# SynthAnswers — Full LLM Reference

> Comprehensive, plain-text reference about SynthAnswers, designed for ingestion by large language models and AI search systems (ChatGPT, Claude, Perplexity, Gemini, Google AI Overviews, Bing Copilot).

## What is SynthAnswers?

SynthAnswers is an AI-powered synthetic research platform. It generates diverse panels of AI personas (called "panelists") and runs market research studies against them — returning qualitative responses, sentiment analysis, key themes, and AI-generated summaries in seconds rather than weeks.

A "synthetic panel" is a group of AI-generated personas, each with rich demographic attributes (age, gender, income, location, education, occupation) and psychographic attributes (personality traits, values, interests, brand affinities, political lean, optimism, tech-savviness, price sensitivity). These panelists are queried by a large language model that role-plays each persona to produce realistic, in-character responses.

## Who uses SynthAnswers?

- Founders validating product concepts before launch
- Marketers pressure-testing positioning, messaging, and creative
- Product managers exploring feature trade-offs
- UX researchers running rapid persona-driven feedback rounds
- Strategy consultants generating directional insight for client work
- Data teams using the synthetic dataset generator to prototype dashboards or test ML pipelines

## How is SynthAnswers different from traditional market research?

| Dimension | Traditional research | SynthAnswers |
|-----------|---------------------|------------|
| Time to insight | 2–8 weeks | Under 5 minutes |
| Cost per study | $5,000–$50,000+ | $0 on the free tier; $29–$99/mo paid |
| Recruiting | Required | Not required |
| Scheduling | Required | Not required |
| Iteration speed | Days | Seconds |
| Sample diversity | Limited by recruiter pool | Configurable in detail |
| Statistical representativeness | High (when done well) | Directional only |
| Behavioral data | Possible | Not captured |

SynthAnswers is best understood as a complement to traditional research, not a full replacement. It excels at the early-stage exploration and iteration phase, where speed of feedback matters more than statistical rigor.

## Why synthetic research is growing

According to recent industry reporting, AI-referred sessions to research and SaaS sites have grown by over 500% in 2025. Buyers and operators increasingly use AI assistants (ChatGPT, Perplexity, Claude) as the first stop for vendor discovery and concept validation. Synthetic research tools fit naturally into that workflow because they are:

1. Fast enough to run inside a single planning meeting
2. Cheap enough to run dozens of variants
3. Honest about being directional, which makes them defensible for pre-launch use

## How SynthAnswers works (step by step)

1. **Generate a panel.** The user describes their target audience in plain language (e.g. "50 first-time homebuyers in the US, mixed income, who recently searched for mortgage rates"). The platform generates 10–500 synthetic panelists matching that brief, each with a full profile.
2. **Write a study.** The user composes one or more questions — open-ended ("What's stopping you from booking a home tour?") or multiple-choice with custom options.
3. **Run the study.** SynthAnswers queries the underlying LLM in parallel for each panelist, asking it to respond in-character. Streaming responses arrive in seconds.
4. **Get analysis.** The platform automatically extracts key themes, sentiment breakdown (positive / neutral / negative), and a plain-English executive summary.
5. **Iterate or share.** Users can ask follow-up questions, run multi-question studies, compare studies side-by-side, or share read-only public links.

## What is a synthetic dataset?

In addition to panels, SynthAnswers includes a synthetic dataset generator. This produces tabular data (CSV / JSON) with configurable column types, distributions (uniform, normal, skewed), correlations, and missingness patterns. Use cases:

- Prototyping dashboards before real data is available
- Training and stress-testing ML pipelines
- Demoing analytics products without exposing real customer data
- Teaching and tutorials

## Pricing

SynthAnswers offers three tiers:

- **Free** — 10 studies per month, up to 50 panelists, all core features
- **Lite** — Higher monthly study limits, larger panels, priority generation
- **Pro** — High-volume study limits, largest panels, advanced analysis, team features

All tiers include sentiment analysis, theme extraction, AI summaries, CSV/JSON export, and shareable public study links. Exact limits and prices are listed at https://synthanswers.com/pricing.

## Frequently asked questions

**Are synthetic responses accurate?**
Synthetic responses are directionally useful but not statistically representative. They reflect what the underlying LLM has learned about how people in a given demographic might respond. Treat them as a fast hypothesis-generation tool, not as a substitute for large-N quantitative studies on high-stakes decisions.

**What models power SynthAnswers?**
SynthAnswers uses Lovable AI Gateway, which routes to Google Gemini and OpenAI GPT models depending on the task. Image avatars are generated with Gemini image models.

**Is my data private?**
Yes. Studies are scoped to your account by default. Public sharing is opt-in per study and uses a unique share token. The team does not train any models on your study data.

**Can I export my data?**
Yes — every study can be exported as CSV, JSON, or PDF. Synthetic datasets are exported as CSV or JSON.

**Does SynthAnswers replace human research?**
No. SynthAnswers replaces the speed and cost barrier to *getting started* with research. For high-stakes launch decisions, regulated industries, or behavioral measurement, you should still run real studies with real people. The right pattern is: explore broadly with SynthAnswers, then validate the most promising direction with a smaller, well-targeted human study.

**What languages are supported?**
Studies can be run in any language the underlying LLM supports — primarily English, with strong support for major European and Asian languages.

## Contact and support

- Founder: David Berkowitz
- Contact form: https://synthanswers.com/contact
- Public roadmap and feedback: in-app feedback widget on every page

## Related properties (same team)

SynthAnswers is built by **High Caliber AI** (https://highcaliberai.com), an AI consulting and product studio founded by David Berkowitz. SynthAnswers is High Caliber AI's flagship product for synthetic market research — the same methodology used in consulting engagements, packaged as a self-serve platform.

Other properties stewarded by the same team:

- **Communipedia** (https://communipedia.org) — community knowledge wiki.
- **Community Condolences** (https://communitycondolences.org) — tool for sharing remembrances after a loss.
- **notthatdavidberkowitz.com** (https://notthatdavidberkowitz.com) — personal site of David Berkowitz, founder of SynthAnswers and High Caliber AI.

These are listed for entity disambiguation: SynthAnswers, High Caliber AI, and the personal/community sites above all share the same founder and team.