Sol logo
Trust & decision integrity

Trust & decision integrity for Sol Helps.

Sol Helps captures real customer questions at the moment users get stuck, answers using your documentation as the primary source, and preserves that raw confusion long enough for teams to reason clearly — without tagging, pipelines, or premature conclusions.

See the clarity report Try with your docs
Consent-first
No storage without opt-in
Docs-led
Answers stay grounded
Decision support
Insight without overconfidence

What Sol Helps is

A decision-support layer that shows where users get stuck — using real questions as evidence.

Informational support for docs, product, or website
Patterns of confusion from real questions (with examples)
Built for founders, product teams, and docs/support leads
Safe to deploy on public-facing surfaces

What Sol Helps is not

Clear boundaries reduce risk and increase trust.

Not a general-purpose chatbot
Not a ticketing system
Not a replacement for GA or event analytics tooling
Not an autonomous agent that takes actions
Not a system that claims certainty when the data is thin

Trust is earned through traceability — not bold claims

Sol Helps is designed so a team can point to evidence, understand context, and decide what to change — without being pushed toward premature certainty.

Evidence-first

If you can’t point to examples, you’re guessing.

Insights link back to real question snippets
Context is preserved (step/page/topic where possible)
Teams can judge the signal for themselves

Truth sequencing

Clarity appears progressively — not instantly.

Early states can be quiet (this is expected)
Patterns emerge as questions accumulate
The product avoids forcing conclusions

Decision integrity

Designed to help teams think more clearly.

Preserves ambiguity where it matters
Avoids outcome guarantees and overconfident language
Supports prioritisation without pretending certainty

Documentation-led answers with consent-first handling

A high-level view of how meaning is preserved — from your documentation, through consented conversations, into patterns a team can reason about.

Content sources• URL crawling (HTML)• File uploads• Static docs sitesPrivate knowledge base• Vector store• Docs are primary source• Guardrails against guessingAssistant answers• Informational only• On-brand tone & constraintsClarity report• Themes, intent, friction signals• Evidence + context you can traceconsent gate
Consent gating applies to conversation storage and downstream interpretation. Without opt-in, the assistant can still run — but chat logs are not retained for analysis, and Sol Helps won’t pretend insight exists.
Content in
Upload files or crawl URLs — content becomes a private knowledge base for grounded answers.
Consent gating
Conversation storage only happens after end-user opt-in.
Clarity report inputs
Themes, intent, and friction signals — derived from consented questions over time.

Supported content sources

Common ways teams bring documentation into Sol Helps.

URL crawling
Crawl pages and convert them into HTML for ingestion (static docs, hosted docs, or exported docs).
Direct file uploads
Upload PDFs, Word docs, Markdown, HTML, and more.
Static documentation sites
Perfect for help sites that are already shipped as HTML outputs.

Supported file types

Common formats are visible by default.

.pdfPDF
.docWord (.doc)
.docxWord (.docx)
.mdMarkdown (.md)
.htmlHTML
.txtPlain text (.txt)

The widget works anywhere JavaScript can execute — static sites, frameworks, or docs platforms.

Common trust questions

Written for product and docs/support leads — with detail where it matters.

Are customer conversations used to train public models?

No. Customer conversations are never used to train public models. Conversation storage is consent-first: without opt-in, chat logs are not retained for analysis.

What happens if the assistant doesn’t know the answer?

Sol Helps is designed so your documentation is the primary source of truth. If an answer isn’t supported by the provided content, the assistant should say so rather than guessing.

What data is used to form patterns and themes?

Sol Helps derives patterns from chat content (with consent), plus minimal context like the page URL/referrer and language/locale — to help teams understand where confusion happens and what to improve next.

Can Sol Helps take actions inside my product?

Not today. Sol Helps is strictly informational. It answers questions and surfaces patterns of confusion — it does not perform actions or make changes inside your product.

When will this become decision-relevant?

You can set up an assistant in minutes, but decision-relevant patterns emerge as real questions accumulate. Early signal may be sparse — that’s expected — and it’s safer than generating false clarity.

How quickly can we validate this internally?

The lowest-effort path is: sign up → paste a URL → select pages → embed the script. Time to a working assistant is typically 5–10 minutes.

Start small. See what this changes in how you decide.

If you already have documentation or onboarding content, you can get a working assistant in minutes. From there, Sol Helps turns real questions into a clearer picture of confusion — with evidence and context — so you can prioritise what to fix next without committing to a platform change.

Want deeper detail first? How it works · Use cases · Contact

See the clarity report Try with your docs