Hireboard
Get real engineering
Emergent Team Management by Hireboard

Cortex reads the work your team is already doing.

Cortex is Hireboard's Emergent Team Management (ETM) platform. It connects to the tools your team already works in — Slack, Jira, GitHub, Google Workspace and more — observes the activity that happens there, keeps every goal's progress up to date on its own, and escalates only the decisions that genuinely need a human.

Talk to Our Team
  • Read-only ingestion by default
  • Connect or disconnect any tool at any time
  • One isolated database per company
What Cortex is

A work layer that infers progress instead of asking for it

Most project tools only know what someone remembered to type into them. Cortex takes the opposite approach: it connects to the tools where the work actually happens, reads what changed, and maintains the goal tree itself. Humans are only interrupted when the system is genuinely unsure.

How it behaves

Executives define a goal once — the title, the requirements, and what "done" actually means.
Teams keep working where they already work. Nobody moves into a new project tool.
Cortex polls every connected tool on a schedule and reads the raw activity it finds there.
Each signal is matched to the goal tree, so a pull request, a Jira ticket and a Slack thread about the same work resolve to one item instead of three.
Progress, status and blockers roll up on their own — no manual status updates, no standup spreadsheet.
Anything ambiguous, low-confidence or conflicting is routed to a Decision Queue for a human to settle.
A continuous signal, end to end
  1. 1SlackA thread in #eng-team discusses splitting the API work in two.
  2. 2CortexInfers a subgoal under the existing "Launch v2 Public API" objective.
  3. 3GitHubPR #482 opens and is linked to that same subgoal — no duplicate ticket created.
  4. 4CortexConfidence is high, so progress rolls up automatically. Nobody was asked for an update.

Cortex never creates a new top-level objective on its own. It can only extend the tree your executives defined — adding a subgoal or a task under an existing goal — which keeps the plan recognisable to the people who own it.

How Cortex works

From connected tool to confident answer

Cortex runs the same five steps on every piece of activity it sees. The cheap, certain checks run first, AI runs last, and a human is only pulled in when the system cannot be confident on its own.

1

Step 1 — Connect the tools your team already uses:

A workspace admin connects each adapter once through a standard OAuth consent screen. Cortex only ever receives the access you approve at that moment, and each connection can be revoked from the workspace or from the provider at any time.
Thirteen adapters are available today, across engineering, comms, docs and CRM.
2

Step 2 — Define the objective, not the task list:

An executive creates a root goal with a title and a definition of done. Optionally, Cortex broadcasts that goal out to the connected tools — opening a tracking issue in GitHub, an epic in Jira, a brief doc in Google Docs, a channel post in Slack — so the team starts from the same place they always do.
From there, the team self-organises. Nobody is asked to maintain a second plan.
3

Step 3 — Cortex reads the raw activity, continuously:

On a fixed schedule, Cortex polls every connected tool for what changed since the last sync and stores that batch against your company. If a sync fails, the cursor is left untouched and the next run picks up from exactly the same point, so nothing is silently skipped and nothing is counted twice.
The kind of activity it reads: -
  • Pull requests, commits and merges
  • Ticket creation, sprint status and epic breakdowns
  • Thread discussions, assignments and hand-offs
  • Document edits, comments and brief updates
  • Deal stages and pipeline movement
4

Step 4 — Every signal is matched back to the goal tree:

Cortex tries the cheapest, most certain route first: if a record carries an identifier the tree already knows — a pull request URL, a Jira key, a CRM deal ID — it links instantly, with no AI involved at all. Free-text sources like chat and email pass a relevance check before anything else happens, so ordinary conversation never becomes a work item.
Only what survives those two filters reaches the AI linking step — and even then, it can extend the tree beneath an existing goal, never invent a new top-level objective.
5

Step 5 — Only real decisions reach a human:

Each inferred change is scored for confidence. Confident changes apply and roll up on their own. Anything low-confidence, blocked or conflicting is flagged and raised in the Decision Queue, where the responsible person approves it, rejects it, or corrects it with an instruction.
What lands in the queue: -
  • Data classification — is this signal really about this goal?
  • Assignment confusion — who actually owns this piece of work?
  • Plan breakdown confusion — where does this belong in the tree?
  • Goal proposals — new work that emerged and needs an owner
Adapters

The tools Cortex reads and writes

Every adapter is opt-in. Connect only the tools you want Cortex to observe, and disconnect any of them later without affecting the rest of your workspace.

