DOCUMENTATION

Mobile App

EXAWin+ mobile app guide: Action Queue, activity logging, voice draft, evidence attachments, recommendations, alerts, and PWA installation

Mobile App

The EXAWin+ mobile app is a mobile-only workspace for checking deal status in the field, recording customer touches or meeting notes quickly, and keeping follow-up actions from being missed. It is not a full copy of every web administration screen. It is a PWA (Progressive Web App) focused on the sales actions people perform most often while moving.

Open the mobile app in a browser at https://[domain]/m. For the production domain, use:

https://exawin.exaeuler.com/m

After opening it in the browser, you can add or install it on the home screen and launch it with an app icon. Android supports installation from Chrome or Edge. iPhone and iPad users install it from Safari using Share > Add to Home Screen. The installed app is configured to open inside the mobile scope (/m).

Activities, comments, likes, and recommendation action records entered on mobile use the same database as the web version. A note recorded in the field is therefore visible later in Decision Console, Activity War Room, and the project timeline on the web.


Screen Layout

The Mobile App is structured with a 5-tab navigation at the bottom:

TabIconFunction
Home๐Ÿ“My project list + Search
Recordโž•Record activity + Bayesian analytics
Activity๐Ÿ“‹Global activity feed
Alerts๐Ÿ””Notification center
Settingsโš™Dark mode / Language / Profile

Header (Common)

  • EXAWin Logo โ€” Displayed in the top left corner.
  • Back โ€” A โ† arrow is displayed on the left for screens other than Home.
  • Right Area โ€” Product page link (๐ŸŒ), Docs link (๐Ÿ“„), and user avatar are positioned here.

Bottom Tab Bar

  • The currently selected tab is highlighted in the brand color, with a blue indicator bar displayed at the top.
  • Alerts Tab โ€” If there are unread notifications, the count is shown with a red badge (shows 9+ if more than 9).

Login

Authenticate using your email and password on the mobile-exclusive login screen.

ItemDescription
EmailThe company email registered at sign-up
PasswordAccount password
Auth FailureDisplays an error message with a red background
Inactive AccountAccounts deactivated by the administrator are blocked from logging in

๐Ÿ’ก If you are already logged in, you will be automatically redirected to the Home screen.


Home Tab โ€” My Projects

The Home screen displays the list of projects you are participating in.

Project Card

Each project is displayed as a card, containing the following information:

ItemDescription
Project NameDeal name. Won ๐Ÿ† / Lost ๐Ÿ“Œ icons are displayed automatically.
Customer๐Ÿข Customer name
CodeAuto-generated project code (PRJ-YYMM-XXXX)
P(Win) BadgeCurrent posterior probability percentage. The color changes based on the probability.
Progress BarVisualizes the stage progression status (0~100%)

P(Win) Badge Colors

P(Win) RangeColor
โ‰ฅ 50%Emerald (Positive)
30~49%Amber (Neutral)
< 30%Gray (Low)

Sorting Rules

Projects are sorted in the order of your most recent activities. This is based on the latest timestamp among activity records, comments, and likes.

Entering a project name in the top search bar enables real-time filtering. Unmatched cards are hidden instantly as you type.

Empty State

If you are not participating in any projects, a ๐Ÿ“‚ icon and an instructional message will be displayed.

Project Card Click

Tapping a card takes you to the Record screen, where you can log activities for that project or view its timeline.


Record Tab โ€” Activity Recording

Record is the screen used most often on mobile. It records meetings, calls, emails, SNS touches, and other sales activities, and it can connect the activity to recommendation or Bayesian analysis context when needed.

Entry paths change what the screen shows:

Entry pathBehavior
Bottom Activity tabOpens a new activity input screen.
Home deal cardOpens activity input for the selected deal.
Feed activity cardOpens the existing activity detail screen.
AlertOpens the related activity or recommendation context.
Decision recommendationOpens with context for handling the recommended action.

Project Selection

The project selector appears at the top of the new activity screen.

  • Standard users can access projects they own or where they are project members.
  • Admins and Super Users can access eligible company projects.
  • New activity entry normally lists active projects.
  • If you enter from a Home deal card, that project is selected automatically.
  • Changing the project refreshes the activity type, stage, signals, and main contact context.

