Platform

Workspaces

Every tab in the sidebar, and the concrete things Kai can do on each one. Kai is route-aware, so a question on Email gets a different default from the same words in Code.

Kaiday replaces the tool stack, not integrates with it

Kaiday ships its own messaging, video meetings, calendar, documents, presentations, and email. The goal isn't to sync with Google Calendar or Slack. It's to give you a single workspace where all of that lives natively and Kai sees the whole picture.

Home

The dashboard you land on after login: live execution health, what needs you across every surface, and the briefing for your day.

What Kai can do here

→Summarize what changed across email, calendar, code, and messaging since you last looked.
→Draft your day. What to focus on first, what can wait.
→Surface what's overdue or stalled, with a one-click handoff to fix it.

Try saying to Kai

›What's on fire this morning?
›What did I miss while I was out yesterday?
›Plan my Tuesday. Top 3 things to ship.

Kai

The full conversation surface. The same Kai that lives on every other tab, with more room to think out loud.

What Kai can do here

→Pull together context from across the workspace (email, clients, code, docs) and answer in one place.
→Run multi-step jobs: 'draft a proposal for this client, attach the latest case study, schedule a follow-up next Tuesday.'
→Use context from every connected surface without switching workspaces.

Try saying to Kai

›Draft a proposal for the Acme renewal, attach the Q3 case study, and book a follow-up next Tuesday.
›What's the status of every open client engagement?
›Show me what's pending across the workspace.

Messaging

Kaiday's own internal chat: channels, DMs, mentions, threads. This replaces Slack. Don't bridge to it.

What Kai can do here

→Summarize long threads into the actual decisions and open questions.
→Draft replies in any channel, in your voice.
→Search every channel and DM by meaning, not just keyword.
→Spin up new channels for new clients, projects, or initiatives.
→Turn a chat decision into an Email or Calendar invite without leaving the thread.

Try saying to Kai

›Summarize the #launch thread from the last 24 hours.
›Reply to the last message in #ops in my voice.
›Make a channel for the momentum study and add the model reviewer.

Automations

Two tabs: Tasks shows every operation Kai is running for you (chat-triggered or workflow-triggered), with status, errors, and a link to the conversation that produced it. Workflows is where you define event-driven or scheduled pipelines.

What Kai can do here

→Build workflows from a description: 'when a study reaches review, notify the assigned reviewer.'
→Schedule recurring jobs: 'every Monday at 7am, post the week's plan to #team.'
→Wire MiniApp events to actions in any other surface.
→Retry, cancel, or open the chat for any task that failed.

Try saying to Kai

›When a study reaches review, DM the reviewer and include its main limitation.
›Every Monday at 7am, post a digest of completed studies to #research.
›Retry the last failed code_implement task with one more iteration budget.

Email

A full inbox. Multi-org, multi-mailbox; mint addresses on any domain you've verified. Kai handles entire conversations, not just one-shot replies: it carries the thread context, the client's history with you, and the outcome you're trying to reach into every draft it produces.

Imagine this. The research inbox is full before market open

Kai has triaged dataset notices, paper alerts, reviewer questions, and model-risk requests. Routine messages have draft replies; one request that would change a live methodology is flagged for human review with the relevant study context.

You read the flag, edit the draft for ten seconds, and send. The 47 other emails were already handled in your voice.

What Kai can do here

→Draft replies in any thread with the relevant study, evidence, and review context.
→Handle multi-message research conversations while tracking answered and open questions.
→Compose new emails from a one-line description: 'ask the data provider about the missing 2018 history.'
→Auto-classify inbound mail (data, research, risk, operations, noise).
→Send follow-up sequences on a cadence you describe.
→Detect when a thread needs you (negotiation, refund, complaint) and escalate instead of replying.

Try saying to Kai

›Handle this data-provider thread, summarize the coverage gap, and stop before accepting new terms.
›Reply to every email tagged 'study review' with the current status, but escalate methodology objections.
›Ask each assigned reviewer for a decision on studies that have been waiting more than seven days.

Calendar

Kaiday's own calendar and video meetings. No Google or Outlook to sync, no Zoom to install. Events, prep notes, live video, and recordings all live here.