  • GitHub

    Ingests PR opens, AI code reviews, commit pushes, and merge signals.

  • Slack

    Ingests thread discussions, assignments & team self-organization signals.

  • Jira

    Reads ticket creation, sprint status, and epic decomposition data.

  • HubSpot

    Tracks marketing/CRM deal stages, contacts, and pipeline movement.

  • Google Docs

    Reads doc edits, comments, and shared brief updates.

  • Confluence

    Reads page publishes, edits, and space activity.

  • Gmail

    Reads stakeholder email threads and status updates.

  • Google Meet

    Reads meeting records, attendance, and scheduled syncs.

  • Linear

    Reads issue state, team, and assignee activity across your workspace.

  • Salesforce

    Tracks CRM accounts, opportunities, and contact activity.

  • Microsoft Teams

    Syncs channel message history, and posts goal updates directly to channels.

  • Microsoft 365

    Syncs Outlook inbox messages, and drafts or sends goal updates via email.

  • WhatsApp

    Receives real-time inbound messages via Meta Webhooks and dispatches goal updates to WhatsApp numbers.

Your data

What Cortex reads — and what it does with it

Cortex reads only from the tools a workspace admin has connected, and only through each provider's own authorisation screen. Across every adapter — chat, issue trackers, code hosting, documents, email, calendars and CRM — it reads the same narrow slice: what changed since the last sync. In practice that means messages and threads, tickets and their status, pull requests and commits, document edits and comments, scheduled meetings, and deal stages.

All of it serves one purpose: keeping your goal tree current. A record is either matched to a goal or discarded — and free-text sources such as chat and email pass a relevance check first, so ordinary conversation never becomes a work item. Where Cortex writes back to a tool — a tracking issue, a channel post, a brief document, a kickoff email or meeting — it does so only when someone in your workspace explicitly chooses to broadcast a goal to that tool.

  • Your data is used only to power the features described on this page. It is never sold, and never used for advertising.
  • It is stored in your company’s own isolated database, with connection credentials encrypted at rest.
  • It is not transferred to anyone else except as needed to provide these features, to comply with applicable law, or as part of a merger or acquisition with notice.
  • People do not read it, except with your explicit consent, for security investigations, to comply with applicable law, or where it has been aggregated and anonymised.
  • Content that reaches the AI step is processed only to produce the result shown in your own workspace. It is not used to develop, improve or train generalised AI or machine-learning models.
  • Any adapter can be disconnected at any time, either from Cortex or from the provider’s own account settings — after which Cortex stops reading from it.

Full details are in our Privacy Policy and Terms and Conditions. Where an adapter connects to Google APIs, Cortex's use of that information also adheres to the Google API Services User Data Policy, including the Limited Use requirements.

Data & safeguards

Built to read your tools without owning them

Cortex sits on top of systems your team already trusts, so the boring parts matter: isolation, encryption, revocability, and never doing the same thing twice.

  • One isolated database per company

    Every workspace gets its own dedicated database. Goals, activity records, decision items and integration connections for one company are never stored alongside another company’s.

  • Credentials encrypted at rest

    Connection credentials and tenant database URIs are encrypted with AES-256 before they are stored, and decrypted only in memory when a sync actually runs.

  • Only the tools you connect

    Nothing is read until a workspace admin connects an adapter through that provider’s own OAuth consent screen — and disconnecting one adapter never affects the others.

  • Nothing is processed twice

    Every record carries a source reference, so a redelivered message or a retried sync is recognised and skipped. A genuine status change is still picked up; a duplicate is not.

  • Role-based visibility

    Workspace admins manage adapters and members. Executives own objectives. Members see the work they are part of. The decision queue routes each item to the role responsible for it.

  • The plan stays yours

    Cortex can add a subgoal or a task beneath an existing objective, but it can never create a new top-level goal — so the structure your leadership defined stays intact.

