What to do when your clients want to cut costs?
Help them cut costs, and be part of a bigger vision beyond cost-cutting.
The coronavirus hasn’t been friendly to most consultants and professional services leaders – as enterprises clamp down on spending. It’s hard to make strong predictions about what the coming months and years will bring – and the only certainty most companies have is the need to conserve capital.
While there might be less money circulating in the economy, there are new sets of technology and strategy priorities emerging (a topic we’ve covered in our previous Emerj Plus articles) – which present new opportunities to professional services firms that are smart enough to get ahold of them.
In this article, we covering three critical factors that will separate the winners from the losers in terms of AI ROI during the crisis – and we explore four unique categories of cost-cutting AI and automation applications – including reference links to help readers find sample AI use-cases quickly.
What Will Separate the AI Winners from the AI Losers in a Time of Crisis?
The AI consultants who come out of this crisis with more clients and more client trust will focus on: Adapting to changing market forces, leveraging an understanding of the full range of AI use-cases, and overcoming adoption challenges.
We’ll explore each of these factors in greater depth below:
Adapting to Changing Market Forces
The list below highlights four changing market forces, and the resulting adaptation that smart AI consultants should consider (note that this same advice applies to innovation or strategy leaders within the enterprise).
- Leaders focus on cutting costs during recessions >> Identify and propose specific cost-reduction projects for clients
- Budgets are falling rapidly for some professional services more than others >> Response: Diversify service portfolios toward mature and growing sectors, or towards markets less impacted by the virus
- Rapid market changes create confusion, stagnating decision-makers >> Map the new landscape, identifying specific risks and actionable opportunities
- Projects and strategies are being re-drawn >> Redraw plans to focus on low-complexity, high-ROI projects, and play a role in helping to shape a client’s longer-term future vision
Understanding the AI Use-Case Landscape
Clients don’t know what specific AI solutions they need. C-suite AI mandates are often based on erroneous factors (“Our competitors are pushing press releases about this kind of AI solution, why aren’t we doing that?”), and they’re often handed down with the supposition that AI can be “plugged in” like an IT solution.
Helping a client reach their goals means understanding enough about the broader AI landscape to make suggestions that make sense. This begins with determining the data sources and business processes a client is working with – and developing a list of plausible AI (or non-AI) solutions to each.
Saying “yes” to the assumed AI “solution” to the client’s problem often means setting oneself up for failure.
First, because the presumed solution may not be a fit given the client’s goals –
Second, because the presumed solution is unlikely to take into account the true state of the client’s data infrastructure and data science talent, and –
Third, because the client wants to be led by a trusted partner who can set them up for success. A poor choice for AI solutions – even if it was one the client was enthusiastic about – means less motivation to move forward with AI, and a potentially soiled relationship between consultant and client.
Our AI Opportunity Landscape research involves constantly mapping and ranking AI startups, AI use-case maturity, and AI ROI across industries. Enterprise leaders often need this ROI data to move forward on AI projects, any smart consulting firm should develop their own methods of assessing the right tools for the job – beyond a Google search for “AI in [client’s industry].”
Overcoming Adoption Challenges
During times of crisis and recession, enterprises redraw project plans in light of shifting strategies. Consider specific opportunities to surmount adoption hurdles during periods of economic contraction through the lens of the diagram below.
When the demand landscape shifts, it’s time for a fresh assessment of strategic initiatives and product positioning goals. In light of falling demand for non-essential services, leaders would be wise to lean towards initiatives and proposals with lower risk, lower complexity, and clearer return on investment. This must be balanced with a future-forward vision that avoids reactive initiatives, which do not support the organization’s long-term goals for product positioning and infrastructure investment.
Thanks to Ian Wilson (former head of AI at HSBC), whose ideas helped to shape the simple diagram above.
As a consultant, or even as an internal strategy leader, consider the following set of factors, draw from our Critical Capabilities AI prerequisites framework:
- Data science talent requirements. What kind of new data science skill will this project require? What kind of subject-matter experts from other parts of our business will we need to be dedicated to this kind of project?
- Data infrastructure requirements. What kind of data, data volume, and data quality would this project require? Is the state of our data sufficient to consider this project in the first place?
- Executive understanding. Does leadership understand the upfront data infrastructure work and iteration that this project will require? Are the expectations of ROI too near-term, and unrealistic?
Enterprise leaders and savvy consultants should be looking for applications that require minimal data sources – and which leverage data that we know we have on-hand (and in reasonably good quality). Leaders should also look for applications with an immediately evident benchmark for measuring success (preferably, a tangible reduction of time or cost). Look for applications with robust case studies, not nascent applications.
More than ever, enterprises should steer clear of underfunded, smaller AI startups who make big promises but have no past clients to speak of (we recommend reviewing our article).
Firms looking for a succinct blueprint for deploying AI in the enterprise, consider getting a copy of our Emerj’s AI Deployment Roadmap Report.
Automation and AI – Applications for Post-COVID-19 Efficiency
So where is the low-hanging fruit when enterprises are working fast to cut costs? We think it’s in front of our faces. Serve organizations to cut their costs. Consultants and professional services leaders who can’t help with immediate priorities of the present won’t be able to get their foot in the door for long-term AI transformation projects in the years ahead.
Below, we’ll explore four automation and AI applications that can not only help with cost-cutting, but they also have a strong potential to remain relevant despite in face of COVID-19 economic crunch.
