How it works

Your AI team, built around how you actually work.

Kolig is a team of AI specialists, each with a role, skills, memory, and the tools they need to work alongside your people every day.

01

Build your team

Create AI roles from templates, clone from LinkedIn, or describe what you need. Kolig maps the skills and you pick the model, skill packs, and instructions.

02

Give them context

Upload company knowledge to Brain: docs, playbooks, guidelines. Scope access company-wide or by team so each role knows exactly what it should.

03

Put them to work

Chat one-on-one, run multi-role meetings, or message your roles in Slack and email. They respond from their role, pulling from Brain and memory.

AI Roles

Specialists that know their job.

Start from a role template for any specialty, clone straight from a LinkedIn profile, or describe what you need in your own words. Kolig understands and maps the skills automatically. Install skill packs, assign an AI model, write custom instructions, and each role is ready to go with the expertise it needs.

  • Role templates for any specialty
  • Clone any role from a LinkedIn profile
  • Skills mapped automatically from profiles and descriptions
  • Installable skill packs per role
  • Per-role AI model selection
  • Custom instructions and system prompts
Elena Vasquez

Elena Vasquez

AnthropicClaude Fable 5

VP of Marketing

Cloned from LinkedIn

Skills

Brand StrategyDemand GenerationMarketing AnalyticsContent StrategyGo-to-MarketPaid Media

Abilities

HubSpotSlackGmailSalesforceCalendarTelegram

Role Memory

ICP: Series B SaaS, 50–500 employees. Brand voice stays direct and proof-led.

Q3 paid ROAS held at 2.4×; lifecycle nurture owns 34% of pipeline. Enterprise messaging leads with SSO and audit readiness.

Meetings

Your whole team, one room.

Start a meeting with any combination of roles and each one responds from its own perspective: Product, Legal, Engineering, Marketing weighing in at once. Kolig auto-generates the summary and assigns action items so nothing gets lost.

  • Any combination of roles in one meeting
  • Sequential responses with role attribution
  • Real-time streaming per participant
  • Auto-generated meeting summaries
  • Action items extracted automatically
Sam, Product Manager
Morgan, Legal Counsel
Alex, Engineer
Jordan, Copywriter
Elena, VP Marketing

Product Launch Review

Meeting · 5 participants

Live
@sam @legal Can we launch Monday?
Sam, Product Manager

Sam

Product Manager

SSO is ready. QA passes Friday.
Morgan, Legal Counsel

Morgan

Legal Counsel

Terms are approved. No blockers.
Alex, Engineer

Alex

Engineer

One fix left. I'll ship it today.

Brain

Shared knowledge for every role.

Upload PDFs, Markdown, or any other file format. Scope knowledge company-wide or by team. Legal's playbooks stay with Legal, brand guidelines go everywhere. Roles pull from Brain automatically when they answer.

  • Upload PDF, Markdown, or any file format
  • Company-wide or team-scoped access
  • Roles cite the source in every answer
  • Document management dashboard
Marketing
Product
Engineering
Legal
Brain
  • PDF
    Brand guidelines
  • MD
    Q3 roadmapCompany-wide
  • DOCX
    Refund policyCompany-wide
  • DOCX
    Security policy
Morgan
Elena
Sam

Policy & Brand Review

Meeting · 3 participants

End meeting
Where is our refund policy documented?
Morgan, Legal Counsel

Morgan

Legal Counsel

Full refunds are available within 30 days of purchase. Pro-rated refunds apply after that window for annual plans.

From Brain ·DOCXRefund policy.docx, section 4.2

What are our logo usage rules?
Elena, VP of Marketing

Elena

VP of Marketing

Logo must have 24px clear space. Never stretch or recolor outside brand palette.

From Brain ·PDFBrand guidelines.pdf, page 12

Multi-Provider AI

The right model for every role.

Not every task needs the same AI. Assign Anthropic Claude, OpenAI GPT, Google Gemini, or Grok per role, independently. Your Security Engineer gets Claude Opus for depth. Your Support Agent gets GPT-4o Mini for speed. You pick, Kolig handles the rest.

  • Anthropic, OpenAI, Google, Grok
  • Economy, Standard, and Premium tiers
  • Per-role model selection
  • Speed, quality, and cost indicators

AI Models

Choose model per role

Anthropic

  • Claude Sonnet 4.6Selected
  • Claude Fable 5
  • Claude Opus 4.6

OpenAI

  • GPT-4o
  • GPT-4o Mini

Google

  • Gemini 1.5 Pro
  • Gemini 1.5 Flash

Grok

  • Grok 3
  • Grok 3 Mini
  • Grok 2

Integrations

Connect to the tools you use.

Your AI roles can receive messages in Slack, respond to emails, read from HubSpot, and query Google Analytics without leaving Kolig. Each integration is an ability you plug into a role. Add Slack to your Customer Success role and it replies to support threads directly.

  • Plug abilities into any role
  • Receive and reply in chat channels
  • Dedicated inbound email per role
  • Read and update CRM records
  • Query analytics from your data stack

Workspace integrations

Connect tools · assign per role

  • Slack

  • Microsoft Teams

  • Google Calendar

  • Salesforce

  • HubSpot

  • Gmail

  • LinkedIn

  • Telegram

  • WhatsApp

  • Google Analytics

  • Mixpanel

7 connected · roles use abilities in Slack, CRM, email & more

Role Memory

Roles that remember.

After every conversation, each role builds a compressed memory of how you work together. Your ICP, brand voice preferences, recurring decisions, all retained. No context bloat, no repeating yourself. The longer you work with a role, the more useful it becomes.

  • Compressed memory per role
  • Updated after every conversation
  • Rebuild or clear anytime
  • Combined with Brain context
  • Persistent across sessions
Elena Vasquez

Elena Vasquez

VP of Marketing · Role Memory

~134 tokens
  • ICP

    Series B SaaS, 50–500 employees. Decision-maker is VP of Engineering or CTO.

  • Brand voice

    Direct and proof-led. Avoid jargon. Never use exclamation points in copy.

  • Pipeline

    Q3 paid ROAS held at 2.4×. Lifecycle nurture owns 34% of pipeline.

  • Preferences

    Prefers bullet summaries over prose. Always include sources in answers.

  • Decisions

    Enterprise tier leads with SSO messaging. Avoid discounting below 15% without CFO sign-off.

  • Recent context

    Launching Product Hunt campaign next Tuesday. A/B test headlines before paid spend.

Updated after every conversation · compressed context, no bloat

Still have questions?

Browse the docs for a full breakdown of every feature, or reach out and we'll walk you through it.

Ready to build your AI team?

Start with 3 roles for free. No credit card required.