Bigmind

What’s new in Bigmind

Every update to Bigmind, newest first. Page 4 of 10.

Introducing Automations, the new home for everything Bigmind does in the background

The recurring jobs Bigmind runs for you (post-meeting summaries, daily digests, signal-driven follow-ups, pipeline alerts, account updates, and everything else that fires without you opening a chat) used to live under a concept called Background modes.

Background modes are now Automations

They were tucked inside individual agents, hard to find, and harder to reason about as a whole.

Today we're pulling all of that into a first-class concept: Automations. Same jobs, same triggers, same outputs, but with a clear home, clear inputs, and clear results.

What you get with Automations

One place to see everything that runs for you – A single Automations page lists every automation in your workspace: what it's triggered by (a meeting ending, a signal firing, a schedule, a deal updating), what it produces (a todo, a document, a CRM update, a message), and whether it's active.

Customize per organization – Every automation has sensible defaults out of the box. Override any of them at the organization level (system prompt, tools, schedule, scope) and your version takes over for everyone in the workspace.

Run any automation on demand – Pair this with the Run now dialog we shipped last week and you can simulate any automation against a real scenario before it ever fires for real, so building sophisticated flows no longer means deploying and waiting.

What's next

This release is the start of a much bigger push to make automations easier to build, both personally (your own automations for your own workflows, like "ping me when this account hasn't had a meeting in 30 days") and organization-wide (shared automations that everyone in the workspace benefits from). The infrastructure is in place. The next few releases will focus on making authoring an automation feel as fast as writing a Slack message.

Find your Automations under Settings, AI, Automations. If you had Background modes configured before, they're already there—nothing to migrate, nothing to re-set-up.

Simulate any automation trigger with Run now

Building a good automation usually means iterating: tweak the prompt, change which tools the agent has, refine when it should fire.

Stop deploying just to see what happens

The painful part has always been the loop in between, where you have to wait for a real trigger to come along (a meeting that ends, a signal that fires, a deal that updates) before you can see whether your change actually worked.

Today we're introducing Run now: a built-in dialog on every automation that lets you fire it on demand against a specific scenario. No more deploying-and-waiting.

How it works

From any automation's settings page, hit Run now. You'll be prompted to pick the context the automation should run against (a specific meeting, account, deal, signal, or scheduled tick, depending on the trigger type), tweak any input you want to override, and execute. The automation runs end-to-end against real data, and you see exactly what it produced (todos created, fields updated, documents written, messages sent) before the next real trigger ever fires.

Iterate fast on prompts and tools – Edit the system prompt, hit Run now against the same context, see what changed. Tighten the loop from days to seconds.

Test edge cases on real data – Pick the gnarly account or the deal everyone's worried about and run the automation against it specifically, instead of waiting to see what happens in production.

Validate before rollout – When you're about to enable an automation across the whole org, simulate it against a handful of representative scenarios first. Catch the surprises before your team does.

Run now is available today on every automation, custom or out of the box, under Settings, AI, Automations.

Email signatures for outbound mail sent through Bigmind

Every rep had a signature in Gmail, but emails drafted and sent from Bigmind arrived without one — until now.

Your signature, on every email you send from Bigmind

When reps send email through the Send Email tool in chat, the message should look like it came from them — not from a blank template. Until now, signatures lived in Gmail or Google Workspace and did not carry over.

You can now configure a plain-text and HTML signature per email account. Bigmind appends it automatically when an email goes out through the Send Email tool: plain text for text emails, HTML (with image upload support) for HTML emails.

Per-account control – Set signatures on org-wide email accounts under Settings, Engage, Email accounts, or on personal accounts under Settings, Profile, Email accounts.

Preview before send – The Send Email composer shows how your signature will render so you can spot formatting issues before the message leaves your outbox.

Configure your first signature under Settings, Engage, Email accounts.

Introducing Skills, focused playbooks the AI Agent loads on demand

With one agent now handling everything, the next question was: how do you keep the depth of guidance you used to spread across multiple specialized agents?

Tailored guidance, without the maintenance overhead

The answer is Skills.

A Skill is a small, focused instruction packet for a specific class of request. The AI Agent picks the right one for what you're doing, loads it on demand, and follows it. You get the steerability of the old multi-agent setup without having to maintain six different agents and modes that all overlap.

Every Skill lives under Settings, AI, Skills. You can read them, edit them, fork them, or add your own. Skills you haven't customized keep tracking the latest version we ship, so you don't fall behind on improvements just because you wrote a tweak six months ago.