FAQs

Cortex FAQs

What is Cortex?Toggle
Cortex is Hireboard’s Emergent Team Management (ETM) platform. Instead of asking people to keep a project tool up to date, Cortex connects to the tools a team already uses — Slack, Jira, GitHub, Linear, Confluence, HubSpot, Salesforce, Google Workspace, Microsoft 365, Microsoft Teams and WhatsApp — reads the activity that happens there, keeps each goal’s progress current on its own, and raises only the decisions that genuinely need a person. It is operated by Fireboard Intelligence, LLC (DBA Hireboard).
How does Cortex know what my team is working on?Toggle
An executive defines a goal with a title and a definition of done. From then on, Cortex polls each connected tool on a schedule for what changed since the last sync. Each record is matched back to the goal tree: first by an identifier the tree already knows, such as a pull request URL, a Jira key or a CRM deal ID, which needs no AI at all. Free-text sources like chat and email pass a relevance check before anything else happens. Only what survives those filters reaches the AI linking step, and that step can extend the tree under an existing goal — it can never invent a new top-level objective.
What is the Decision Queue?Toggle
It is where Cortex sends anything it is not confident about. Every inferred change is scored; confident changes apply and roll up on their own, while low-confidence, blocked or conflicting ones are flagged and raised for a human. Queue items cover data classification (is this signal really about this goal?), assignment confusion (who owns this?), plan breakdown confusion (where does this belong?), and goal proposals (new work that needs an owner). The responsible person can approve, reject, or correct the inference with an instruction.
What data does Cortex read from a connected tool, and why?Toggle
Only what changed since the last sync, and only from tools a workspace admin has connected through that provider’s own authorisation screen. Depending on the adapter, that means messages and threads, tickets and their status, pull requests and commits, document edits and comments, scheduled meetings, or deal stages. It all serves one purpose: keeping your goal tree current, so a record is either matched to a goal or discarded. Cortex writes back to a tool — a tracking issue, a channel post, a brief document, a kickoff email or meeting — only when someone in your workspace explicitly chooses to broadcast a goal there. Where an adapter connects to Google APIs, our use of that information also adheres to the Google API Services User Data Policy, including the Limited Use requirements.
Does Cortex send data to an AI model?Toggle
Only where it is needed to make a linking or breakdown decision, and only after the cheaper checks have already failed to resolve it. That content is processed by Google Gemini models through Vertex AI purely to produce the result shown in your own workspace. It is not used to develop, improve or train generalised AI or machine-learning models, and it is not sold or used for advertising.
Is my company’s data kept separate from other customers?Toggle
Yes. Every company gets its own dedicated database — goals, activity records, decision items and integration connections are never stored alongside another company’s. Connection credentials and tenant database URIs are encrypted with AES-256 before storage and are only decrypted in memory when a sync runs.
Can we disconnect a tool later?Toggle
At any time. Each adapter is connected individually through that provider’s own OAuth consent screen, and disconnecting one has no effect on the others. Access can also be revoked directly from the provider’s own account settings, after which Cortex stops reading from it.
Will Cortex create duplicate tickets in our tools?Toggle
No. Cortex matches an incoming record against identifiers the goal tree already holds, so a Slack thread, a Jira ticket and a GitHub pull request about the same piece of work resolve to one item rather than three. Every record also carries a source reference, so a redelivered message or a retried sync is recognised and skipped, while a genuine status change is still picked up.
How do we get access to Cortex?Toggle
Cortex is currently offered to teams through Hireboard. Book a call with our team and we will walk through your tooling, what Cortex would read from it, and how a workspace is set up for your company.

See what Cortex reads from your stack

Book a walkthrough and we will map your current tools, show exactly what Cortex would observe in each of them, and set up a workspace for your company.

  • Works with the tools your team already uses
  • No status standups, no second plan to maintain
  • Every adapter is opt-in and can be revoked at any time
Book a Call with Our Team

Cortex is a product of Fireboard Intelligence, LLC (DBA Hireboard). Hireboard, Lago Vista, TX 78645 United States

Questions about Cortex or your data? support@hireboardhq.co · Privacy Policy · Terms and Conditions