30-sec log and Full log

Record provides two input modes.

ModeWhen to use itScreen behavior
30-sec logUse it to leave only the essential activity quickly while moving.Focuses on activity type, stage, key signals, short summary, and notes. Detailed areas such as contacts, AI assist, and evidence attachments are hidden.
Full logUse it for meeting notes, evidence attachments, and detailed records.Shows contact, detailed notes, evidence attachments, local AI assist, and the full input area.

30-sec log is not a different saving method. It is the same activity save flow with a simplified screen so the user can skip nonessential detail fields. The selected mode can be remembered by the browser.

Activity date and type

FieldDescription
Date & timeDate and time when the activity occurred. The default is the current time.
Activity typeMeeting, Video, Call, Email, SNS, and similar types.

Contact

In Full log mode, the project main contact can be displayed automatically. The contact identifies the customer person involved in the activity. In 30-sec log mode, this area can be hidden for faster entry.

Sales Stage

Mobile distinguishes selectable and disabled stages.

  1. You cannot move backward to an earlier non-terminal stage.
  2. Stages already passed are disabled.
  3. Terminal stages such as Won or Lost can be selected for final transition.
  4. Selecting a terminal stage can close the project.
  5. For closed projects, only the allowed latest-activity flow can add closing-correction or terminal-stage information.

Stage selection affects Bayesian updates and project state. Select only the stage that was actually reached in the sales process.

Signal Selection

Signals are positive, negative, or neutral evidence observed during a customer touch. Multiple signals can be selected.

TypeMeaning
PositiveEvidence that increases win probability.
NegativeEvidence that decreases win probability.
NeutralObservation with weak or neutral direction.

Selected signals are linked to the activity and may be used as Bayesian evidence. Incorrect signal selection can distort the analysis, so choose signals based on confirmed facts.

Summary

Summary is the short title of the activity. It is shown first on cards, timelines, and Feed. Keep it short and clear, for example:

CFO budget review meeting, PoC scope reconfirmed after technical demo

Notes

Notes contain meeting details, customer statements, next actions, and internal remarks. Good notes include customer requirements, budget, timeline, competitors, decision makers, next meetings, and the reason why selected signals were observed.

Field-level voice input

When supported by the browser, Summary and Notes can use voice input. The recognized text is appended to the existing field; Summary appends like a sentence, and Notes are easier to record with line breaks.

Mobile voice input supports the same 7 UI languages: en, ko, vi, zh, ja, es, and id. Recognition quality depends on browser, device, noise, and network state. Always review important customer statements before saving.

Quick voice note

Quick voice note is different from field-level dictation. It is used to speak a longer field note during a meeting or while moving, accumulate the recognized text into Notes, and then create an activity draft from that note.

Flow:

  1. Start Quick voice note.
  2. Speak the meeting note or field memo.
  3. Final recognized sentences accumulate in Notes.
  4. When dictation ends, Create draft from voice note is displayed.
  5. Pressing it switches to Full log and uses the Notes text to prepare Summary, action-style wording, and signal candidates.

This is not a server LLM meeting-summary feature. The current mobile behavior uses local draft logic from the browser-side Notes text: first sentence, important keywords, signal names, and keyword matching. Treat it as a draft helper, not a finished meeting-minutes generator.

Browser Local AI Assist

In Full log mode, Browser Local AI Assist can appear. It helps create a title draft or summarize the current tab from user-entered content. It does not overwrite the original text automatically; the user must review and apply results.

Evidence attachments

In Full log mode, the Evidence section can attach files or camera captures.

InputUse case
CameraBusiness card, site photo, whiteboard, visit photo.
FilePDF, document, image, spreadsheet, presentation, or zip file.

Current mobile limits are: up to 5 files per activity, 10MB per file, 30MB total, and allowed extensions pdf, doc, docx, xls, xlsx, ppt, pptx, jpg, jpeg, png, gif, webp, zip.

Uploaded attachments can be opened again from the existing activity detail screen as download links. Files attached on mobile are also visible as evidence on the same activity in the web version.