What ships out of the box

Every workspace gets a curated library of Skills covering the most common sales workflows:

  • Account summary – A quick read on an account: who they are, who you know, recent activity, and open opportunities.
  • Account outreach – Drafting an email or planning outreach to contacts at a specific account.
  • Account plan writer – Creating an account plan from a library template.
  • Contact summary – A quick read on a single contact: their role, your relationship history, and what to talk about next.
  • Deal coaching – Moving a specific deal forward: current state, risks, and the next action.
  • Pipeline summary – Reviewing a pipeline (yours, a report's, or the org's) including risk, commits, and what to push on.
  • Meeting prep – Preparing for an upcoming meeting with context on attendees, prior history, and talking points.
  • Meeting follow-up – Recapping a meeting, drafting action items, or composing a follow-up email.
  • Meeting coaching analysis – Manager-led coaching on a meeting a rep ran: strengths, gaps, and a specific coaching plan grounded in the transcript.
  • Meeting self-review – Coaching feedback on your own meeting: what went well, what to improve, where you lost or won the room.
  • Meeting reflection – A personal reflection on a meeting you ran: what went well, what to improve, what to try next time.
  • Enablement lookup – Process, methodology, product, or "how do I" questions answered from your company library.
  • Library document editor – Searching, writing, or editing library documents like playbooks, ICPs, templates, battlecards, or process docs.
  • Website overview – A quick read on a company while browsing their website.
  • Custom tool builder – Building, editing, or refining a custom tool definition.
  • Mini-app builder – Building, editing, or testing a Bigmind mini-app.

Make them yours

Skills are the place where your team's specific guidance lives: the framework you use to coach reps, the checklist that goes into an account plan, the order you research a new account in, the rules for which knowledge source to consult first. Edit any built-in Skill in place to add your own context, fork one to create a variant, or write a brand-new Skill from scratch for a workflow we haven't covered. Each one is a one-line edit, not a redesign.

Sync WhatsApp Web conversations to Bigmind

Bring customer WhatsApp conversations into Bigmind from WhatsApp Web, so important context can live alongside meetings, emails, notes, todos, and CRM activity.

Customer conversations do not all happen in email

Many commercial teams use WhatsApp for the quick customer conversations that happen between meetings: confirming next steps, chasing an answer, sharing a document, or clarifying what was promised. Those messages often matter, but they used to stay outside the account timeline unless someone copied them over by hand.

Now you can sync WhatsApp Web conversations into Bigmind with the Bigmind browser extension. Open a WhatsApp thread, preview the visible messages, choose what should be brought into Bigmind, and sync it as account activity.

How it works in WhatsApp Web

When the WhatsApp integration is connected for your workspace, the Bigmind extension adds a sync action inside WhatsApp Web. Open a conversation, click the Bigmind button, and you will see a preview of the messages currently loaded in the thread.

From there, you can select all messages, clear the selection, or choose only the messages that should be synced. If you need older context, scroll back in WhatsApp first so those messages are visible before opening the sync preview.

What gets stored in Bigmind

The conversation thread - Bigmind stores the synced WhatsApp thread and its selected messages as communication activity.

Message details - Synced messages include the visible text, direction, author label, and timestamp available from WhatsApp Web.

CRM attribution where possible - When Bigmind can identify a phone number for the conversation, it can associate the activity with the matching contact, lead, account, or deal based on your attribution settings.

Group chat support - For group conversations, Bigmind lets you choose the participant to sync against when WhatsApp exposes a usable phone number.

Where to find it

First, connect the WhatsApp integration under Settings, Communication data, WhatsApp. Then open WhatsApp Web in the same browser where the Bigmind extension is installed and signed in.

The sync action appears in the WhatsApp conversation header when a thread is open. This is separate from Bigmind on WhatsApp: syncing WhatsApp Web messages brings existing customer conversations into Bigmind activity, while Bigmind on WhatsApp lets users message the Bigmind AI Agent directly.

A more seamless agentic experience

When we first built the agentic experience in Bigmind, we leaned on specialization: a meeting agent for prep and recap, a deal agent for coaching, an account agent for summaries and outreach, an enablement agent for lookups, and an Auto agent on top that routed your request to the right sub-agent.

From a router with sub-agents to a single capable agent

That shape made sense at the time. LLMs were less capable on their own, the surface area we needed to cover was narrower, and giving each domain its own dedicated agent made it easier to specialize prompts, tools, and behavior per workflow.

What that design gave us in clarity, it took back in how chats actually felt. Routing was visible and clunky—you'd see the Auto agent hand off, then a sub-agent take over, sometimes hand off again. Conversations that crossed domains (a meeting question that became a deal question that became a CRM lookup) didn't always flow naturally. And under the hood, the same guidance often had to live in two or three places to be available wherever a relevant request might land.

Both sides of that equation have changed. Models are stronger and more reliable orchestrators, and Bigmind's tooling has matured to the point where a single capable agent, well-guided, can do what we used to spread across five.

What's different today

Chats now happen with a single AI Agent. It handles meeting prep, deal coaching, pipeline review, account summaries, drafting a handover, looking something up in your knowledge base—all in one place, without visible handoffs.

Suggested prompts wherever you start – When you open the chat, you'll see a curated set of conversation starters tied to the page you're on. Looking at a deal? You'll see deal-shaped prompts. On an account? Account ones. The most common workflows are now one click away.

