Semarize
Start building

How it works

Build conversational intelligence with Semarize

Semarize is the Conversational Intelligence API that turns your sales calls, customer emails, and support chats into structured fields your CRM, BI tool, or automation can read. Define how you want calls assessed into Bricks, bundle them into dedicated call review Kits, and call the API to get the same fields back on every conversation.

Start buildingExplore the docs
Bricks into Kits One API call Same fields back
SKitsDiscovery Signals240 credits

Discovery Signals Kit

DraftReady
kit_code6AD3F0B6versionv2.3

Kit composition

Build this Kit from ordered blocks.

Discovery block
Brick Group
Budget confirmedbooleanCustom

Did the buyer confirm budget is allocated?

MEDDICC scorescoreCustom

Overall qualification strength, 0 to 100.

Next stepcategoryCustom

The agreed next action after the call.

StakeholderslistCustom

People named as part of the decision.

Add Brick Add Kit
5 credits per runSave draft

Building in Semarize

Bricks · Kits · Call

Create reusable bricks, bundle them into kits, and call the API once. Structured signals come back on every run, no engineering team required.

Step 1
Build reusable Bricks
A Brick is one typed question about a conversation - one input, one consistent answer. Define it once, run it across every call, email, and chat that flows through Semarize.
  • Returns a yes/no, score, category, or list, not a paragraph
  • Works on sales calls, customer emails, support chats, and meeting transcripts
  • Versionable and reusable across teams, segments, and playbooks
Brick library
Reusable semantic checks
Next step agreed
Confirms a clear next action and owner.
boolean
Stakeholders identified
Detects roles and decision makers.
score 0-100
Pain is specific
Checks for measurable pain, not vague interest.
score 0-100
Timeline mentioned
Finds timing signals and urgency.
string_list
Pricing discussed
Detects pricing talk and captures context.
boolean + evidence
Objections raised
Identifies objection themes mentioned.
string_list
Step 2
Bundle Bricks into Kits
A Kit is a set of Bricks grouped for a specific workflow - a discovery rubric, a QA scorecard, a renewal call audit. Run the same Kit on every conversation in that workflow and you get directly comparable results.
  • Mix and match Bricks per stage, team, or playbook
  • Standardise scoring across reps, segments, and regions
  • Reuse Kits across products and pipelines without rebuilding
Kits
Stacks of bricks per workflow
Inbound Qualification Kit
Turns first calls into routing signals.
kit
Pain is specific
score 0-100
Timeline mentioned
string_list
Next step agreed
boolean
Stakeholders identified
score 0-100
Discovery Quality Kit
Scores how well the rep ran discovery.
kit
Pain is specific
score 0-100
Stakeholders identified
score 0-100
Objections raised
string_list
Next step agreed
boolean
Step 3
Run a Kit with one API call
Send a transcript and a Kit code. Semarize runs every Brick in the Kit against the content and returns the result. Send transcripts when they're ready by webhook, by polling on a schedule, or by pushing them straight from your stack.
  • Inputs: a transcript plus the Kit code
  • Three delivery modes: webhook, polling, or direct push
  • Outputs: structured JSON with value, confidence, and evidence per Brick
API call
input + kit_code
POST /v1/runs
{
"kit_code": "A1B2C3D4",
"input": { "...transcript..." }
}

Your input + a kit code.

Response
{
"run_id": "run_...",
"status": "succeeded",
"output": { "bricks": { ... } }
}

Predictable schema for downstream systems.

Step 4
Same fields back, every time
Every Run returns the same shape. The Bricks you defined come back as fields your CRM, BI tool, or automation can read on the first try, with the exact transcript line that supported each answer attached.
  • run_id and kit_version_id for auditability
  • output.bricks: value, reason, evidence, and confidence
  • Push into Salesforce, HubSpot, Snowflake, BigQuery, Zapier, Make, or n8n
Structured response
output.bricks + evidence
output.bricks
"next_step": {
"value": true,
"confidence": 0.93,
"reason": "Next action and owner agreed",
"evidence": ["...we'll send the proposal Friday..."]
}
What you can do next
Automation
If next step missing - create follow-up task.
Reporting
Trend discovery depth by segment and rep.
Action
Route high intent leads to the right owner.

Inputs

Connects to the tools that produce your transcripts

Semarize sits on top of your existing capture layer. Anything that produces text becomes a Run.

Note-takers and diallers

Meeting bots, outbound calls, and sales activity.

Meeting platforms

Video calls and webinar recordings.

Revenue platforms

Call libraries and sales engagement activity.

Semarize doesn't record, join meetings, or replace your capture tools. It processes the transcripts they already produce.

Outcomes

Signals that drive action

Structured outputs are easy to automate. These are the patterns RevOps and data teams build on Semarize signals to move CRM enrichment, forecast accuracy, and QA coverage.

Ops and RevOps

Turn signals into the next action.

MEDDICC.score < 70
Create "fix gaps" task and notify manager
next_step.exists = false
Auto-email rep coaching prompt
competitor_mentioned = true
Add battlecard to Slack thread

CRM enrichment

Write the fields reps never get to.

Stakeholders extracted
Update buying committee in CRM
Budget confirmed
Set budget_status = confirmed
Timeline mentioned
Adjust close_date on opportunity

Data and analytics

Trend, correlate, and forecast on real evidence.

Brick scores collected
Trend by rep, team, segment, region
Skill signals tracked
Correlate with conversion rates
Risk flags aggregated
Build early warning dashboards

Ready to build

Ready to build?

Start with Bricks, assemble Kits, and get structured signals back from your first conversation in minutes.

No card required. Start testing in minutes.

Start buildingExplore the docs

Questions

Frequently asked questions

Common questions about API-first conversation intelligence, ingestion, and how Semarize compares to the platforms teams already use.

A Conversational Intelligence API turns recorded conversations - calls, meetings, emails, support chats - into structured fields your other tools can read. Where a conversation intelligence platform gives a person a dashboard, an API gives a system a consistent answer. Semarize is that API layer. You define what to evaluate as Bricks, group them into Kits, and run a Kit against any transcript to get the same fields back every time.

From the blog

Further reading

Practical pieces for teams turning conversation content into Bricks, Kits, API runs, and structured JSON workflows.

Browse all posts