Local draft

Record can keep meaningful unsaved input in browser local storage, including Summary, Notes, stage, and signals. File inputs are not restored by local draft because of browser security rules. Offline activity submission is blocked; server save requires internet access.

Saving

When you press Save, the app performs the following checks and actions:

  1. Required values are validated.
  2. Project, contact, stage, and signals are checked against company scope and user permissions.
  3. Invalid stage changes, such as moving backward to an earlier non-terminal stage, are blocked.
  4. A new activity is created or an existing editable activity is updated.
  5. Selected signals are linked to the activity.
  6. Attachments are saved after file count, size, and extension checks.
  7. If the screen was opened from a recommendation, the activity can be linked to the decision ledger execution record.
  8. After a successful save, the activity detail screen opens.

New mobile activities are saved as draft-style records. Approval, lock, deletion, and broader management tasks remain better suited to the web version.


Viewing and Editing Existing Activities

Existing activity details can be opened from Feed, Alerts, the project timeline, or contextual Home entries.

ItemDescription
Activity NoActivity number.
ProjectThe project that owns the activity.
AuthorUser who recorded the activity. A Me badge can appear for your activity.
Activity timeTime of the customer touch or activity.
Activity typeMeeting, Video, Call, Email, SNS, and similar types.
StageSales stage selected for that activity.
SignalsObserved evidence linked to the activity.
Summary / NotesShort title and detailed notes.
AttachmentsDownload links for saved evidence files.
Likes / CommentsCollaboration reactions and discussion.

Editing permissions

SituationMobile behavior
My unapproved activityEditable.
Admin or Super UserCan edit accessible activities.
Project ownerCan edit project activities.
Approved activityNot editable.
Another user's activity without permissionRead, like, and comment oriented.
Past activity of a closed projectUsually not editable; closed projects are limited to the latest-activity flow.

When you do not have edit permission, the screen becomes read-only. Likes and comments can still be used within access permissions.

Likes

Tapping Like applies immediately. Tapping again removes it. Liking another user's activity can create an alert for the author.

Comments and replies

Activities support comments and replies. Use comments for next-action requests, internal checks, and customer-statement clarification. Comments and replies can create alerts for related users.


Recommendation and Decision Context

Mobile Record can connect activity entry with recommendations created or adopted in Decision Console.

Decision Needed deals appear when a recommendation has been adopted but has not yet been linked to a real sales activity. Opening such a deal can show recommendation title, reason, project, recommended next action, acceptance/ignore state, and the information that will be connected when the activity is saved.

ActionMeaning
Accept recommendationAdopts the recommendation.
Ignore recommendationIgnores the current recommendation.
Log sales actionRecords the recommendation as a real sales activity.
View Decision ConsoleOpens the broader web Decision Console context.

Follow-up Due deals appear when there has been no customer contact for the required period. Risk Rising deals are not selected simply because P(Win) is below the project threshold. Risk classification uses Bayesian insight states such as freshness, stale, cooling, and momentum. A low P(Win) may mean more meetings or evidence are needed, not necessarily that the deal is risky.


Bayesian Analytics and Saved Analysis

The mobile Record screen prioritizes fast field entry. It does not create heavy analysis automatically every time you select a project. Instead, the analysis area is opened only when the user needs it, through two separate buttons.

ButtonBehaviorServer processingWhen to use it
View saved analysisLoads the latest analysis already saved on the server.Does not create new analysis. It reads the latest Bayesian update, Decision/Ontology recommendation ledger, and saved AI draft data.Use it when you need to quickly review the current final analysis in the field.
Generate new analysisRuns a new analysis from the current project, latest activities, stage, signals, recommendation history, and ontology relationship data.Runs Bayesian insight and Decision/Ontology reasoning, then saves a recommendation ledger. The generated result becomes the latest saved analysis.Use it after recording a new meeting on mobile or tablet and you want to refresh the analysis immediately.

View saved analysis is read-only. Pressing it does not run Bayesian calculation, ontology reasoning, or an AI runtime. This keeps the Record screen fast and lets you reuse the final analysis already created from web or mobile.

Generate new analysis runs only when you explicitly press the button. It can take time, similar to generating analysis on the web. When generation completes, a new recommendation ledger is saved. After logging out and logging back in, View saved analysis should show that latest result. If the saved view does not match the analysis you just generated, check cache, save failure, permissions, or ledger update issues.

