A WORKING PROJECT / BUILD LOG + WAITLIST
Journoa weighs a B2B team's funnel evidence to identify the problem that deserves attention first. It recommends one play and shows the reasoning before the team acts.
Build notes only when there is something worth showing. Unsubscribe anytime.
CURRENT BUILD
This console tracks what has been implemented, what changed, and the tests behind each milestone.
LAST UPDATED JULY 21, 2026
FOLLOW THE BUILD
Updates go out when there is a meaningful change to inspect.
Confirmation by email. Unsubscribe anytime.
FROM INPUT TO NEXT MOVE
Seven stages run behind each recommendation. Journoa condenses their decisions into a short trace: what was observed, where the largest loss occurred, what to try next, and how confident the result is.
62 accounts were created. 14 reached first value.
The largest measured loss is between account creation and first value.
Shorten the path to first value and contact accounts that stall.
Directional. The sample is still small.
Missing evidence stays unknown. Journoa does not fill the gap with a confident answer.
HOW IT WORKS
THE PULSING DOT MARKS THE LARGEST MEASURED DROP IN THIS EXAMPLE.
Each observation keeps its unit, source, timestamp, and sample size. Missing fields stay missing.
Funnel measurements become signals and condition checks. When evidence is missing, the result can remain unknown instead of being recorded as false.
The picker removes plays that do not fit the workspace, ranks the eligible options, and records why the final play was selected. The wider methodology defines 33 play specifications.
The engine currently accepts validated workspace input in code. Direct CSV, analytics, CRM, messaging, and email connections are still planned.
HOW JOURNOA WILL FIT
The engine code stays proprietary, but access will be free. Every surface will call the same diagnosis logic, so the recommendation does not change because you entered through a different tool.
Send structured funnel context and receive the diagnosis, recommendation, and trace.
Use the same engine from Claude and other tools that support MCP.
Bring packaged Journoa workflows into the coding tools you already use.
Use Journoa directly when you do not want to connect another tool.
JOURNOA'S ROLE
Turns structured funnel context into a diagnosis.
Identifies the largest problem supported by the available evidence.
Selects one play that fits the company context and prior attempts.
Returns the recommendation with its confidence, missing inputs, and trace.
YOUR EXISTING STACK
CRM, analytics tools, and spreadsheets continue to hold the underlying data.
Product and funnel events continue to be collected by the tools already in place.
Your team reviews the recommendation and decides whether to use it.
Campaigns, messages, and spend change only after approval.
RESEARCH AND LEARNING
Journoa's methodology currently includes 13 benchmarks graded by source quality. The engine can already read supplied outcome records and calculate win rates.
Capturing those outcomes and building a consented learning pool are not implemented yet. Journoa does not currently use one company's data to inform another.
The longer term aim is a community informed engine for better B2B growth decisions. Contributed cases and consented outcomes can show where the method holds, where it fails, and what needs to change.
13 benchmarks, each graded by source quality.
Win rate reads from supplied outcome records.
Any future shared learning pool will require consent and exclude internal records.
A NOTE FROM THE BUILDER
Claude Code and Codex help me write and review it. I use published research to shape the method, then frozen specs, typed contracts, tests, and a separate verification pass to decide when a piece of work is actually done.
Jai, building in the open. The waitlist gets the build notes first :)
QUESTIONS AND ANSWERS
Journoa is a B2B funnel diagnosis engine. It identifies the largest problem the available evidence supports, records how it reached that conclusion, and recommends one play to examine next. The first version focuses on person level funnel diagnosis.
It is for B2B founders, growth marketers, and SaaS operators who have funnel data but still need to decide what deserves attention next.
An assessment triggers a set of candidate plays. The engine removes any that do not fit the workspace, ranks the rest, and records why the final play was selected. The active pack stays deliberately small while that logic is being verified.
Access to the engine is planned to be free. The API will be the first surface, followed by a remote MCP server, plugins for Claude Code and Codex, and a hosted workspace for people who do not want to connect another tool.
No. The engine code and methodology remain proprietary. Access will be free so more B2B teams can use the engine and contribute cases and consented outcome data that help test its decisions.
No. Journoa recommends a play, but it does not launch campaigns or send messages. Any future execution feature will require approval before it runs.
Yes. Journoa is meant to improve through real B2B situations, not a closed set of assumptions. Edge cases, contextual conditions, and well documented outcomes can all help test the engine. Shared learning will use only consented records. You can send them to connect@jai-k.com.
Jai uses Claude Code and Codex to help write and review the code. Published research shapes the method. Frozen specs, typed contracts, tests, and an independent verification pass decide whether a work package is complete.
