Factors.ai Product Updates and Releases

Accounts can now be graded against a fixed score threshold instead of percentile. Set your own Hot/Warm/Cool/Ice boundaries so categories hold their meaning over time rather than shifting when other accounts become more or less active.

Customers can now self-install the Factors Salesforce app using a direct installation link, without needing an AppExchange listing. Install for Admins Only, drag the factorsaiEmbed widget onto the Account page, and save.

Credit calculation now reflects LLM pricing for reused context, and the savings are passed to customers. The same Scout work generally consumes fewer credits, with longer conversations benefiting the most.

Scout can now show what changed in LinkedIn and Google Ads campaigns and when, with field-level diffs (before and after values) across campaign settings, budgets, bids, targeting, keywords, and more.

Milestones now uses actual LinkedIn impression and click volumes instead of grouping by ad group and day. This gives a more accurate picture of how much LinkedIn exposure it takes to drive conversion.

Press Ctrl+Shift+U (or Cmd+Shift+U on Mac) on any page to copy a shareable URL that includes the project ID. Recipients are automatically switched to the correct project when they open the link.

You no longer have to select the maximum number of Bombora topics allowed by your plan. Choose only the topics relevant to your business to reduce noise and improve enrichment quality.

When setting up a contact workflow in HubSpot or Salesforce, you can now choose Blitz as an alternative to Apollo. Blitz requires no API key and has no additional cost.

Ask Claude, ChatGPT, or any MCP-compatible LLM about a saved Factors report by name and it will fetch and return the data inline. Supports pulling multiple reports at once for combined analysis.

Every completed agent run now shows credits consumed, domain researched, AI model used, execution mode, status, timestamp, prompt sent, and full response, all in the Activity Log.

You can now regenerate your Factors MCP personal access token from Settings > AI Features. Regenerating immediately invalidates the previous token. Tokens cannot be regenerated again for 24 hours.

The Account Identification page now lets you choose between Factors Deanonymization and Bring Your Own Key in one place, with a new Waterfall Fallback option that tries additional providers when your selected provider cannot identify a visit.

Run a full sync of all accounts in a chosen Factors segment to Salesforce on a recurring weekly schedule, regardless of activity. Supports property mapping and an optional daily account cap. HubSpot support is next.

Account research agent outputs delivered to Slack now render tables, headers, and lists natively instead of arriving as plain text. Output structure is controlled by the format you define in the agent itself.

The Factors widget now embeds directly in Salesforce (Account object) and HubSpot (Company record). Sales teams can view the account journey timeline, Birdeye snapshot, and run account research agents without leaving the CRM.

Scout now shows a credits badge per conversation in the chat header. Settings > Plans & Billing > AI Credits shows a breakdown by teammate and by agents for the current calendar month.

Push any account signal (web research, social intent, custom events) into Factors via API. Signals now appear as named events on the account journey and work across segments, alerts, routines, agents, and reports. Supports up to 10,000 new accounts per request.

Failed agent runs now return specific, actionable error messages explaining the cause and next steps, plus a support reference code for faster diagnosis.

Describe what you want your agent to do and the AI Planner automatically selects the right events and properties. It updates as your instructions change and never overrides manual selections.

Turn any natural-language instruction into a scheduled report or a condition-based alert delivered to Slack, email, or API. Launches with 20+ ready-to-use instructions across GTM monitoring and paid media ops.

Web Research now uses Exa Search for faster, sharper results. A new Deep Web Research tool researches accounts in depth, reading full pages and pulling from specialized sources to surface hiring changes, funding rounds, and company initiatives.

Account search now matches on company name as well as domain, checking the resolved Factors name, Salesforce name, and HubSpot name. No more needing to know whether an account is keyed as acme.com or acme.io.

Automate LinkedIn bid changes on a schedule. Create rules with a time window and a bid increase or decrease by percentage; the bid reverts automatically when the window ends.

A new model selector in the agent builder lets you match the model to the job: heavier models like Claude Opus 4.6 for complex reasoning, lighter models like Claude Sonnet 4.5 for simple, well-scoped tasks.

Define exclusion rules once and apply them across multiple ad accounts and campaigns automatically, combining company exclusions, synced segments, and attribute-based exclusions like job title or seniority.

Control which website visitors get enriched instead of enriching every visitor by default. Choose to identify all accounts or set custom rules to focus enrichment, and your quota, on high-intent traffic.

Select multiple event types at once when filtering the timeline view, instead of reopening the dropdown and picking one option at a time for each event.

The Factors app is now officially listed on the HubSpot Marketplace and will be featured in the New Apps section for 30 days.

Admins can now control which users have access to Scout Chat and Scout Agents through role-based permissions, with independent View or Edit access levels for each.

Sandbox is live for all Scout users, letting Scout run code directly against datasets and increasing analysis capacity from roughly 1,000 rows to 250,000 rows for full account journeys and large exports.

Users can now run significantly more analyses, investigations, and visualizations with the same credits as before, letting credits last much longer without increasing spend.

Milestone Analysis now supports journeys with a single event, like becoming an MQL or requesting a demo, removing the need to add an extra event just to unlock analysis.

Agent runs that previously took around 5 minutes now complete in 1 to 2 minutes. A move from character-based to token-based context limits lets agents process up to 180 days of event data in a single pass without trimming.

Both views now default to the last 3 months of activity, with a new filter to switch between 30 days, 3/6/12 months, or all time. Fixes slow loads on accounts with long engagement histories.

Share any Scout conversation with your team via a read-only link. Chats stay private by default; switching to Team generates a shareable URL. Teammates can view full reports, canvases, and visualizations.

