People and agents work from the same customer records

Where teams get stuck handing CRM work to agents

1. Our agents have no way to read or write customer data

   Your AI tools and your own agents work in Trackit directly

   AI tools like Claude and Cursor connect with a URL and a sign-in, and the agents you built connect through the CLI and the public API. Agents find customers, add records in bulk and update values. They see only what the connected person can see.

   Related features

   • MCP
   AI chat

   Add this week's 6 web form leads and link them to companies

   records.create-batch

   Written to Trackit · 6 leads

   Added 6 leads and linked each one to its company record.

   Message

   Leads

   NameCompanySource

   Seojun LeeNaverWeb form

   Mina ChoiLG ElectronicsWeb form

   Haneul JungTossWeb form

   Doyun HanKakaoWeb form

   Jisoo OhSK hynixWeb form

   Chaewon YoonHyundai MotorWeb form

2. Every agent and every person judges by a different standard

   Write your team's guidelines once, and every call follows them

   Give a custom agent its role and the order to reason in, and each new lead is checked the same way against records, emails, notes and the web. The next block saves the result to the customer record as a score or a note.

   Related features

   • Custom Agents
   Lead scoring agent

   Guidelines

   Score inbound leads by our team's criteria

   1. 1Company size
   2. 2Company email domain
   3. 3Specific inquiry

   Web searchOn

   New leads

   LeadScore

   Yujin KangSamsung Electronics

   92

   Reason · Enterprise · work email · clear scope

   Minho SeoToss

   74

   Reason · Work email · seat count only

   DDaeun BaePersonal

   31

   Reason · Personal email · one-line ask

3. Our own agents and internal systems live apart from the CRM

   Your systems plug in at both ends of a workflow

   Start a workflow from data your systems send by webhook, and call external APIs with the HTTP request block to fill records with the response. Put AI blocks and custom agents in between to classify, summarize and decide in one flow.

   Related features

   • AI Workflows
   • Integrations
   Your service→{ "event": "signup", "domain": "lge.com" }

   Run results

   1. Webhook receivedYour service → Trackit
      LG Electronics · sign-up
   2. Custom agentClassify the inquiry
      Evaluating
   3. HTTP requestPOST internal API /accounts
      200 · account found
   4. Create or update recordLead · match by company domain
      LG Electronics lead updated

How teams use it

Customer stories

• 

  An AI agent sells through the Trackit MCP

  “We connected the Trackit MCP to our Hermes agent. It qualifies leads and runs outbound sequences.”

  NotiflyMinyong Lee, CEO · Kyungseo JungRead the story(opens in a new tab)
• 

  From sales management to customer support

  “Trackit helped us set up the CRM we needed for Naver Place, and now MCP takes us into customer support.”

  Place&Unsik Oh, Team LeadRead the story(opens in a new tab)

Use cases

• 
  B2B SaaS

  Routing leads to owners by score

  An agent scores leads from web forms and the meeting scheduler, then hands each to an owner on the matching team with a call request or a first email.

• 
  Sales organization

  Weekly sales reports written by an agent

  Every Monday morning, an agent reads last week’s sales activity and leaves a team report.

• 
  Public bidding team

  From bid notice to tax invoice

  AI reads each notice and picks the ones that fit, then the contract and tax invoice land on the same record after you win.