Checking project access…

Flovia Dashboard - Overview

Skip to market content

Northwind (Demo)

Measurement period

Workspace

Discovery & Routing

Discovery Findings Product Results Test Questions

Agent Experience

AX Findings AX Evidence

Overview

Understand priority issues and competitive position.

Loading results

Skip to market content

Northwind (Demo)

Measurement period

Workspace

Discovery & Routing

Discovery Findings Product Results Test Questions

Agent Experience

AX Findings AX Evidence

Overview

Prioritize measured product actions, then inspect the evidence behind each decision.

Export

Export this page as Markdown

Decision workspace

Next actions

Choose an action to inspect its measured impact and supporting evidence.

Review findings

6 actions

ActionImpactBased on

ImpactDiscovery & routing

Selection · Exploratory listed-choice test

8%83%+75 pp

Before1/12

After10/12

Recommended action

Test the unbranded Transaction Submission API entry path

Current docs already explain multi-path routing; test whether an unbranded task entry reaches that explanation.

Review supporting evidence

Competitive analysis

Where Northwind leads—and where competitors are chosen

Compare results by task and provider choices by model.

Open Prompt Evidence

How does northwind compare by purpose?

Purpose0%50%100%Leader / target

Real-time streaming57%

Agent experience

What changes after a product is selected?

Message tests and documentation delivery show where routing can still change.

Open Agent Experience

How does routing differ before and after product-description tests?

Product description testRouting selectedChange

Transaction submission+75 pp

28%

Leader: northwind 57%

Historical analytics60%15%18%Leader: northwind 60%

Production RPC65%32%Leader: northwind 65%

Transaction delivery53%25%18%Leader: northwind 53%

Wallet and portfolio asset data86%Leader: northwind 86%

northwindTop competitorNext competitorOther providers

Share of valid runs; no-selection outcomes stay in the denominator · Search OFF · 72 runs per purpose

Which providers does each model choose?

ProviderTerra Sonnet 5 Gemini 3.6 Flash Grok 4.3

northwind83%76%57%41%northwind: Terra 75 of 90 runs, Sonnet 5 68 of 90 runs, Gemini 3.6 Flash 51 of 90 runs, Grok 4.3 37 of 90 runs.

Competitor A0%0%17%39%Competitor A: Terra 0 of 90 runs, Sonnet 5 0 of 90 runs, Gemini 3.6 Flash 15 of 90 runs, Grok 4.3 35 of 90 runs.

Competitor B1%1%10%2%Competitor B: Terra 1 of 90 runs, Sonnet 5 1 of 90 runs, Gemini 3.6 Flash 9 of 90 runs, Grok 4.3 2 of 90 runs.

Competitor E Crypto0%12%0%0%Competitor E Crypto: Terra 0 of 90 runs, Sonnet 5 11 of 90 runs, Gemini 3.6 Flash 0 of 90 runs, Grok 4.3 0 of 90 runs.

Competitor F0%0%0%8%Competitor F: Terra 0 of 90 runs, Sonnet 5 0 of 90 runs, Gemini 3.6 Flash 0 of 90 runs, Grok 4.3 7 of 90 runs.

Others16%11%

No selection0%0%0%0%No selection: Terra 0 of 90 runs, Sonnet 5 0 of 90 runs, Gemini 3.6 Flash 0 of 90 runs, Grok 4.3 0 of 90 runs.

Search OFF · 90 runs per model.

Real-time streaming+25 pp

Address history+25 pp

BeforeAfter

How does target-API selection vary by documentation setup?

Documentation setupTarget API selected

Nothing provided29%
Docs only, unannounced25%
Index in AGENTS.md45%
Index only, no docs100%
Index + docs100%
Docs + auto-loaded index with the address history API entry removed100%
No local docs; public web docs reachable (llms.txt)100%

17%

10%

Others: Terra 14 of 90 runs, Sonnet 5 10 of 90 runs, Gemini 3.6 Flash 15 of 90 runs, Grok 4.3 9 of 90 runs.