Imagine this. Every meeting becomes 3x more useful

You have a discovery call at 3pm. At 2:55, Kai DMs you a one-page prep brief: who's joining, every email they've ever sent you, what they bought last time, the open question from your last conversation, and three good questions to ask.

You join the meeting. Kai records and transcribes silently in the background. The call ends at 3:42. By 3:45, the client has an email summarizing the next steps and a draft proposal is in your Knowledge Base waiting for you to glance at.

You didn't take a single note.

What Kai can do here

→Create events and recurring meetings from a sentence: 'book me with Sara next Tuesday afternoon.'
→Find open time across calendars and schedule meetings.
→Run meetings on Kaiday's video stack. No third-party app to install.
→Record meetings and transcribe them automatically.
→Turn a recording into action items and follow-up emails.
→Prepare prep notes before every meeting: who's coming, what changed since you last talked, what to ask.

Try saying to Kai

›Book a 30-minute methodology review and prep me with the study's assumptions and limitations.
›Record the investment committee meeting and send the action items to reviewers afterward.
›Find an open 60-minute slot Tuesday or Thursday for the model-risk review.

Apps

Hosted financial workflow apps live here. Kai customizes a code-backed starting point and publishes it under your domain. MiniApps do not process payments or execute trades.

Imagine this. Saturday morning, an idea, a working app by lunch

You need a review dashboard that puts benchmark, holdout, sensitivity results, limitations, and reviewer state on one screen.

At 9:14am you say to Kai: "build a study review dashboard at research.example.com with benchmark, temporal holdout, sensitivity, and decision status."

The dashboard is deployed to the research domain and the first completed study is ready for review.

What Kai can do here

→Build a MiniApp from a description: 'evidence library with citations and paper roles.'
→Host financial workflow apps under your verified domain (e.g., research.example.com).
→Update an app by talking: 'add transaction-cost assumptions to the review form.'
→Wire apps to approved research data and namespaced app storage.

Try saying to Kai

›Build a hypothesis intake with benchmark, costs, and rejection criteria.
›Add a limitation field and reviewer decision to the study dashboard.
›Publish the review app at research.example.com.

See the full MiniApps section.

Knowledge Base

Kaiday's own documents, sheets, and presentations. Searchable, editable, shareable. This is where the living memory of the company lives, and where Kai writes when you ask it to put something on paper.

Imagine this. The pitch deck you've been avoiding all week

You have a board meeting Monday. You've been dreading the deck. You say to Kai: "Build a 12-slide board update, Q3 results pulled from production, biggest open risks, and our 18-month plan. Use the brand template in the KB."

Twenty minutes later the deck is in your Knowledge Base. Numbers from the database. Risk language pulled from last week's exec meeting transcript. The brand template applied. You spend Sunday tightening three slides instead of building 12 from scratch.

What Kai can do here

→Create documents, sheets, and decks from a description.
→Edit existing documents. Rewrite a section, restructure an outline, add a chart.
→Generate presentations: investment-committee decks, portfolio reviews, board updates, and methodology training.
→Edit slides by talking: 'remove slide 4, tighten the headline on slide 6, add a competitive landscape slide.'
→Generate decks from meeting notes, study packages, or a single-sentence brief.
→Search every document by meaning, not keyword; cite documents inline when you ask a question in any other surface.

Try saying to Kai

›Generate a 10-slide investment-committee deck from the approved study package.
›Turn yesterday's methodology review transcript into a one-page decision memo.
›Add a sensitivity-analysis slide to the portfolio review deck. Pull from the Knowledge Base.

Code

Connect a GitHub repo and Kai writes code for you. Not autocomplete. Not snippets. Real pull requests, opened on real branches, with the diff written, the tests run, and a PR description that summarizes what changed and why. You review and merge. Kai never touches main on its own.

Imagine this. Tuesday, 2:47pm

A researcher reports that the return series shifts by one day after a timezone change. You ask Kai to reproduce it and prepare a fix.

Kai reproduces the timestamp misalignment, traces it to the normalization boundary, and opens a draft pull request with the fix and a regression test.

You read the diff, merge, and reply to the reviewer that it's deployed. Total time on your side: four minutes.

