Before selecting an SPM software based on its AI functionality, understand all the potential use cases.
Sales leaders have numerous technology options for sales performance management, from point solutions to expansions of enterprise-wide software solutions. At Spaulding Ridge, we have the privilege of working with multiple software providers, and we often help our clients select the best software option to manage sales planning, territory design, quota setting, or incentive compensation management.
Increasingly, AI appears to have become table-stakes in SPM. From boosting forecasting accuracy to automating manual processes, there are numerous ways that AI can help you with your SPM needs. But AI can’t do everything (yet). For example, while it’s smart to consult your favorite LLM to start researching SPM software options, remember that LLMs are limited by what they can’t see. ChatGPT can tell you what an SPM vendor’s marketing copy says, and research customer reviews, but it can’t see under the hood and tell you whether the software actually delivers.
Where to start? Here are four major SPM-specific AI capabilities to look for in your next SPM software provider, and what to ask to make sure you find the right fit.
Increased Accuracy in Forecasting
What to look for:
Improving your forecasting accuracy is an undervalued benefit. The first-order benefits are obvious: If an organization can improve its forecasting accuracy, that means it can set more accurate and fair quotas. However, the second-order impact of setting more accurate quotas means that the organization can get more confident with their quota setting and have a good sense of where most of the reps will finish the year vs. their quotas. This means you can financially afford to set more aggressive overachievement accelerators for the top performers in your organization. This drives your best sellers to be even better while improving overall company performance and retention.
In the context of SPM, a software provider’s time-series forecasting capabilities are key, and most providers have a library of options. A machine learning algorithm that can cycle through and select the best-fit forecasting approach for your data significantly enhances sales forecast and quota-setting accuracy.
What really separates the great from the good is the ability for the forecasting product to ‘explain itself,’ not just to sales operations but also to business users. Can the forecasting algorithm justify the drivers of the forecast down to the lowest level of detail? This is often what helps secure buy-in from leadership and the field.
What to ask:
Ask about the depth and flexibility of the model library to deploy different forecasting algorithms at key intersections of your business that behave differently. For example, in our Enterprise customers with a global footprint, we see different forecasting algorithms end up being the best fit for different geographies and/or business models, depending on their level of maturity.
Ask about the features for forecasting explainability and justification, as well as how findings can be presented to various stakeholders.
Effectiveness through “In-flow” Automations
What to look for:
If you’re responsible for running payroll at your organization, how often have you analyzed the top and bottom earners and found anomalies ten minutes before payroll? How many times was your payroll delayed due to one-off investigations? And how much easier would it make your life if you could avoid even half of those issues?
“In-flow” automations refer to “human-in-the-loop” operations that augment the payroll professional. Payroll still needs to be reviewed and approved by a human—it’s too mission-critical to set and forget—but it can be significantly accelerated. Commission anomaly detectors can speed this process up. Without waiting for payroll day, you can set up standard anomaly detector algorithms to run in the background and proactively detect and highlight the key issues. For enterprise customers with a global footprint, this type of “in-flow” automation enables payroll workflows to be decentralized across regions while still holding to a high standard.
What to ask:
When selecting an SPM software, ask what out-of-the-box anomaly detector algorithms and packages are available to deploy. Make sure you determine what workflows and processes they can support, and that can be configured. Then, make sure these AI capabilities can be trained on your data and set up with custom guardrails instead of a black box. Acceptable anomaly thresholds differ from company to company, so be sure you can set your own threshold in the system.
Time Savings through Agent-based Analysis / Dispute Resolution
What to look for:
In the world of SPM, “death by a thousand cuts” often means “death by a thousand investigations.” When I ran payroll, I used to spend all my time answering questions like “Why didn’t I get paid for this deal?” Every question meant another investigation and hours spent chasing down data. It took up so much time that I wished I had a robot assistant to help me.
Agent-based dispute resolution gives you that robot assistant: An LLM-powered agent that conducts dispute resolutions and pay investigations for you. It starts by operating under human oversight. The dispute resolution agent conducts the investigation, but a human reviews its response before it’s sent back to the end-user. Gradually, as these models get trained, they can handle requests independently.
Agents are good for more than dispute resolution. Many SPM leaders also use them for territory and quota design or compensation plan design. Through scenario modelling algorithms, AI analyst agents can test capacity planning or spiff scenarios with low manual effort required. They can answer questions in plain language, such as “what would be the total account load and quota per rep if we reduced our headcount by three percent?” or “what would this spiff have cost us last year?”, allowing leaders to skip the fumbling with reports and calculations to get the answers they need quickly. This gives sales leaders more time to think about their strategic approach to capacity planning, territories, and quotas, leaving the tactical calculations to the agents.
What to ask:
Today, almost every technology vendor offers agents built into their software. Ask about your ability to configure and train task-specific agents, and how you can set guardrails and security access restrictions so that the agents only access and retrieve what the end-user has access to. SPM agents need to work with sensitive payroll and customer information, and you need your agents to keep that data secure.
Efficiency through natural-language comp plan configuration
What to look for:
Getting an SPM system working is an important step, but it’s not the last step. If your organization doesn’t plan for updates and maintenance down the road, you could wind up with a “day two” problem. SPM/ICM platforms require platform-specific knowledge to maintain and update, and compensation plan designs change significantly over the years (usually very close to year-end). When this happens, reconfiguring the ICM plan is no small ask, regardless of the platform.
In the old days, this meant hoping the person who coded the ICM plan still works at the organization and then asking them to make changes on a tight deadline. But today, natural-language comp plan configuration means an agent trained on the platform’s building blocks can effectively create the underlying building blocks or code (think SQL, etc.) that will enable configuration.
Buyer beware: This capability is much more robust if the software platform has ICM-specific functionality that it can call upon. If not, the configuration agent will likely have many more configurations to make, so the likelihood of hallucinations or subpar configurations increases significantly. The more the agent has to do, the more you’ll need a human analyst in the loop.
What to ask:
Before buying your SPM system, ask about the configuration process and whether you will have access to a builder agent. Is this agent working with pre-existing functions or building blocks? Or is it expected to architect an entire system? Can you access its outputs and audit its code and configurations, or is it a black box? Understand your internal technical resources before deciding how to approach this challenge.
Finding the right SPM software is no small feat.
As sales and operations leaders decide which SPM software to invest in and how to harness the power of AI within their organizations, they have more options than ever, from their existing enterprise vendors’ extensions to new point solutions. While almost every SPM software provider boasts AI capabilities, there are significant differences in what each SPM system can deliver. There’s no one correct answer, but a strategic partner can help you find the best fit for your needs.
Spaulding Ridge works with several carefully selected technology vendors, allowing us to understand how your SPM system fits within your overall data flows. Whether you’re just dipping your toes into SPM research or looking for a full AI SPM transformation, we can help you take the next step. Reach out to discuss how we can build the next-generation tech stack for SPM.