No more local package setup. Add the hosted MCP URL as a custom connector in Claude or any MCP-compatible client. Setup takes about 2 minutes with a personal access token.

Basic segments now build in 1 to 2 hours instead of 8 to 10. A complete rebuild of the processing pipeline means faster iteration and same-morning results after creating or editing a segment.

Report on which paid channels, campaigns, ad sets, and keywords influenced MQLs, SQLs, deals, and pipeline. Features a collapsible reporting structure and drill-downs on conversions.

Scout can now research beyond domain-based queries. Ask about competitors, market trends, product launches, and buying signals from across the web, powered by Perplexity.

Click Analyze on any milestone journey to get impact-ranked touchpoints, verdicts (Working / Weak / No Signal / Negative), and recommended actions. Follow-up questions let you explore specific transitions in depth.

New projects launch with a pre-configured signal library covering HubSpot, Salesforce, G2, LinkedIn, and website activity. Customers get meaningful account scores immediately after onboarding with no manual setup.

Collapsible months, switchable granularity (Month / Week / Day), inline milestones, event filtering, and CSV export. Each collapsed row surfaces the top event types at a glance.

Anonymous Users and Account Activity streams now appear as parallel columns against a shared timeline, eliminating the need to scroll past irrelevant sections.

Renamed based on customer feedback for clarity. Top Identified Users is now Top Identified Contacts to align with industry-standard terminology.

A new tab in the account side drawer lets you browse, search, and run all published agents for that account. Full-width output layout replaces the narrow column view in segment tables.

When a visitor submits a form with their work email, Factors enriches their profile with LinkedIn details including role, company, skills, bio, and connection count, via Blitz.

A client-side getData() call exposes IP-based firmographics and Factors segment membership to your GTM stack. Pipe segment data into VWO, Optimizely, or any personalization tool. Enabled per project via feature flag.

Set a ceiling on how much any single signal contributes to an account score regardless of frequency. Prevents noisy low-intent signals from drowning out high-intent activity like demo requests.

Every campaign-influenced touch (emails, landing pages, forms, CTAs, social broadcasts, and more) now syncs to Factors as events attributed to the right account. Filter by campaign in segments, agents, alerts, and attribution reports.

Deanonymize US-based B2B website visitors at the person level, surfacing name, title, company, work email, and LinkedIn URL. Available across Timeline, Agents, Segments, Reports, Alerts, and Workflows.

Use Segment's anonymous_id as an event property across Explain, KPIs, Events analysis, Segments, and Alerts, closing the loop between Segment's CDP identity and Factors account intelligence.

Sync users from a Factors segment to a LinkedIn Audience alongside company-level sync. Toggle company and user sync independently from the same rule. Live for all customers.

Takes a domain and returns a ranked list of champions (name, title, LinkedIn URL) via waterfall enrichment. Prioritizes C-level and VP-level marketers, then Directors, then Managers.

Mention @people_search in any agent to fetch enriched employee data for a given domain: names, titles, LinkedIn URLs, and work emails. Filter by country, continent, job level, and function before the agent runs.

Password-based sign-ins now require a verification code sent to email. SSO users are unaffected. Prevents password-breach scenarios reported by multiple customers.

Third Party Intent Visit renamed to Bombora Intent Surge retroactively. All historical data, segments, reports, and workflows referencing the old event name continue to work.

Run Factors account research agents from any MCP-compatible client: Claude, ChatGPT, Gemini, Cursor, Cowork. Chain with CRM, calendar, and Slack MCPs for end-to-end workflows without context switching.

Agents can now pull directly from Salesforce including firmographic data, deal stage, amount, close date, and custom fields, alongside the existing HubSpot tools.

Extended from 30 to 180 days to capture full enterprise sales cycles. Use targeted event filters for longer windows to keep agent output relevant.

Customers can now track exact credit consumption by Scout and Agents from two places in the product, addressing the top transparency request.

Agents, Assistant, and Apps (coming soon) are consolidated under the Scout umbrella, a first step toward them working together in a single flow.

Map any agent output (competitive signals, deal loss reasons, governance flags) directly to Salesforce account fields via Automations, Workflows, and the Salesforce account workflow.

Filter accounts based on topics they are actively researching across the web: Attribution Models, Company Surge, Demand Generation, Demandbase, and more. Available under account properties in the segment builder.

Agents and Scout now share a single credit pool. Recommended starting point for Scout and Agents customers is 100k credits. All older separate add-ons are archived.

Call any published agent from inside another agent. One outer agent can invoke multiple inner agents in a single run, enabling modular, composable research workflows with no manual stitching.

Build complex segments without workarounds. Previously limited to 1 OR per statement, now stack up to 10 across firmographics, engagement, and intent signals in a single readable group.

Duplicate any segment with all its filters in one click. Edit the copy for a new vertical, company size, or territory without recreating complex conditions from scratch.

Two one-click visualizations: an interactive account journey timeline (channel-coded events, buying signal spikes, milestones) and a seniority x function heatmap of the buying group. No setup required.

@Segments returns all segments a domain belongs to. @Part Of Segment is a boolean check for conditional agent logic: is this account in segment X? Both work alongside existing tools.

When an agent outputs a classification or label, that value is written back as an account property you can filter on. Combine agent output like Tier 1 or Disqualified with event data to build precise segments without manual tagging.

Use the Add or Update Company Object workflow to write any agent output (last funding round, competitive signals, lifecycle stage) directly into HubSpot company fields.

Hover over any account or company property across the product to see a plain-language description, data type, and example values. Covers All Accounts and Company Identification categories initially.

Click tools inside the editor to view and change their config. Each tool now includes a plain-language description. Testing view is separated from the editor with its own scroll so long outputs don't push the prompt offscreen.