How it works

Your team builds the context. Your choice of AI uses it.

The people who know what your data means are the ones who define it — in an easy-to-use app, not in code.

What's in the full context

Data, semantics, and process.

An agent needs more than your records. It needs the knowledge that isn't in any field, and it needs to know how your team does the work. ContextWorks manages these three aspects of the context, with semantic instructions throughout.

Entities

The records

Curated fields, rollups over child records, filtered related lists, the parent record, and computed signals — assembled in real time.

Briefs

The business knowledge

How you segment, what a stage means, your brand voice, what your policies are. The business knowledge that isn't in the data, but is critical understanding and guidance.

Skills

How the work gets done

The steps and processes your teams follow every day, so agents and AI tools can effectively collaborate and perform their tasks.

Instructions, across all three

Instructions provide the judgment your team applies without thinking: what a value means, when it matters, and how it may be used. They attach per field, per entity, and per domain — so the same field can read one way in a pipeline view and another in a renewals view.

Who writes them, and how

Example Field Instruction

“Health score below 60 warrants outreach. Sourced from the customer success platform, refreshed nightly — internal only, never quote to a customer.”

Building your context layer

An easy-to-use app for business users — not a job for developers.

The people who know the data and meaning best are the ones to build the context. ContextWorks helps you get started, and once it is live the app shows how your context is being used and performing.

  1. 01

    Build your entities

    Entity definitions start with your page layouts, and you choose which fields to keep, to remove, and to document with instructions. Rollups over child records, filtered related lists, and computed signals get added the same way.

  2. 02

    Add briefs and skills

    Briefs are the business knowledge that is outside the data — how you segment, what a stage means, your brand voice. Skills are how your team does the work, so every agent does it the same way.

  3. 03

    Preview with real records

    See exactly what an agent will receive, rendered from records that you choose.

  4. 04

    Define self-contained domains

    A domain is a self-contained lens: the set of entities, briefs and skills specific to a team or an agent. It allows Finance and Sales to look at the same records in the ways that are most relevant for their roles.

  5. 05

    Deploy to your AI tools and agents

    Pick where the domain is delivered — Agentforce, an MCP client, Prompt Builder, Flow, Apex. Adding another later is a connection, not a rebuild.

ContextWorks · Build

Opportunity 360

Active

A deal as RevWorks sees it. Reach for this when judging whether a deal is healthy, stuck, or about to slip.

Core record

Opportunity

The same object, at the depth the question needs.

List 6fields
Standard 15fields
Detailed 15fields

Enrichment

4
Deal Health scored
Sales Process status
Stage Progression trend
Activity Timeline recent

Computed on the way out. None of it is stored on the record.

Referenced object

Account embedded at standard level

Related objects

2
Products OpportunityLineItem 25 rows max
Contacts & Roles OpportunityContactRole 10 rows max

The two lists that matter, capped — not the whole tree.

Metrics

1
Open Tasks count of Tasks IsClosed = false

Counted in the query. The agent gets one number, not every task.

Building an entity: the object's fields on the left, the curated set on the right, and an instruction written against one of them.

Studio · Entities — Opportunity 360

1 / 3
Studio · Entities — Opportunity 360
Delivering your context

Where your users and agents already work.

You build the context once. Every AI tool and agent your team uses reads it, so the same question gets the same answer in all of them.

Agentforce Claude ChatGPT Slack Teams Any AI agent

Claude, over MCP

Reps and managers conversationally explore their data and manage their pipeline.

Agentforce

Service and sales agents answer customer questions from the same account context, inside the Einstein Trust Layer.

Prompt Builder

Outreach and summary templates draft from the record, using the curated fields and the guidance attached to them.

Flow

Automations flag at-risk renewals and route them to the right owner, using the same definition of at-risk.

See ContextWorks in action.

A short walkthrough of how context gets built, what an agent receives, and how it reaches the tools your team already uses.

Schedule a demo

Akoonu is trusted by

HoneywellThalesAvery DennisonTeamworksKOREStreetLight Data

We use cookies to understand how you use our site and improve your experience.