BETA
Join the waitlist Plan · asks firstPull 142 replays, 38 ticketsWalk the mobile checkoutRank hypothesesApprove
Research
Ask a question in a sentence. Get a cited brief back, or a moderated study on any topic.
Fix what leaks. Ship what earns.AI finds, designs, tests and ships on one visual canvas your whole team works in.
Fix the problems that cost you revenue and build the new experiences that earn it, the way your best team would: evidence, hypothesis, test, proof. Run by AI, directed by you, on one canvas.
Scout AI
+++
Token Left: 500KThinkingThinkingThinkingThoughtkeyboard_arrow_down
Send a followup
+
Skills
SkillsMonitor
+++
!−6.2% at addressResearch
+++3of 4 stuck
Browser
+++2issues found
Analyze
+++PapprovedDesign
+++
Rpicked BTest
refresh
arrow_forward
Code
+++
SmergedDecision
+++Token Left: 500KScout · just now
Decision record #0412 is ready. Variant B won by 6.7 points.
91.2%address step,
back to normal
back to normal
I’ll keep watching this fix.
P
Priya · now
Turn it into a slide for leadership?
Turn it into a slide for leadership?
Send a followup
+Skills
SkillsSlide
+++↗ready to share
Preview
+++✓live
Reading your prompt and the goal behind itThe problem
AI made shipping fast. Knowing what to ship, and proving it paid off, is still slow.
Problems it solves
Pick the number you need to move, or the experience you need to build.
ActivationMore new users reach their first winactivation rate · first order · D1
EngagementUsers come back, and back againweekly active · repeat purchase · DAU/MAU
RetentionFewer leave, and you know whyday-30 retention · renewals · churn
ConversionEvery funnel step watched and fixedsignup, cart or search funnels
Cross-sell & upsellMore revenue from the customers you haveexpansion ARR · AOV · attach rate
Trial to paidPricing, packaging and paywalls that convertpaid conversion · subscription starts
New experiencesShip a new flow, feature or product line, with proofnew ARR · GMV · revenue per visitor
ReactivationWin back the users who leftrecovered MRR · reorders · reactivated users
See it run
One goal, from a sentence to a shipped result.
Scout walks your product, finds the leak, designs the fix in your system, builds it in your code and proves it worked. Your team approves every step.
01
The philosophy
The way your best team does it. Run by AI.
Most AI tools take a prompt and give you output. Your best team takes a problem, or an idea for a new experience, and gives you proof: evidence, a hypothesis, a test, a result, written down. Scopeout is the second, run by the first.
01
Teamsthe method, slowly
02
AI todayoutput, no proof
03
Scopeoutthe method, run by AI
6 WEEKS · 5 HANDOFFS
W1Analytics
W2Research
W3Figma
W4Jira
W5Code
?
Did it work?Nobody measured after launch
How teams did it.
The right method, five tools and six weeks. Nobody checked if it worked.
PROMPT IN · OUTPUT OUT
›Fix the checkout drop
Generated screen
EvidenceTestRecord
How AI does it today.
Prompt in, output out. No evidence, no test, no memory of what happened.
THE METHOD · RUN BY AI
✓Evidence
✓Hypothesis
✓Test
✓Proof
✓Record
↑
+6.7 pts at addressShipped · decision 0412
How Scopeout does it.
Your team’s method, run by AI: evidence, hypothesis, test, proof, record. Fix what leaks, ship what earns. You direct.
02
The same canvas, eight buckets
Everything you need, on one canvas.
Each bucket is a real area of the canvas, not a separate tool. Open one to see it work.
Experience map · mobile checkoutLand100%Sign up82%Payment41%Done38%
Address46%
Analyze
Turn replays and tickets into experience maps and audits.
Design
Generate screens on your design system. Edit them inline.
Variant B → pull request #48212 tests passedPreview readyBehind a flagMerge
Build
Take the chosen variant to a pull request, behind a flag.
A / B · day 6 of 14A · 3.9%B · 5.8%97% significance
Test
A/B/n with guardrails, plus AI-moderated user interviews. Winners promote once they clear your bar.
Staged rollout · variant BGuardrailsholding · 3 / 3
10%Mon
25%Tue
50%today
100%needs approval
Publish
Roll winners out in stages, with one-click rollback.
Address step · conversion vs baseline bandbaseline bandrelease v4.12−6.2% · outside band
Monitor
Watch every funnel step and release against its baseline.
Run #118 · recordProblemaddress step −6.2%Evidence142 replays · 38 ticketsDecisionship variant BApproverpriyaOutcome+3.4% · promoted
Improve
Keep the record: problem, evidence, decision, outcome.
03
Start from a workflow
Start from a workflow.
Every run starts with a sentence. Pick one and it opens on your canvas, already wired.
Find why a funnel step drops
“A step in our funnel is losing people. Find out why and fix it.”
Funnel · replays→Monitor→Research→Analyze→Fix shipped
Try it
Get new users to their first win
“New users stall before they get value. Find the blocker and test a fix.”
Replays · tickets→Journey map→Pain points→Variant→A/B test
Try it
Turn trials into paying customers
“Find the moment users are ready to pay and test the upgrade there.”