What Kai can do here

→Open real draft pull requests on GitHub. Diff written, tests run, PR description filled in.
→Diagnose bugs from a reviewer report, a stack trace, or a failed study run.
→Read your codebase to answer technical questions: 'where is survivorship bias controlled?' 'why is this study slow?'
→Refactor when you ask: 'this file is too long, split it into modules.'
→Review teammate PRs and flag risks before you merge them.
→Never force-push. Never touch main without your merge.

Imagine this. A Sunday research review

You notice the factor chart needs a log-scale option. You don't want to open the laptop. You DM Kai from your phone: "add a log-scale toggle, remember the reviewer preference, draft PR only."

By the time you get home, the PR is waiting. You merge it Monday morning over coffee. The feature you'd have shipped "someday" is in production this week.

Try saying to Kai

›The daily returns shift around the DST boundary. Reproduce it and open a draft PR with a regression test.
›Add a log-scale toggle to the factor chart. Draft PR only.
›PR #421. Review it and tell me if anything's risky before I merge.
›Split the auth.ts file into modules. Draft PR.

Connections

Plug in your production database, Postgres, MySQL, or MongoDB, and Kai treats it like part of your workspace. You don't write SQL. You don't open a BI tool. You ask in English, and the answer comes back as a number, a chart, a sheet, a slide, or an email. Whatever shape you asked for.

Imagine this. Thursday 4pm, board update due Monday

You don't open the database. You don't export a CSV. You don't fight with the BI tool. You say to Kai:

"Pull last quarter's attribution, exposures, drawdowns, and benchmark deltas from production Postgres, and build a 12-slide portfolio review deck plus a backup sheet. Cite the queries."

By the time you leave for the day, the deck and the sheet are sitting in your Knowledge Base. The deck looks like your brand. The numbers are real. Every slide has a footnote with the exact query Kai ran, so when the board asks "where does this come from?" you can show them.

What Kai can do here

→Answer ad-hoc questions against your production data in plain English.
→Turn data into the artifact you actually need. A deck, a sheet, an email, a doc, a slide for a meeting.
→Build recurring reports that auto-run on a schedule and land in Messaging or your inbox.
→Compare two market-data sources and flag mismatched observations.
→Cite the queries it ran, so the audit trail is real.
→Read-only by default; writes require explicit per-connection approval.

Imagine this. A committee asks a question mid-call

You're on a Kaiday investment-committee meeting. They ask: "what drove active return in the top decile?" You open Kai and type: "active return by sleeve, top-decile positions, last four quarters, chart it with benchmark context."

By the time the investor finishes their next sentence, the chart is on your screen with the underlying numbers. You share it. The call ends differently than it would have.

Try saying to Kai

›Pull Q3 attribution and risk metrics from production Postgres and build a 12-slide review deck. Cite the queries.
›Which EU positions breached their risk limits in the last seven days? List the source and as-of date.
›Every Friday at 5pm, post an exposure-and-drawdown digest to #portfolio-review.
›Active return by sleeve, top-decile positions, last four quarters. Chart it.

Control (admin)

Transparency and audit: every task Kai ran, every override, every memory edit. Admins only.

What Kai can do here

→Surface anything that looks unusual in the audit log.
→Answer 'what did you do last Thursday?' with the actual trace.

Try saying to Kai

›What did you do last Thursday between 2 and 5pm?
›Anything unusual in the audit log this week?

Connect data & code Automations

Platform

Workspaces

Every tab in the sidebar, and the concrete things Kai can do on each one. Kai is route-aware, so a question on Email gets a different default from the same words in Code.

Kaiday replaces the tool stack, not integrates with it

Home

The dashboard you land on after login: live execution health, what needs you across every surface, and the briefing for your day.

What Kai can do here

→Summarize what changed across email, calendar, code, and messaging since you last looked.
→Draft your day. What to focus on first, what can wait.
→Surface what's overdue or stalled, with a one-click handoff to fix it.

Try saying to Kai

›What's on fire this morning?
›What did I miss while I was out yesterday?
›Plan my Tuesday. Top 3 things to ship.

Kai

The full conversation surface. The same Kai that lives on every other tab, with more room to think out loud.

