The Hive CFO Forum Workshop Series: Continuous Revenue Management with AI

by Kamesh Raghavendra, Chief Product Officer, The Hive

On February 6, 2019, The Hive Think Tank launched the CFO Forum Workshop Series organized as private events that explore emerging applications of artificial intelligence (AI) in enterprise finance functions. This series brings together thought leaders in CFO’s offices of Fortune 500 companies, industry experts from big advisory firms and entrepreneurs creating new AI-driven enterprise finance applications. Workshops in this series will address CFO office business themes like Revenue Management, Strategy & Planning, Accounting, Internal Control, Audit, GRC, Treasury etc.

The first of the CFO Forum series was a workshop about Continuous Revenue Management with AI. This workshop explored the emerging role of AI & automation in high-fidelity revenue forecasting, revenue recognition, optimization, dynamic re-evaluation, and pricing. It was led and moderated by Gargi Ray, VP & Global Controller at Infosys. It was co-sponsored by Live Objects. The participants of the workshop shared their experiences and best practices in delivering on forecasted revenues by using agile optimizations of pricing, internal control, and incentives as dynamic levers. This blog describes the key insights that came from the session.

The session explored the challenges in revenue management across a broad spectrum of industry verticals represented by the panelists in the workshop. These challenges emerged in the following three broader themes:

a) Detecting and preventing revenue leakage

b) Revenue Closure Cycle and disclosures & determination of Revenue as per GAAP

c) Contractual Compliance

The panelists represented four broad types of industry verticals. We have captured below the trends in revenue management challenges across these four kinds of industries:

A. Service Industry

  1. Missed billing of hours which could be due to
  2. Not maintaining a complete control on the hours, or
  3. Scope creep — where the customer changes the requirements and those are accepted without any corresponding billing. It’s difficult to track whether the project is consuming more efforts for the same scope of there has been a scope creep
  4. Unstructured conversations leading to business decisions are often not traceable back to the system. Becomes difficult to replicate the rationale for the decision as we cannot trace those conversation threads.
  5. Pricing discipline — This is typically seen as an issue because there is a vast difference between the rack rate/ list price and the realized price.

B. SaaS Industry

  1. Diversity in Subscription — Difficult to capture the variation in the subscription models which are offered to consumers. Mostly manual excel tracking needs to be done for variations as the systems may not be geared up to capture and compute the revenues. This may lead to a revenue leakage.
  2. SaaS + Professional Services — Typically the models are hybrid nowadays with a professional service element attached to the same. Professional services may be difficult to be priced on a comparative basis.
  3. Time evolution of Billing — Managing the revisions applied to subscriptions as customers continuously modify their plans to optimize spend
  4. Seat/Usage Based Billing — Billing models may be dependent upon the seat/ usage. However, there could be inappropriate sharing of licenses and there may not be robust enough systematic tracking to detect usage.
  5. Gainsharing model — Revenue models also be dependent upon the client customer’s usage of data. This data is extremely vulnerable to the accuracy of reporting by the client of the end consumer’s data.
  6. Royalty-based revenue model — In royalty-based revenue models tracking actual usage is often difficult.

C. Consumer Electronics Industry

  1. E-commerce — This industry deals with high volume low order values. The incidence of giving chargebacks is extremely common in this industry. Incorrect capturing mechanism of these chargebacks lead to fraud and revenue leakage.
  2. Multi-channel/Multi-vendor Participation — Since the entire chain has many participants there is a risk of fraud at every juncture.
  3. Component Cost-down Model — In many circumstances, the component costs may be really low per unit. Due to lack of transparency of the entire supply chain it gets difficult to understand where there are losses in the cycle.