Funnel→Research→Design→A/B test→Ship
Try it
Stop users from leaving
“Retention is slipping. Rank the reasons people leave and fix the top one.”
Cohorts→Recordings→Causes→Fix→Test
Try it
Bring users back more often
“Find what makes people come back, and give them more of it.”
Usage→Segments→Hypothesis→Nudge→Test
Try it
Grow revenue per customer
“Find where customers are ready to buy more and test one offer.”
Orders→Moments→Offer→A/B test→Ship
Try it
Win back inactive users
“Find the users who went quiet, why they left, and what brings them back.”
Inactive users→Last action→Reason→Win-back→Test
Try it
Build what customers ask for
“Turn the most requested feature into a shipped, tested experience.”
Requests→Design→Code→Flag→Test→PR
Try it
04
What's next
Give it a goal. Watch the number move.
One sentence sets the revenue goal and the guardrails. The canvas finds the leaks, writes the hypotheses, runs the tests, ships the winners, and learns your product as it goes. You approve the calls and watch the number move.
THIS QUARTER
from 5 problems Scout found and fixed
Goal $120K68%
Week 6 of 13 · guardrail: refunds ≤ 2%
PROBLEM SCOUT FOUNDWHAT SHIPPEDRESULTREVENUE
Mobile checkoutAddress step −6.2% after 4.12.0
Autofill fix + saved addresses
+6.7 pts
+$22KShipped
Trial to paidPaywall shown before the first win
Paywall moved after first report
+3.1 pts
+$19KShipped
CheckoutPromo code errors up 3×
Fixed code validation on mobile
−71% errors
+$15KShipped
CartExpress pay hidden below the fold
Apple Pay above the fold
+2.8%
+$14KShipped
Delivery“Can’t change delivery date” tickets ×2
Self-serve date change
−48% tickets
+$11KShipped
SignupAndroid stalls at email verify
Magic-link login, testing
day 4 of 7
—Testing
Detect
Watches every funnel around the clock and flags the leak before you ask.
Hypothesize
Turns replays, tickets and data into one cause, with how sure it is.
Test
Runs the fix against today, inside your guardrails, until it’s significant.
Ship & learn
Ships the winner once you approve, and remembers what worked for next time.
05
FAQ
Questions, answered.
What does Scopeout do that my analytics, Figma and an AI coding tool don't?
Those tools each cover one step. Scopeout runs the loop between them: it notices the drop, finds out why with cited evidence, draws the fix on your design system, ships it behind a flag, tests it, promotes the winner and writes down what happened. You direct; it runs; the record stays.
What do I need to get started?
Anything. An idea in a sentence, a URL, a screenshot, a Figma file, a repo or a pile of tickets: start with whatever you have and Scout picks it up from there. Connect analytics, your repo, Figma and Slack when you want the full loop; each connector adds a step it can run.
How fast do I get something useful?
A cited research brief in under an hour on day one. A first variant behind a flag in the first week, once the repo and design system are connected. The weekly digest starts the first Monday.
What runs on its own, and what waits for a person?
Every step has a dial. Out of the box: research and analysis run after you approve the plan; design proposes variants for you to pick; build opens pull requests your engineers review; test promotes only past the bar you set; publish waits for a named approver. Turn any of these up as trust grows.
Can it ship something bad to production?
No change reaches users without a flag, a guardrail and a person. Tests stop on a guardrail breach (refunds, errors, support volume), rollouts go out in stages, and any change rolls back in one click. Nothing merges on its own.
Where does the code go, and who reviews it?
Into your repo, as a normal pull request with tests and a preview link, written in the stack and conventions the repo already uses. Your engineers review and merge it like any teammate's PR.
How does it use our design system?
Connect Figma or Storybook and Design generates on your components and tokens, sized to real content, so what the team reviews is what ships. Edit inline or send it to Figma and back.
How does it decide what to work on?
From the goal you give it and from your product model: what your users do by segment and what has worked for whom. It proposes loops ranked by expected lift with the evidence attached; you approve, reorder or drop them.
What is "the record", and who can see it?
Every run writes problem, evidence, decision, approver and outcome as a block on the canvas, linked to the replays, tickets and PRs behind it. Anyone on the canvas can read it; roles decide who can approve. It exports to Notion, Confluence or PDF.
Is our data used to train models? Where does it live?
Your data is used only for your product and never to train models for other customers. It is encrypted at rest in Scopeout's cloud; enterprise plans add private deployment, SSO and a full audit trail.
What does it connect to?
Analytics (GA4, Mixpanel, Amplitude), support (Zendesk, Intercom), design (Figma, Storybook), code (GitHub, GitLab), docs and chat (Notion, Confluence, Slack), and a browser agent for anything with a URL.
What about a product with years of history?
That is where it does best. It learns your product from what already exists (code, docs, tickets, past experiments) so the first hypotheses come from your evidence, not a blank page.
What does it cost?
Free to try, no credit card, cancel any time. Paid plans scale with the number of runs. Enterprise adds private deployment and SSO; contact us for that.
Stop running the business. Start directing it.
Join the waitlist
ScoutPrivate beta. We onboard teams in small batches. Or contact us.