A lighter, faster chat under the hood – As part of the same rollout we've shipped a leaner chat experience. Conversations open faster, stream more reliably, and hold up better on long sessions. You shouldn't have to think about it—that's the point—but it's the foundation everything else sits on.

Custom agents are still fully supported

Specialization still matters for some workflows. If a use case really does need its own agent—a different model, a custom system prompt, a locked-down tool list, dedicated modes—that's a first-class option. You can spin one up under Settings, AI, Agents whenever it warrants it. The single AI Agent is the new default, but it's not a ceiling.

Subscription rules to automatically manage who follows signals

Previously, staying on top of signal activity for an account required each user to manually subscribe to the objects they cared about.

Let admins define who's subscribed to signals—at scale

If you wanted every account owner to automatically follow signals on their accounts, there was no way to set that up—it was entirely self-serve and easy to forget.

Now, admins can create subscription rules that automatically subscribe the right people to signals across your CRM. Set the recipient to "CRM Owner" and every account, deal, or lead owner is automatically subscribed to signals on the objects they own—no manual action needed. Or pick "Specific Users" to ensure a manager or team lead is always subscribed to signals across a set of records.

Each rule also defines channel (email, in-app message, or both) and timing (immediate or daily digest), so admins control not just who is subscribed but how they receive updates.

Granular signal filtering. Rules include a signal filter section to scope subscriptions precisely. Pick specific signals so the rule only applies to events you care about—like "Contract Renewal" or "Product Usage Drop." Filter by insight type (opportunity, risk, action, or information) to target different teams. And set a minimum urgency threshold so subscribers are only notified when it truly matters.

Users can still personally subscribe to individual CRM objects—with the same filtering options—for anything not covered by admin rules. The system combines both, so everyone is subscribed to exactly what's relevant.

Head to Settings → Automations → Subscription Rules to create your first rule.

Signal-driven AI agents that react to events in real time

Previously, signals generated insights that appeared on account and deal pages—but acting on them was still a manual step.

Turn signals into automated actions with background agents

You had to read the insight, decide what to do, and then take action yourself. There was no way to have the AI automatically respond to incoming signal data.

Now, signals can spawn AI agents. When you configure a background mode with the Signal use case, that agent will automatically run every time new signal events arrive for an account. The agent processes the signals in context—understanding the account's full history, CRM data, and prior interactions—and takes action using the tools you've assigned.

What can signal-driven agents do?

Create todos – Automatically generate follow-up tasks when a risk signal fires or a customer hits a usage milestone.

Update account overviews – Keep account summaries current by having the agent write new insights directly into the account overview.

Trigger any tool – Equip the background agent with any tool in your toolkit—CRM lookups, enrichment, document creation, or custom tools—and let it orchestrate a response.

To set this up, go to your agent's settings, create or edit a background mode, and enable the Signal use case. Assign the tools you want the agent to use, write a system prompt that instructs how to react to different signal types, and you're done. Every new signal event for every account will be processed automatically.

Revamped Todos with Kanban, grouping, and batch actions

Previously, the Todos view was a simple flat list—functional, but limited.

A complete redesign of how you manage your tasks

As your task count grew, it became harder to find what needed attention, prioritize effectively across accounts, or quickly clear a batch of completed items.

Now, Todos has been completely redesigned. Switch between list view and kanban view to work the way you prefer. Kanban gives you a drag-and-drop board with columns for Backlog, To Do, In Progress, and Done—making it easy to visualize your pipeline of work at a glance.

Group by anything. Organize your todos by status, account, priority, type, assignee, or due date. Grouping by account is especially useful for seeing all outstanding tasks for a specific customer before a meeting or review.

Filter by status and account. Quickly narrow your view to only in-progress items, or filter to a specific account to focus your efforts. Active filters display as chips that can be cleared in one click.

Batch resolve. Select multiple completed todos and mark them as done in a single action—no more resolving one by one.

The new Todos view is designed to scale with your workload, whether you manage ten tasks or ten thousand.

Signals Ingestion API for custom data sources

Previously, Bigmind's signal intelligence was limited to data it could pull from your existing integrations—CRM, email, calendar, and calls.

Send any event from your application into Bigmind

If your team had valuable product usage data, customer health metrics, or operational events locked inside internal systems, there was no way to feed that context to your AI agents.

Now, the Signals Ingestion API lets you push structured events, metrics, and state changes directly into Bigmind from any source. Send a signal when a customer visits your pricing page, when a support ticket is escalated, when product usage drops below a threshold, or when a contract is renewed—and Bigmind's AI will analyze each event in the context of the associated CRM record to generate actionable insights.

The API supports three signal types: events for discrete occurrences, metrics for numeric values aggregated over time, and state signals for tracking current conditions like health scores or integration status. Each signal can have a JSON Schema for payload validation, with field descriptions that tell the AI how to interpret the data.

Create signals, define schemas, and configure trackers from Settings → Enrichment → Signals, then start sending data with a simple REST call. Full documentation is available in the developer docs.