D. Online Marketplaces

  1. High-value transactions typically change hand, especially when the goods/services exchanged are regulated; so, it becomes extremely essential to maintain traceability of the entire transaction. This is currently a challenge is there is a break in the cycle.
  2. Cancellation Policies — Usually in this industry, the cancellation policies are not in the favor of the vendor hence it becomes important to determine at which point does the obligation for one party end. Different event updates need to happen in the system to evidence the same.
  3. Long Transaction Cycle — There are multiple parties involved and the transaction cycles are usually long hence there may be leakages due to delay or non -closure of pending issues. There could be issues like partial payments made so the transactions are on hold.
  4. Data Syncing issues — There is a huge data flow between multiple vendors and more often than not there are data syncing issues which lead to revenue leakage
  5. Promotions/ Offers, unaccounted — There are on the spots promotions given which may be ratified later but not traceable in the system.

Several revenue management challenges emerged unanimously across all the industry verticals represented at the session:

  1. Adherence to revenue recognition standards set by ASC606, especially interpretation of contracts and events across multiple industry verticals (especially non-software related verticals)
  2. Expense recovery from contracts in a common area of leakage as adequate internal controls may not be there
  3. Fraud detection techniques need to be more robust and should evolve with different revenue models and industries.
  4. Adherence to new leasing standards set forth by ASC842, while segregating leasing and non-leasing related aspects of contracts.

Revenue Closure Cycle

The session also had a segment dedicated to the revenue closure cycle process. Revenue closure resonated as a common issue for all panelists. Accurate and timely closure can help the entire book closure and financial statement review process. If we can close early it saves times to analyze the company performance, accentuate the strong areas and reason out why certain areas did not fare well. This helps in positioning the company story to the external stakeholders.

Common points of discussion which came up in the Revenue closure cycle are:

  1. Multi-system input and output — For most companies, there are peripheral systems providing inputs to the final Financial ERP data. There is a lot of time required to ensure exhaustive and accurate check-ins and check-outs.
  2. Revenue leakage points if can be tacked this would automatically streamline the closure cycle
  3. Human collaboration and judgment if can be integrated in the system it would help in future decision making.
  4. Exception dashboards throughout the life cycle and on a real-time basis would make the closure process very simplistic and at any point in time we would know the revenue for the business.
  5. Today’s GAAP requirements have made the determination and valuation of performance obligations even more difficult, if we can insert structured and unstructured data and decisions into the system it would help in getting better data points for both determining and value PO more consistently.

Role of Artificial Intelligence (AI) in Revenue Management

AI can drive three types of applications in addressing enterprises’ revenue management challenges:

1. Automation: AI brings robustness, resiliency and intelligence to automation, enabling such automation to address much broader variations and complexities in revenue recognition & closure cycle tasks. AI also drives the ability to reinforce automation by learning from human-driven resolution of new exceptions.

2. Optimization: AI can also be applied over goal-based optimization of rules, workflows and configurations for contract interpretation, exception detection/remediation and fraud prevention. Such optimization is learned over measurements performed over existing rules/workflows and detecting sub-optimal performance (e.g. revenue leakage, un-prevented fraud etc.), while using any human-driven remediation as feedback to train new rules.

3. Validation: AI-driven verification of compliance against rules and regulation can be leveraged to replace sampling-based validation with 100% fact-check based validation. This involves curation of compliance related artifacts (data elements), collaboration with human subject matter experts when relevant and aiming 100% fact-based coverage.

The above applications of AI can bring radical transformations in revenue management:

  • Deliver continuous revenue management at high volumes of transaction postings
  • Address spikes during quarterly cycles without loss in SLAs or creating spikes in work for the CFO organization
  • Non-routine contracts can be processed with minimal training
  • Review timelines would be shortened and would be more consistent
  • Helpful for quickly adopting new standards like ASC606 and ASC842

Conclusions

The workshop vividly brought out the dynamics of the revenue management space, and the crucial role technologies like AI can play in addressing key challenges. However, it will take a strong coalition between the CFO office thought leaders, CIO office enablers, technology/SaaS platforms and startups to effectively solve these challenges with efficient applications of AI. Given the strong influence of big-4 advisories in the CFO offices, they will continue to play a crucial role in accelerating the adoption of technologies such as AI for such use-cases.