How Machine Learning Transforms Gray Work to Smarter Dynamic Work

Jamie knew she had plenty of work to do when she took the Director of IT role at a midsize HVAC services company. 

Over the last decade, the company grew organically and expanded regionally by acquiring smaller service companies but had done very little to modernize its business operations. The business was largely running on a highly customized ERP, a basic CRM, spreadsheets, and emails, and she knew there was a significant opportunity to transform this gray work with automation and machine learning capabilities.

Gray work refers to dysfunctional business operations backed by platforms that don’t integrate, have dysfunctional user experiences, and contain poor data quality. When tools don’t integrate, business managers must copy/paste or swivel chair between screens to complete tasks requiring multiple platforms. When there’s poor data quality and limited reporting capabilities, they export data to spreadsheets, manually crunch numbers, and email outdated results to their colleagues.

I know Digital Trailblazers like Jamie and have consulted with HVAC and construction companies seeking to become operationally resilient with dynamic work management. It’s a feasible transformation for many industrial companies, even with small IT departments and limited budgets. They replace the gray work by building dynamic workflows with no-code platforms and leverage built-in machine learning and artificial intelligence capabilities without hiring software developers and data scientists. 

If I spoke to Jamie, I’d tell her about the recent Forrester Total Economic Impact study showing a 315% ROI and a 15% reduction in wasteful operational spending by transitioning to dynamic work management with no-code and citizen development.

ML-enabled dynamic work use cases

A key strategy for Digital Trailblazers looking to achieve significant ROI is to seek digital transformation force multipliers. Instead of investing in multiple SaaS tools to support small departmental workflow challenges, IT establishes a center of excellence for citizen development and uses a no-code platform for multiple business challenges.

Digital Trailblazers use no-code platforms to improve operational efficiencies but also target revenue generation and growth opportunities. They recognize that dynamic work lies at the intersection of peoples’ abilities to make smarter and faster decisions based on integrated workflows and machine learning capabilities.    

I’ll illustrate with two examples of how Jamie can accelerate digital transformation.

Managing dynamic fleet scheduling and maintenance

A midsize HVAC company can easily have a fleet of over 100 vehicles that require juggling daily work and fleet maintenance schedules. Optimizing these schedules can be challenging, especially in regions where cooling is needed during the 6-9 hot weather months and where a spike in service calls occurs during heat waves.   

Building a dynamic workflow requires connecting operational systems used to schedule teams and equipment, asset management systems, and weather forecasting data. In Jamie’s case, a CRM stores the customer service request, and the ERP stores the fleet information, but two employees manage the equipment scheduling and maintenance in different spreadsheets. These tools don’t leverage weather information, and teams scramble to make changes whenever there is an unexpected spike in service calls.

Here’s how Jamie and a citizen developer can build a dynamic workflow with machine learning to address this challenge.   

• An integration pipeline runs whenever the CRM captures data on a new or modified service call.
• A second pipeline connects to the ERP to capture adding or decommissioning vans and other equipment.
• A third pipeline pulls in updated weather forecasts every hour.
• The citizen developer creates one dynamic workflow to replace the crew and fleet scheduling spreadsheet. It leverages mapping to optimize routes and machine learning to predict the time required to complete a job by its type and sizing metrics.
• A second dynamic workflow schedules van maintenance and uses machine learning to forecast optimal times when demand is low and weather-related spikes are not in the forecast.
• The IT team provides access to the workflows and information through a mobile application tailored to each department’s needs.
• This dynamic workflow improves customer satisfaction when crews arrive at jobs on time, optimizes how the company uses its fleet during peak periods, and ensures maintenance is performed with minimal operational impacts. 

Creating workflows for marketing events

Transforming gray work to dynamic work doesn’t just apply to operational workflows. Many front-office functions, especially in marketing, provide the opportunity to streamline processes and enable data-driven results.

Jamie’s organization has a CRM, but the tool only supports sales and customer service workflows. The platform has limited marketing capabilities, leaving this team to manage its workflows and data with a hodgepodge of SaaS tools and manual data integration steps.

Jaimie discovers that the primary marketing activity centers around events, including events that the marketing department hosts for prospects and commercial events where they collect leads at the company’s booths. Over a typical three-month period, the marketing department hosts eight events, connects with hundreds of prospects, and has one gargantuan spreadsheet capturing all this data. There is a manual push to get the data into the CRM, and one consultant does this step manually about once a week. Marketing wants better tools to forecast which events to attend, who to invite, and how to create a closed-loop connection with the CRM.   

A citizen developer addresses this gap by creating a no-code events management application. He loads data from the spreadsheet into two tables, one for the events and one containing prospects who attended them. Instead of one person managing the data, the whole events team gets access and spends time adding more event and attendee dimensions. The citizen developer then adds a third table to track follow-up activities and configure the platform’s data analyzer to score prospects.

When a marketing manager adds new prospects after an event, machine learning predicts how likely each prospect will convert into a sales opportunity. The event’s marketer pushes the hottest prospects directly to the CRM’s sales funnel through a pipeline, and they optimize invites to future events based on the ML model’s predictions.   

How all businesses, from SMBs to large enterprises, transform gray work

From these two examples, I hope you see how organizations of all sizes and across industries can accelerate their digital transformation with dynamic work management and an ML-enabled no-code development platform. The key to success starts by identifying the organization’s gray work, establishing citizen development capabilities, and building ML predictions into dynamic workflows.

This post is brought to you by Quickbase.

The views and opinions expressed herein are those of the author and do not necessarily represent the views and opinions of Quickbase.