What Kai can do here

→Pull together context from across the workspace (email, clients, code, docs) and answer in one place.
→Run multi-step jobs: 'draft a proposal for this client, attach the latest case study, schedule a follow-up next Tuesday.'
→Use context from every connected surface without switching workspaces.

Try saying to Kai

›Draft a proposal for the Acme renewal, attach the Q3 case study, and book a follow-up next Tuesday.
›What's the status of every open client engagement?
›Show me what's pending across the workspace.

Messaging

Kaiday's own internal chat: channels, DMs, mentions, threads. This replaces Slack. Don't bridge to it.

What Kai can do here

→Summarize long threads into the actual decisions and open questions.
→Draft replies in any channel, in your voice.
→Search every channel and DM by meaning, not just keyword.
→Spin up new channels for new clients, projects, or initiatives.
→Turn a chat decision into an Email or Calendar invite without leaving the thread.

Try saying to Kai

›Summarize the #launch thread from the last 24 hours.
›Reply to the last message in #ops in my voice.
›Make a channel for the momentum study and add the model reviewer.

Automations

What Kai can do here

→Build workflows from a description: 'when a study reaches review, notify the assigned reviewer.'
→Schedule recurring jobs: 'every Monday at 7am, post the week's plan to #team.'
→Wire MiniApp events to actions in any other surface.
→Retry, cancel, or open the chat for any task that failed.

Try saying to Kai

›When a study reaches review, DM the reviewer and include its main limitation.
›Every Monday at 7am, post a digest of completed studies to #research.
›Retry the last failed code_implement task with one more iteration budget.

Email

Imagine this. The research inbox is full before market open

You read the flag, edit the draft for ten seconds, and send. The 47 other emails were already handled in your voice.

What Kai can do here

→Draft replies in any thread with the relevant study, evidence, and review context.
→Handle multi-message research conversations while tracking answered and open questions.
→Compose new emails from a one-line description: 'ask the data provider about the missing 2018 history.'
→Auto-classify inbound mail (data, research, risk, operations, noise).
→Send follow-up sequences on a cadence you describe.
→Detect when a thread needs you (negotiation, refund, complaint) and escalate instead of replying.

Try saying to Kai

›Handle this data-provider thread, summarize the coverage gap, and stop before accepting new terms.
›Reply to every email tagged 'study review' with the current status, but escalate methodology objections.
›Ask each assigned reviewer for a decision on studies that have been waiting more than seven days.

Calendar

Kaiday's own calendar and video meetings. No Google or Outlook to sync, no Zoom to install. Events, prep notes, live video, and recordings all live here.

Imagine this. Every meeting becomes 3x more useful

You didn't take a single note.

What Kai can do here

→Create events and recurring meetings from a sentence: 'book me with Sara next Tuesday afternoon.'
→Find open time across calendars and schedule meetings.
→Run meetings on Kaiday's video stack. No third-party app to install.
→Record meetings and transcribe them automatically.
→Turn a recording into action items and follow-up emails.
→Prepare prep notes before every meeting: who's coming, what changed since you last talked, what to ask.

Try saying to Kai

›Book a 30-minute methodology review and prep me with the study's assumptions and limitations.
›Record the investment committee meeting and send the action items to reviewers afterward.
›Find an open 60-minute slot Tuesday or Thursday for the model-risk review.

Apps

Hosted financial workflow apps live here. Kai customizes a code-backed starting point and publishes it under your domain. MiniApps do not process payments or execute trades.

Imagine this. Saturday morning, an idea, a working app by lunch

You need a review dashboard that puts benchmark, holdout, sensitivity results, limitations, and reviewer state on one screen.

At 9:14am you say to Kai: "build a study review dashboard at research.example.com with benchmark, temporal holdout, sensitivity, and decision status."

The dashboard is deployed to the research domain and the first completed study is ready for review.

What Kai can do here

→Build a MiniApp from a description: 'evidence library with citations and paper roles.'
→Host financial workflow apps under your verified domain (e.g., research.example.com).
→Update an app by talking: 'add transaction-cost assumptions to the review form.'
→Wire apps to approved research data and namespaced app storage.