AI draft appears only when AI draft data is stored in the recommendation ledger. Generating a new Bayesian/Decision/Ontology analysis updates the analytical record, but it should not be understood as always creating a new AI-written interpretation.

8.1 When the Analysis Panel Opens Automatically

The detailed Bayesian Analytics panel can appear directly when an analysis context is present.

  • You moved to mobile Record from a Decision Console recommendation.
  • You opened Record from a recommendation-related alert.
  • The URL contains analysis context such as decision_id or analytics=1.
  • You pressed Generate new analysis in the Record screen.
  • You pressed View saved analysis and requested the saved result.

If you simply open a new activity form from the bottom Activity tab, the heavy analysis panel may not appear automatically. This is intentional to protect mobile input speed. You can still use View saved analysis or Generate new analysis when needed.

8.2 Where Saved Analysis Data Comes From

Saved analysis is not a single summary sentence. It combines several saved and reconstructed data sources.

DataMeaningWhere it appears on mobile
Bayesian updateLatest stored P(Win) calculation from activities, stages, signals, and silence penalty.P(Win), change, 95% CI, Impedance cards
Recommendation ledgerStored Decision/Ontology recommendation and decision record.Recommendation, Decision detail, Tracking, Data basis
Decision payloadDecision state, decision band, next action, and drivers stored inside the ledger.Recommendation label, decision band, reason for the recommended action
Ontology ops payloadDisplay context rebuilt from the current project's relationships, bottlenecks, missing links, and evidence paths. It does not create a new ledger in saved-analysis mode.Why constrained, Bottleneck, Next action, Evidence summary, Missing evidence
AI draftSaved draft state and text when AI draft generation has been stored in the ledger.AI draft area

In short, View saved analysis brings the latest saved analysis back onto the screen. It does not create a new decision, but it combines the saved ledger with safe display context so the user can understand the result as cards.

8.3 Overall Screen Structure

Depending on what is saved, the analysis area can show the following cards.

AreaWhat you seeHow to read it
P(Win)Current win probability and change from the previous update.Use it as a quick view of the deal's current strength. Do not decide from the probability alone.
95% CI / ImpedanceUncertainty range for P(Win) and confidence against the threshold.A wide range means weak evidence. Impedance shows how convincingly P(Win) clears the decision threshold.
RecommendationProposed next action and recommendation state.This is not an order. It is the next sales action most likely to improve the state based on current evidence.
Decision detailWhy constrained, Bottleneck, Next action, and Tracking.Check why the deal is constrained, where the bottleneck is, what to do next, and whether the recommendation is tracked.
AI ContextEvidence summary and missing evidence prepared before AI execution.This is fact-based context before AI writing. Use it to know what to confirm with the customer.
Data basisCounts for evidence paths, missing links, bottlenecks, and tracking state.Use it to understand what data structure supports the analysis. It also helps support and verification.
AI draftStored AI draft state, output code, and text.Shown only when a stored AI draft exists. Otherwise it appears as not generated.

8.4 How to Read Bayesian Metrics

Users new to Bayesian analysis often look only at the P(Win) number. In EXAWin+, judgment should combine P(Win), credible interval, evidence volume, recent momentum, silence penalty, and threshold confidence.

