://agent.parse.gl/mcp`

[Prices](https://parse.gl/pricing.md) · [Map of the site](https://parse.gl/llms.txt)
