---
title: "1916 Release - 2019-05-02"
date: "2025-09-12T05:57:24+00:00"
summary: "AgilOne 1916 release: Actions improvements, ML model enhancements, platform fixes, and new admin UI for user access management."
image:
type: "announcement"
url: "/customer-data-platform/release/2019-05-02/79811-1916-release-2019-05-02"
id: "80f1a5d0-dba6-49b5-8177-c456fdd74e0e"
---

**Area**

**AgilOne JIRA**

**Description of functionality after the fix**

Actions

AGO-11768

In this release we have improved Actions Advanced Segmentation to enable the use of the ‘In’ and ‘Contains’ (i.e. Like) operators when segmenting an audience.

Actions

AGO-11978

Previously, when trying to use the ‘Address Type - Residential or Business’ filter in Actions, you were incorrectly provided with the option to filter by ‘Unknown’. The correct value in the data for unknown is actually ‘U’. We have fixed this bug so that you are presented with the accurate filter options of: ‘R’, ‘B’, and ‘U’.

Actions

AGO-12006

We have patched a bug in Actions that was causing the ‘Were included in an AgilOne campaign’ filters to incorrectly display the campaign ID instead of the campaign name after saving. These filters now properly surface the campaign name.

Actions

AGO-12015

The Actions audience count filter was previously not taking the optimizer into account when choosing to read from cache (fast) or refresh the counts from the data (slow). We have fixed this issue so that the optimizer is considered when deciding whether to read from cache or the underlying data.

Actions

AGO-12016

In addition, we fixed a separate bug with our Actions audience count refresh cache that was causing the cache to not be read properly on certain occasions. For scalability reasons there are two separate cache services maintained behind the scenes. Before 1916, the two services could have gotten out of sync, but now they will remain in sync and return the same values.

Actions

AGO-12103

We have fixed a bug in Campaigns+ that was preventing users from naming the output file for SFTP destinations.

ML

AGO-11828

For some ML models (Likelihood to Buy, Behavior Clusters, and Product Clusters) we were not properly storing the cluster names of historical values: ‘as of this month’, ‘1 month ago’, and ‘2 months ago’, instead we were storing the raw data. For example, instead of storing the human readable name of ‘High LTB - Decile 1’, we were storing and surfacing the raw data of ‘1’. This has been fixed, so over the next 2 months the values for LTB, BC, and PC ‘as of this month’, ‘1 month ago’, and ‘2 months ago’ will be more human readable and usable.

ML

AGO-11976

Additionally, we have enhanced our LTB, BC, and PC ML models to enable them to train on even larger datasets (>2 billion records).

Platform

AGO-12034

In this release we patched a bug with our customersummary.emailStatus logic that was causing both ‘Unverified domains’ and ‘Unknown’ emails to get assigned an email status of ‘U’. Now, only unverified domains will receive an email status of ‘U’ and unknown emails will be assigned a status ‘Unknown’.

Settings

AGO-11922

For admin users, we have introduced a new user-interface for reviewing who has access to your AgilOne environment, and at a high level what access they have. It is our hope that this UI will better enable our clients to review who has access to the AgilOne application, and thereby request any necessary adjustments. This application is currently in a beta state. Access to this application will require a request from the business sponsor to [helpdesk@agilone.com](mailto:helpdesk%40agilone.com) before it will appear in the settings tab.