MetricPlain explanationUser judgment
P(Win)Current win probability calculated from stored evidence. For example, 65% means the current evidence is stronger on the winning side, not that the deal is guaranteed.If high, strengthen closing preparation and decision-maker confirmation. If low, first check whether evidence is sufficient before giving up.
ChangeDifference from the previous Bayesian update in percentage points. 42% to 48% is +6%p.Identify which activity raised or lowered probability.
95% CIUncertainty range of P(Win). A wide range such as 38% - 72% means the system cannot judge confidently yet.Collect more decisive evidence such as key meetings, decision-maker confirmation, or budget confirmation.
AlphaAccumulated positive evidence strength. Positive signals, positive responses in later stages, and valid customer contact can increase it.Rising Alpha means win-side evidence strengthened. Confirm that selected signals match facts.
BetaAccumulated negative or loss-side evidence. Negative signals, competitor advantage, delays, and silence penalty can increase it.Rising Beta means you should check customer response, missing data, or long gaps without contact.
Evidence strengthEvidence maturity, broadly Alpha plus Beta. With little evidence, P(Win) can move sharply but still be unreliable.In early deals, the key question is not just low P(Win), but what evidence is still missing.
Stage / SWVSales stage changes the weight of evidence. The same positive signal can matter more in Proposal or Negotiation than in early discovery.Do not choose a later stage unless the deal actually reached it.
Silence penaltyA silence effect reflected toward Beta when there has been no customer contact for a configured period.It turns lack of contact into a measurable warning. Recording a customer touch can change future analysis.
Momentum P(Win)Recent-evidence-weighted flow metric. If it is below current P(Win), recent momentum may be cooling.Check whether you are relying on old positive evidence while recent response is weakening.
ImpedanceDecision confidence showing how convincingly P(Win) passes the project threshold.Around 50% means near-threshold ambiguity. Higher values mean stronger grounds for action at the current stage.
TrendMovement of P(Win) across multiple Bayesian updates.Look for repeated rises or drops after specific meetings and plan the next activity accordingly.

A P(Win) below threshold does not automatically mean the deal is risky. Early deals with little evidence may simply need more meetings and confirmation. Risk classification also uses freshness, stale, cooling, and momentum states.

8.5 Recommendation Items

The Recommendation card shows the next action proposed by the saved Decision/Ontology analysis.

ItemMeaningHow to use it
Recommended actionThe most likely next sales action to take now, such as recording customer follow-up, confirming the decision maker, or checking missing conditions.If you will do it, record it as an activity in Record. If not, ignore or review it in web Decision Console.
Response stateWhether the recommendation is pending, accepted, ignored, and so on.If accepted but no real activity is linked, the deal may remain in Decision Needed.
Decision bandA decision zone such as below threshold, near threshold, or clearly above threshold.Use it to understand the judgment zone faster than reading probability alone.
Confidence labelHow mature and reliable the current evidence is.If confidence is low, additional confirmation may be needed even when a recommendation exists.
Freshness labelHow fresh the latest activities and evidence are.If stale or cooling, customer contact and updated information come first.
WhyCore reason for the recommendation.Convert it into questions or notes in Summary and Notes.

Recommendation does not replace user judgment. It shows which action is most likely to improve the analytical state based on the stored evidence.

8.6 Decision / Ontology Detail Items

Decision detail compresses the core of web Decision Console and Ontology analysis into mobile cards. Because mobile cannot show the full graph or every tab, it focuses on four field-action questions.

CardQuestionExplanation
Why constrainedWhy is this deal constrained now?Gives the shortest explanation of the bottleneck, such as no official customer response, no buying-center link, or insufficient recent evidence.
BottleneckWhat structural bottleneck is blocking judgment?Shows whether relationship, activity, evidence, buying center, or decision stage is weak. It does not simply mean P(Win) is low.
Next actionWhat should be done next?Converts the recommendation into an executable sales action, such as contacting the customer, scheduling a meeting, or confirming a condition.
TrackingIs the recommendation being executed and tracked?Shows whether an adopted recommendation has been linked to real sales activity and can affect later analysis.

Recommended reading order:

  1. Read Why constrained to understand what is blocking the deal.
  2. Read Bottleneck to see which data or relationship is weak.
  3. Turn Next action into a real customer touch or internal preparation task.
  4. After saving the activity, check whether Tracking links the activity to the recommendation.

8.7 AI Context and Missing Evidence

AI Context is not a free-form AI opinion. It is the evidence context confirmed by the server before AI execution. On mobile it appears as Evidence summary and Missing evidence.

ItemMeaningHow to use it
Evidence summaryFacts currently used by the analysis.Check what the system is basing its judgment on. If it differs from what you know, update activity records or master data.
Missing evidenceLinks or evidence that ontology checks found insufficient.Use it as the next customer-question list: buying-center member missing, customer contact missing, evidence path missing, and so on.
Recommended inputContent worth adding in the next record.Reflect it in Summary, Notes, signals, and contact links.

