decisions at scale

does your business automate decisions?

we do!

we wish!

As you know, systems that automate

lending · fraud flagging · transaction risk assessment · lead prioritization · pricing ·  search ranking · matching & dispatch ·  product recommendations · trading · . . .

are fragile and opaque, often losing money and trust.

Untitled drawing (3) (1).png

decision diagnostics

AI that learns the full data graph behind every automated decision in production and alerts precisely what to fix to recover lost value.

spot complex graph


  • decimates time to discover and fix valuable issues

  • guards against decay due to

    • evolving markets​

    • fragile pipelines

    • imprecise definitions

    • swapped data sources

    • unstable logging

  • lets you deploy models and algos with confidence

  • frees several Eng/DS to invent, not reinvent

  • increases trust in your Data Org

  • fills Eng/Data/Prod accountability gap

  • last line of defense, watching the watchmen


  • alerts as data issues are discovered

  • estimates of business value of each issue

  • graphical view of global production data

  • spans batch/async pipelines, too

  • installs simply

  • recalls full historical data context

  • data organized ideally for analysis

  • interact through UI, API, and CLI

Smart business needs smart decisions:

  • which transactions to block?

  • what prices to display?

  • which customers to target?

But staffing Data Science teams and ML experts is an expensive gamble.


decisions as a service

AI-generated decisions that optimize your key business metrics.



  • decimates time to scaling good decisions

  • lets you inspect the full rationale for each decision

  • easy to call or integrate with existing codebase

  • eliminates cost of expensive DS & Eng talent

  • eliminates hidden of devops to maintain uptime

  • continuously optimizes for your biz objectives

  • promotes clarity of business objectives

  • no-hassle

  • organizes data for easy analysis


  • handles virtually all business objectives

  • works for most decision "levers"

  • self-optimizes

  • reports performance

  • will eventually run own AB tests

  • recalls full historical data context

  • data organized ideally for analysis

Existing solutions aren't solutions.

data quality monitoring:

searches the wrong haystack.

Data quality monitors look for changes in stored data, mostly unrelated to strategic, automated decisions.

dcyd hones in on data relevant to decisions that move the business.

model monitoring:

misses the big picture.

Model monitoring sees isolated pieces of the puzzle, and can't connect local performance to business objectives.


dcyd connects data all the way from sources to decisions.

anomaly detection:

too noisy, too needy.

Anomaly detection has no sense of priority and wakes you up for nothing.

dcyd alerts only when anomalies are impacting your business objectives.

is the future of business decision-making @scale.
  • observes data flowing in production.

  • learns behaviors and data dependencies.

  • analyzes the impact on business objectives.

  • alerts only business-critical issues.

better decisions at scale

let's get started

good dcysion!