Try saying to Kai

›Build a hypothesis intake with benchmark, costs, and rejection criteria.
›Add a limitation field and reviewer decision to the study dashboard.
›Publish the review app at research.example.com.

See the full MiniApps section.

Knowledge Base

Imagine this. The pitch deck you've been avoiding all week

What Kai can do here

→Create documents, sheets, and decks from a description.
→Edit existing documents. Rewrite a section, restructure an outline, add a chart.
→Generate presentations: investment-committee decks, portfolio reviews, board updates, and methodology training.
→Edit slides by talking: 'remove slide 4, tighten the headline on slide 6, add a competitive landscape slide.'
→Generate decks from meeting notes, study packages, or a single-sentence brief.
→Search every document by meaning, not keyword; cite documents inline when you ask a question in any other surface.

Try saying to Kai

›Generate a 10-slide investment-committee deck from the approved study package.
›Turn yesterday's methodology review transcript into a one-page decision memo.
›Add a sensitivity-analysis slide to the portfolio review deck. Pull from the Knowledge Base.

Code

Imagine this. Tuesday, 2:47pm

A researcher reports that the return series shifts by one day after a timezone change. You ask Kai to reproduce it and prepare a fix.

Kai reproduces the timestamp misalignment, traces it to the normalization boundary, and opens a draft pull request with the fix and a regression test.

You read the diff, merge, and reply to the reviewer that it's deployed. Total time on your side: four minutes.

What Kai can do here

→Open real draft pull requests on GitHub. Diff written, tests run, PR description filled in.
→Diagnose bugs from a reviewer report, a stack trace, or a failed study run.
→Read your codebase to answer technical questions: 'where is survivorship bias controlled?' 'why is this study slow?'
→Refactor when you ask: 'this file is too long, split it into modules.'
→Review teammate PRs and flag risks before you merge them.
→Never force-push. Never touch main without your merge.

Imagine this. A Sunday research review

You notice the factor chart needs a log-scale option. You don't want to open the laptop. You DM Kai from your phone: "add a log-scale toggle, remember the reviewer preference, draft PR only."

By the time you get home, the PR is waiting. You merge it Monday morning over coffee. The feature you'd have shipped "someday" is in production this week.

Try saying to Kai

›The daily returns shift around the DST boundary. Reproduce it and open a draft PR with a regression test.
›Add a log-scale toggle to the factor chart. Draft PR only.
›PR #421. Review it and tell me if anything's risky before I merge.
›Split the auth.ts file into modules. Draft PR.

Connections

Imagine this. Thursday 4pm, board update due Monday

You don't open the database. You don't export a CSV. You don't fight with the BI tool. You say to Kai:

"Pull last quarter's attribution, exposures, drawdowns, and benchmark deltas from production Postgres, and build a 12-slide portfolio review deck plus a backup sheet. Cite the queries."

What Kai can do here

→Answer ad-hoc questions against your production data in plain English.
→Turn data into the artifact you actually need. A deck, a sheet, an email, a doc, a slide for a meeting.
→Build recurring reports that auto-run on a schedule and land in Messaging or your inbox.
→Compare two market-data sources and flag mismatched observations.
→Cite the queries it ran, so the audit trail is real.
→Read-only by default; writes require explicit per-connection approval.

Imagine this. A committee asks a question mid-call

By the time the investor finishes their next sentence, the chart is on your screen with the underlying numbers. You share it. The call ends differently than it would have.

Try saying to Kai

›Pull Q3 attribution and risk metrics from production Postgres and build a 12-slide review deck. Cite the queries.
›Which EU positions breached their risk limits in the last seven days? List the source and as-of date.
›Every Friday at 5pm, post an exposure-and-drawdown digest to #portfolio-review.
›Active return by sleeve, top-decile positions, last four quarters. Chart it.

Control (admin)

Transparency and audit: every task Kai ran, every override, every memory edit. Admins only.

What Kai can do here

→Surface anything that looks unusual in the audit log.
→Answer 'what did you do last Thursday?' with the actual trace.

Try saying to Kai

›What did you do last Thursday between 2 and 5pm?
›Anything unusual in the audit log this week?

Connect data & code Automations