AI Context is more basic than AI draft. Even when no AI draft exists, AI Context and Missing evidence can tell you what to supplement next.

8.8 Data Basis

Data basis shows the saved records and ontology summary behind the analysis. Users do not need to memorize IDs, but they are useful when verifying a result or asking for support.

DisplayMeaning
Bayesian update idThe Bayesian update record used to display P(Win).
Recommendation ledger idThe recommendation/decision record used for Recommendation and Decision detail.
Decision UUIDIdentifier used to track the same decision result across Decision Console and Ontology analysis.
Evidence paths countNumber of evidence paths found by the analysis.
Missing links countNumber of relationship links considered missing by ontology checks.
Bottleneck countNumber of bottleneck candidates currently displayable.
Tracking stateWhether the recommendation is accepted, ignored, linked to execution, and so on.

View saved analysis reads the latest ledger. If a new analysis is created from web or mobile, the ledger id or decision UUID may change.

8.9 AI Draft

AI draft appears only when an AI interpretation draft has been saved.

ItemMeaning
StateProcessing state of the AI draft, such as generated, waiting for review, applied, or rejected.
Output codeCode used to classify the draft result, such as success, runtime limit, or no valid output.
TextStored AI draft text or summary.
Recorded atTime when the AI draft was recorded.

If AI draft is missing, it does not mean Bayesian/Decision/Ontology analysis failed. Mobile saved analysis can still show P(Win), recommendation, bottleneck, and missing evidence from stored mathematical and ontology data.

To record a new meeting and review analysis on mobile or tablet:

  1. Select a deal from Home and open Record.
  2. Enter activity type, stage, signals, Summary, and Notes based on the real meeting.
  3. Attach evidence such as business cards, site photos, or documents if needed.
  4. Save the activity.
  5. Reopen the project in Record and press Generate new analysis.
  6. Review P(Win), 95% CI, Recommendation, Why constrained, Bottleneck, and Next action.
  7. Turn Next action into customer contact, internal preparation, or decision-maker confirmation.
  8. Later, use View saved analysis to quickly reopen the same final analysis.

If time is short, save the activity first and run Generate new analysis later on a stable network.

8.11 Common Misunderstandings

MisunderstandingCorrect understanding
Low P(Win) always means a risky deal.Not necessarily. Early deals with little evidence may show low P(Win). Risk classification also considers freshness, stale, cooling, and momentum.
High P(Win) means no more action is needed.No. If recent activity stopped or the decision maker is unconfirmed, high P(Win) can weaken.
View saved analysis creates a new analysis.No. It only reads the latest saved result from the server.
Generate new analysis always creates a new AI sentence.No. It updates Bayesian/Decision/Ontology analysis storage. AI draft appears only when separate draft data exists.
Wide 95% CI is fine as long as P(Win) is high.No. A wide interval means the evidence base is unstable. Collect key evidence first.
Bottleneck means the deal failed.No. It means there is a weak link or missing evidence blocking judgment. It can improve if supplemented.

8.12 Help and Detailed Analysis

Mobile saved analysis cards compress the key items for field use. Web Activity War Room and Decision Console provide broader panels for Bayesian metrics, timeline, recommendation tracking, ontology graph, and evidence paths.

Mobile does not always expose every web-style metric help popover. For deeper formulas and theory, use the web Bayesian panel, Decision Console, or official BAT Parameters and Threshold/Impedance documents.


Activity Timeline

If activities have been logged for the project, a timeline is displayed at the bottom of the Record screen.

Timeline Card

ItemDescription
Activity NoAuto-generated number (in brand blue)
DateActivity date/time (MM/DD HH:MM)
Type Icon๐Ÿค Meeting / ๐Ÿ“น Video / ๐Ÿ“ž Call / ๐Ÿ“ง Email / ๐Ÿ’ฌ SNS
SummaryActivity title (1-line truncated)
StageIndicated with a brand color
AuthorUser's name. A "Me" badge is added for your own activities
โค๏ธ/๐Ÿ’ฌLike/Comment count (shown if there is at least one)

Status Display

  • Activity Currently Being Viewed โ€” Highlighted with a vertical brand-color line on the left
  • Activity with Unread Notifications โ€” Blue background + pulsing blue dot

Tapping the card redirects you to the detail screen of that activity.


