AI driven Natural Language Queries to access any of your systems and data

...without code

And query it all using ChatterBox, Slack, Teams or API calls.

Use Cases

Ziggy helps teams access data anywhere in your organization using natural language queries. So your actual use cases are defined by whatever data you have in your company.

It doesn't matter if your data is in your CRM, ERP, Data Warehouse, Database or hidden behind an API.

Slack, Teams, ChatterBox and your own front end can be customised in all respects to search and retrieve exactly the data you need from anywhere.

Slack, Teams or any front-end

ChatterBox

Financial Data

Natural languages queries on databases. ChatterBox is an ideal, out-of-the-box front-end

Product Information and Catalogs

Query your PIM catalog using all available fields with fuzzy matching and even sentiment searching on your product descriptions

Support

Use your own knowledge base and ticketing system to deliver fast and easy answers. Using ChatterBox, this offers a superior experience to traditional chat windows.

CRM/ERP data

Use Ziggy's ETL features to extract data ready for natural language queries. Ideal for many in-house business applications such as vendor information, inventory queries

Legal & Compliance

Build a legal document corpus of contacts, regulatory guidance, proesseses etc. then find them using natural language queries and semantic search.

Survey, general Data analysis

Any data you may have from surveys or other general data sets can be made searchable in minutes.

Train and call Models

We, or you, can train models using your own data for predictive or other uses cases.

Use these models for live responses for a single entity from your CRM, ERP, ChatterBox, Slack or API.

You can any system in bulk or incrementally to have an up-to-date predictors and indicators within your own data.

Model Details

ChatterBox has called the model with the selected customer. Model results are then shown in the green box below.

Model Training

We can train models for you to perform predictive or other queries. You can also provide your own trained model. Each model is fully customised to your requirements.

Querying a Local Model

Ziggy has a special Block 'AI Local Model' that takes a query and returns JSON data.

Integrating a model

A Ziggy Flow provides a simple, API callable way for your own call to pass data and get back the required indicators.

Custom Slack Apps

Custom Slack Apps can be created with Ziggy that let your users access any data using Natural Language Queries.

Hover for full screen

Watch a user searching for Hubspot data. Your CRM or other systems often can't perform the exact queries you want without complex coding.

Ziggy's Slack Apps let you have the data quickly and easily with as many / commands as you need.

Add custom URL shortcuts. Save frequently used custom searches.

This one, simple Flow handles standard searching across multiple Hubspot objects. The principle is the same for other CRMs, ERPs and data sources.

If you need to perform custom or complex queries, then you can point to another Flow.

This Flow fetches a selected deal when the user presses a button in Slack, gets associated line items before packaging and sending to Slack.

The Ziggy Slack Blocks Formatter Block lets you customise the appearance in Slack.

ChatterBox

ChatterBox is Ziggy's configurable, customizable and embeddable front end. The source code is also available.

Of course, you can also integrate your own front-end applications.

Display formats

ChatterBox displays data as lists, tables, cross tabs, pivot tables and charts.

Query History

Both queries and responses are retained so you don't have to requery your data or retype your queries.

Multiple Data Targets

If you need to query more than one data target, you can filter and switch between them in the ChatterBox header.

Customization

All aspects of ChatterBox can be customized from the styling to the behaviour. We can also enhance ChatterBox to meet any querying and behavioral requirements.

Ingest

If you've already got properly searchable data, then this step is not required.

If not, Ziggy comes with extensive ETL capabilities, so it's easy to create and maintain perfectly structured data stores for search and generative applications.

Flows

A Flow is an ETL pipeline. An Ingest Flow can

load data in bulk
be scheduled to load from SFTP or other sources
load updates from a WebHook/API call.

ETL

Ziggy has extensive ETL capabilties

Read from and write to data sources at scale.
Manipulate data using one of our many Blocks.
There's even a JavaScript Block for edge cases

Validate, Transform and Map

We provide Blocks and Edge Rules to help you validate, transform and map your data. This simplifies ingesting data from any data source,

Multiple Targets

Currently supported...

Elastic Search
Vector Databases
SQL Databases
Data Warehouses

Query Configuration

Querying is both highly and quickly configurable. It is suitable for chat, search and geneartive applications.

Flows

A Search Flow consists of

an AI Search Prompt Block
a platform specific Query Block

AI Search Prompt Block

This can be used for search and generative queries.

Translates the natural language query into a platform specific query
Each platform type comes with its own standard prompt template.
The prompt can be customized and tested from the Block

Edge Inspection

You can fully test and debug the Flow by inspecting the edge data between the Blocks.

Each Flow can be called by API. Data is returned in standardised formats for

Lists
Aggregations

Calling Hubspot or any other platform

The screenshot shows a Flow configured to query HubSpot using a HubSpot Block.

You can query any platforms that sits behind an API in the same way using the API Block where we don't offer a platform specific Block.

Here you can see the Flow being used by ChatterBox.

You can get counts and even associated objects such as Companies -> Contacts or Deals -> Line Items

ETL Capabilities

Ziggy has a powerful set of ETL features making it ideally suited for ingestion but also writing to other platforms, databases and APIS

ETL Details

Security and Performance

Security

You should be able to control all aspects of security.

Integrated secrets manager
All secrets, API keys etc. are fully encrypted
No sensitive data written to logs
Deploy to your own servers or Cloud

Deployment

Companies can control where Ziggy is deployed.

Deploy to any server/VPC
Dockerized, so easy to deploy
Control access at the DevOps level

Performance

Ziggy is fast but system load and 3rd party APIs need looking after.

API rate limit protection for third party APIs
Batching ensures that you can work with the largest datasets without overloading the system or APIs
Queuing ensures that Webhooks get a fast response and system is managed
Ziggy can be scaled at the infrastructure level to meet your performance requirements.