Machine Learning to Enhance Navy Service Desk
Navy SBIR 2018.3 - Topic N183-143
SPAWAR - Mr. Shadi Azoum - [email protected]
Opens: September 24, 2018 - Closes: October 24, 2018 (8:00 PM ET)

N183-143

TITLE: Machine Learning to Enhance Navy Service Desk

 

TECHNOLOGY AREA(S): Information Systems

ACQUISITION PROGRAM: PMW 250 DS-CRM Navy 311

OBJECTIVE: Develop a capability that improves upon current Information Technology (IT) help desk and customer relations, specifically in parsing and analyzing help desk communications (text and speech), reports, and logs. The resulting data set will be used to employ technical enhancements to the IT support and customer relations management space.

DESCRIPTION: Navy 311 is a Customer Relationship Management (CRM) component of the Navy’s Distance Support (DS) capability, managed by Program Executive Office Enterprise Information Systems (PEO-EIS). Navy 311 is a single point of customer service entry into the Naval ashore infrastructure and network of fleet support providers. Through Navy 311, the Fleet, Sailors, military families, and civilians can get on-demand information assistance for non-emergency, non-tactical issues. This gateway to comprehensive support can assist with issues including but not limited to: systems and equipment (e.g., hull, mechanical and electrical, weapon systems, information technology, technical data), quality of life (e.g., medical and chaplain care), personnel (e.g., career, manpower, training), supply and logistics (e.g., requisition follow-ups ordnance, food service, household goods), and installations and facilities (e.g., environmental, public works, community support). Navy 311 supports various communication mediums that include phone, email, chat, and mobile texts.

The desired capability shall have the following characteristics:
• Ability to search by meaning and context of desk communications (text and speech), reports, and logs to include:
    - Considering user search, email, chat, and browsing history to guide and enhance current search results
    - Promoting search results that other users selected when entering similar search criteria to the current user’s query 
    - Promoting search results for emergent, common issues when several other users’ enter similar search queries, predicting instantaneous recently-arisen, widespread trends in user searches 
• Ability to communicate to the customer base via bots (e-mail or chat) to include:
    - Automatically responding to end-user’s emails or chats describing their problems with the most likely solutions to the resulting tickets, closing the ticket, when appropriate, without human involvement
    - Automatically linking the user to a Frequently Asked Questions (FAQ) page that is known to address the issue the systems has determined the user is encountering
    - Automatically identifying, retrieving, and pre-populating forms the user will need to submit to address the issue at hand
• Ability to predict future workloads and resource utilization, allowing routine IT support tasks to be enhanced by automation, supplementing (not replacing) service desk agents. This includes:
    - Pre-populate emails and other communication forms, from helpdesk personnel to end-users, with the text and other content that has successfully addressed the same issue in the past
    - Search IT service logs and content to determine the most likely cause(s) of the user’s issue and display the historically-known fixes (with supplementary information) to the helpdesk personnel—who then selects the appropriate option
• Ability to automate routing and workflow of new issues to the appropriate personnel or electronic resource to meet end-user expressed needs, based on an understanding of the meaning and context of the issue, past successes (with similar issues), and the availability of different support resources
• Ability to predict future IT service trends by predicting the demand for new/existing IT services, or the future levels of IT support personnel needed, entailing:
    - Predicting upcoming spikes in resource utilization that will require extended labor demands, expanded electronic resources utilization, etc. 
    - Predictive analytics employed to predict future levels of customer satisfaction based on the past impact of various contributory variables
    - Crawling the web for upcoming Security Technical Implementation Guides (STIG) or STIGs that will likely impact operations

The desired capability should also be built to improve upon its own activities, improving over time. Once built, the capability should be able to process data to train itself, making improvements that will be tested and either kept or discarded by the system automatically. The system should also be capable of tuning by expert system operators, and then tested and deployed.

Current state-of-the-art technologies that can address this capability include machine-learning cloud services. However, these current services only provide general tools to begin an approach that address the above requirements. These tools only analyze stored state data sets within online searchable databases and the product is an online response which cannot be consumed as a service by Navy 311. A specific implementation is desired.

PHASE I: Complete a feasibility study describing a novel design for an analytics capability capable of performing tasks specified above with an environment to be proposed by the Small Business Concern (SBC). Include simple proof-of-concept prototypes and research-backed mockups.

The Government will provide a subset of communications (text and speech), reports, and audit logs related to Navy IT support help desk and customer relations that the SBC’s capability can process to implement the desired characteristics above.

Develop a Phase II plan describing the costs and technical effort required to implement the design described in the study. The plan should include visual depictions of the products’ features and general user experience.

PHASE II: Based on the results of the Phase I effort, develop a pilot implementation of the capability’s desired characteristics as mentioned above, integrating in a limited fashion with Navy 311.

The Government will provide a subset of communications (text and speech), reports, and audit logs related to Navy IT support help desk and customer relations that the SBC’s pilot implementation can process to implement the desired capability’s characteristics. The Government will also provide data necessary for the SBC to integrate its capability with Navy 311 in the limited fashion agreed to in the Phase II contract.

The SBC will present a plan to build the final product and integrate its use in the wider Navy IT community (100+ service desks serving a wide array of IT capabilities including: pay, logistics, records management, recruiting, retirement, healthcare, etc.). The plan should include projections of reduction in cost and lost time by employing these solutions in support of Navy 311 activities.

The following performance attributes will be assessed during the Phase II effort:
• Fidelity of cost projection to pilot implementation
• Fidelity of schedule projection to pilot implementation
• Effectiveness of software in enhancing capabilities from a help desk and end-user perspective to include:
    - Average time to assign ticket (NAVY 311 Support Center KPP)
    - Time to assign ticket to final actor (i.e. resolver) 
    - Average time to resolution (NAVY 311 Support Center KPP)
    - % of requests with mandatory fields blank (Metrics KPP) 
    - % of requests with invalid entries in fields (Metrics KPP)
    - Number of reassigned tickets

PHASE III DUAL USE APPLICATIONS: Complete necessary engineering, system integration, packaging, and testing to field the capability into Navy 311. Support Navy test and evaluation activities during software transition to production and software effectiveness after deployment is complete. Following testing and validation, the end product is expected to produce results outperforming the current Government agency business processes and ad hoc methods in use today.

The capability described in this SBIR topic could have private sector commercial potential for any IT business that needs to improve upon IT service support capabilities.

REFERENCES:

1. Mann, Stephen. “5 Use Cases for AI on the IT Service Desk”. SunView Software, Inc. 07/10/2017.  https://www.sunviewsoftware.com/blog/5-use-cases-for-ai-on-the-it-service-desk

2. Ragupathi, Ashwin R. “Machine Learning and ITSM: Helping Help Desks”. DevOps. 07/24/2017.  https://devops.com/machine-learning-itsm-helping-help-desks/

3. Fletcher, Colin and Lord, Katherine. “Apply Machine Learning and Big Data at the IT Service Desk to Support the Digital Workplace”. Gartner, Inc. 02/29/2016. https://www.gartner.com/doc/3232017/apply-machine-learning-big-data

4. Navy 311 Systems Engineering Team.  “Navy 311 Systems Engineering Plan (SEP)”.   v3.2 January 17, 2017 (uploaded to SITIS)

5. PMW 250 Logistics.  “Life-Cycle Sustainment Plan (LCSP)”.  v1.1 October 6, 2016 (uploaded to SITIS)

KEYWORDS: Machine Learning; ML; Artificial Intelligence; AI; Natural Language Processing; NLP; IT Support; Help Desk; Service Desk; Call Center; Customer Relationship Management; CRM

Return