Activity Tab โ€” Global Activity Feed

An integrated feed where you can see the activities of all the projects you are participating in, in one place.

Feed Card

ItemDescription
Project NameThe project the activity belongs to (๐Ÿ†/๐Ÿ“Œ status indicated)
DateMM/DD HH:MM
Type IconActivity type emoji
SummaryActivity title
Activity NoAuto-generated number
StageBrand color
AuthorName + "Me" badge (Your activity)
โค๏ธ/๐Ÿ’ฌLike/Comment counts

Sorting Rules

  1. Activities involving you are at the top: Activities you authored, or left comments/likes on, appear first.
  2. After that, ordered descending by happened_at (activity timestamp).
  3. Up to 50 items are displayed.

Tapping the card takes you to the Record detail screen for that activity.


Alerts Tab โ€” Notifications

Notifications are generated when other users leave likes, comments, or replies on your activities.

Alert Card

ItemDescription
Read Status๐Ÿ”ต Blue dot (Unread) or Gray dot (Read)
ActorAvatar + name of the user who triggered the alert
Action๐Ÿ’ฌ "left a comment" / โ†ฉ "left a reply" / โค๏ธ "liked"
PreviewFirst 60 characters of the comment or activity summary
TimeMM/DD HH:MM

Behaviors

  • Unread Alerts โ€” Highlighted with a blue background and a pulsing dot
  • Read Alerts โ€” Displayed at 60% opacity
  • Tapping an alert will:
    1. Automatically mark it as read (via AJAX).
    2. If there is an associated activity, automatically navigate to the Record detail screen.
    3. If there is no associated activity, it will just mark it as read.

Sorting Rules

Unread alerts are pinned to the top, followed by chronological ordering. Displays up to 50 items.


Settings Tab

Profile Card

  • User Avatar โ€” Name initials displayed in a gradient circle
  • Name and Email
  • Role Badge โ€” Admin / Super User / Standard, etc.

Appearance

  • Dark Mode Toggle โ€” ๐ŸŒ™ icon and switch to toggle between light/dark mode
  • Settings are saved to localStorage and will persist on your next visit.

Language

Supports 7 languages. Represented alongside their flag icons:

CodeLanguageFlag
enEnglish๐Ÿ‡บ๐Ÿ‡ธ
koํ•œ๊ตญ์–ด๐Ÿ‡ฐ๐Ÿ‡ท
viTiแบฟng Viแป‡t๐Ÿ‡ป๐Ÿ‡ณ
zhไธญๆ–‡๐Ÿ‡จ๐Ÿ‡ณ
jaๆ—ฅๆœฌ่ชž๐Ÿ‡ฏ๐Ÿ‡ต
esEspaรฑol๐Ÿ‡ช๐Ÿ‡ธ
idBahasa Indonesia๐Ÿ‡ฎ๐Ÿ‡ฉ

The currently selected language is highlighted with a brand-color border + โœ“ icon. Tapping a language switches it instantly.

App Info

  • Version โ€” Current app version (1.0)
  • Company Name โ€” Affiliated company

Logout

Tapping the red "Sign Out" button ends the session and navigates to the login screen. An Audit Log is automatically recorded.

Open Web Version

Tapping the "Open Web Version" link switches over to the full-featured desktop web browser.


PWA (Progressive Web App) Support

EXAWin+ mobile runs as a PWA. A PWA is a web app that can be launched like an installed app after being added to the home screen.

Android installation

  1. Open https://exawin.exaeuler.com/m in Chrome or Edge.
  2. Press the install button on the login screen or header.
  3. Confirm the browser installation prompt.
  4. An EXAWin+ icon is created on the home screen or app list.
  5. Launch the mobile app from the icon.

The wording can vary by browser and Android version: Install app, Add to Home screen, or similar.

iOS installation

  1. Open https://exawin.exaeuler.com/m in Safari.
  2. Tap Share.
  3. Select Add to Home Screen.
  4. Confirm the name and add it.
  5. Launch from the EXAWin+ icon.