If you’re looking to find cost-cutting opportunities for clients, or proven AI use-cases to drive efficiencies, these resources can help you explore actionable opportunities.
Robotic Process Automation
Robotic Process Automation (RPA) is a relatively mature technology (compared to almost all enterprise AI applications), enabling cost savings with automation capabilities across a broad and growing spectrum of use cases.
RPA is not AI – it is generally a set of if-then steps followed by a machine, in order to replicate and streamline a manual workflow by a human worker. Automation is automation, however, by AI or by any other name.
While RPA often requires a rethinking of existing workflows, it doesn’t involve the same complicated data concerns or cross-functional team concerns that AI poses. RPA also generally requires drastically less iteration and experimentation than enterprise AI applications. It is
Most of the largest RPA vendors are already purporting to leverage AI in their workflows – and RPA systems are likely to become more and more capable as they begin to leverage machine learning and handle more complex or varied workflows.
RPA Articles from Emerj
- The State of RPA in Banking – With Charts and Graphs
- Robotic Process Automation (RPA) in Insurance – Current Applications
- Robotic Process Automation (RPA) in Healthcare – Current Use-Cases
- Robotic Process Automation (RPA) in Finance – Current Applications
Fraud, money laundering, and other cybercrimes often increase in times of economic strife, and the pandemic is no different.
In light of the coronavirus crisis, we believe that fraud detection applications are among the AI use-cases that are most likely to be adopted and deployed even when funds dry up for other kinds of more long-term, strategic innovation investments.
In addition to being cost and risk-oriented (as opposed to moonshot or revenue-oriented), fraud applications often have a number of other advantages in terms of ease of deployment, including:
- Measurability. If I build a recommendation engine to help my wealth management clients find information for their trades, how many months (or years) do I have to test that application in order to determine it impacts on customer lifetime value or retention? With a fraud application, it is often possible to measure a reduction or increase in false positives and false negatives (for payment fraud, for account fraud, etc), in order to assess the potential ROI of an application in a span of a few months – or maybe even a few weeks.
- Use-case maturity. While most enterprise AI applications are nascent, fraud detection. Our AI Opportunity Landscape in Banking research has shown that fraud and cybersecurity AI vendors have raised more money than customer service, marketing, and sales applications combined. That specific trend is from our banking industry research only, but it’s indicative of a broader truth about AI in the real world. Namely, it works well on pattern recognition – and fraud can be seen at its core as a pattern recognition (or anomaly detection) problem.
- Integration and data complexity. Hardly any AI application is “easy” to integrate in a real enterprise environment. That said, fraud applications (say, in payment fraud) often involve a bounded, succinct range of range of data (who paid, how much did they pay, at what location or URL did they pay, at what time, etc). Contrast this with other AI applications that must draw from half a dozen data sources, often bespoke from client to client.
Fraud Detection Articles from Emerj
- AI for Fraud Detection in Retail – 2 Powerful Use Cases
- AI-Based Fraud Detection in Banking – Current Applications and Trends
- How to Deploy an AI-based Fraud Detection Solution in Financial Services
- Machine Learning for Fraud Detection – Modern Applications and Risks
- Artificial Intelligence for Anti-Money Laundering – An Analysis of Solutions
Many large insurers are finding ways to digitize parts of their business process in preparation for future projects involving machine learning. This is especially true in claims processing, which could become faster and less error-prone if claims adjusters did not have to search through large amounts of data or paper documents manually.
Claim volumes – and insurance fraud – increase during a recession. Discover how to parse legitimate claims from excluded or fraudulent claims with AI-driven automation.
Claims Processing Articles from Emerj
- AI for Claims Processing and Underwriting in Insurance – A Comparison of 6 Applications
- AI-Based Enterprise Search for Claims Processing and Fraud Detection
- What do Insurance Experts Think about AI in Claims Processing?
- Machine Learning Applications in Large US Insurance Companies
Lending and Underwriting
At its core, lending is a big data problem, making it a business naturally suited for machine learning.
Part of the value of a loan is tied to the creditworthiness of the individual or business that took out the loan. The more data you have about an individual borrower (and how similar individuals have paid back debts in the past), the better you can assess their creditworthiness.
Instead of using traditional FICO scores, new machine learning models might use new data sources about applications (purchase activity, social activity, a combination of demographic factors, etc) to find more narrow pockets of risk, or pockets of opportunity, that sweeping rule-based scores might have missed. In addition, AI has promise for automating or streamlining the loan application process, helping lenders to more with less, and allow humans to focus on more cognitively demanding work (i.e. examining complex applications, not manually processing claims or denying applications that had no chance of being approved in the first place).
Banks are seeing a flood of new lending and refinancing applications due to COVID-19. Automated screening of inbound applications is an opportunity to handle the flow, and strong risk scoring algorithms pre-empt costs of default. Not only are
Lending and Underwriting Articles from Emerj
- AI Startups in Auto Lending – 2 Well-Funded Examples
- What Banking Leaders Think About AI – and What They’re Missing
- AI Applications for Lending and Loan Management
Emerj for AI Professional Services Providers
Emerj helps AI consultants, strategists, and enterprise innovators catalyze AI adoption for a high return on investment. Discover which AI technologies, AI products, and AI services best support the business priorities of your clients and stakeholders.
Leverage Emerj’s proprietary research and data to inspire confident investment in Robotic Process Automation, Fraud Detection, Claims Processing, Lending, Underwriting, or explore other opportunities.
Learn more about Emerj Research Services.