Service worker and cache

The mobile app uses a service worker to cache some static assets and mobile GET screens. The current cache name is exa-sales-v5.

  • Login and authentication pages are not cached.
  • Mobile GET screens can be cached after successful network responses.
  • Write operations such as saving activities, comments, likes, and recommendation handling require internet access.
  • Old caches are cleaned when the new version runs.

When an old screen remains visible

  1. Fully close the app and launch again.
  2. Hard refresh in the browser.
  3. Delete site storage for the browser being used.
  4. If the installed PWA remains stale, remove the icon and install again.

If QR scanning opens Edge, delete Edge site data and cache. To test in Chrome, open the mobile URL directly in Chrome.


Permission System

The mobile app follows the same permission principles as the web version. Users can only see company data they are allowed to access.

FunctionPermission basis
Project viewProject owner, project member, Admin, Super User.
Activity recordingAllowed for accessible projects.
Activity editingAuthor, Admin, Super User, or project owner; approved activities are excluded.
Closed project editingLimited to the latest-activity flow and requires the required permission.
Likes / commentsAvailable on accessible activities.
Recommendation handlingAvailable for recommendations linked to accessible projects.
Attachment downloadOnly attachments of accessible activities can be opened.

Unauthorized projects or activities are hidden or blocked on access.


Data Relations

Home
  -> Action Queue
     -> Decision Needed: adopted recommendation not yet linked to an execution activity
     -> Follow-up Due: customer-contact silence exceeded the configured threshold
     -> Risk Rising: Bayesian insight cooling/stale/momentum caution state
     -> My Active Deals: active, non-closed deals

Record
  -> Select project
  -> View saved analysis: read saved Bayesian/Decision/Ontology/AI draft data only
  -> Generate new analysis: explicitly create new analysis and save recommendation ledger
  -> Enter type, time, stage, signals, Summary, Notes
  -> Link selected signals to the activity
  -> Save attachments
  -> Link decision ledger execution if entered from a recommendation
  -> Open activity detail after saving

Feed
  -> Recent accessible activities, up to 50
  -> Prioritize my activities or activities I interacted with

Alerts
  -> General alerts and recommendation alerts
  -> Open related activity or recommendation context

Settings
  -> Language, dark mode, app info, sign out

Differences from the Web Version

ItemWeb VersionMobile App
LayoutMulti-panel work screensFocused tab structure, one task at a time
Deal reviewDecision Console and Activity War Room detailed analysisHome Action Queue prioritizes deals requiring action
Activity inputFull fields and rich editing30-sec log / Full log for field entry
Activity editingDetailed management with approvals and locksEditable when permitted; approved activities are restricted
Closed projectsFull state managementLimited latest-activity and terminal-stage related entry
AttachmentsDesktop upload interactionsCamera and file picker evidence upload
Voice inputLimited or not primaryWeb Speech field input and Quick voice note
RecommendationsDecision Console orientedFast connection of recommendations to real sales actions
Bayesian analysisWide panels and Decision/Ontology analysis generationFast saved-analysis view plus separate Generate new analysis button
OfflineNormal web behaviorSome screen cache and local draft; server save requires online access
SettingsSystem and administration settingsLanguage, dark mode, app info, sign out

Important Notes

  1. The mobile app is optimized for field records and fast actions. Use the web version for bulk management, system settings, and detailed analysis review.
  2. Stage and signal choices affect Bayesian analysis. Enter only sales facts actually confirmed.
  3. Moving backward to an earlier non-terminal stage is not allowed. If a project was closed by mistake, review the allowed latest-activity flow or use web management.
  4. P(Win) below threshold does not automatically mean the deal is risky. Risk classification also considers freshness, momentum, stale, and cooling states.
  5. Quick voice note draft generation is a local helper. It is not a high-quality server LLM summary for long meeting minutes; review before saving.
  6. Attachments follow file count, size, and extension limits. Use the web version or another sharing path for unsupported large files.
  7. Offline mode cannot save to the server. Local drafts must be submitted after network recovery.
  8. If an installed app shows an old screen, fully close it or delete browser site storage and reopen.
  9. On shared devices, always press Sign Out after use.
  10. Mobile browser voice, camera, and file features can vary depending on device and browser permission settings.