Securing executive buy-in is an integral step in bringing your artificial intelligence (AI) project to fruition. However, it often takes more than a good idea to acquire funding. Not every business is wired to reward change, and even those that pride themselves on innovation may lack the budget necessary to finance every good idea that comes their way. Convincing investors, executives, and other high level leaders to fund your project can be a challenging process, one that benefits from ample leg work and preparation.
To increase your chances of securing executive buy-in, prepare a strong proposal that takes into account the motivations of your business leaders, and makes a case for the technical feasibility of your AI project. With that in mind, here are five tips to consider when attempting to secure executive buy-in.
1. Show How AI Will Move the Needle
First and foremost, understand the motivations of those authorizing and funding your AI project. Your proposal should appeal directly to the incentives of your company higher-ups. Company leaders ultimately want to see a return on investment (ROI) when deciding whether or not to fund a proposed project. However, the long-term profitability of an AI solution may not be immediately clear. Speak to your solution’s ability to move the needle on metrics (KPIs) indicative of future ROI outcomes.
Most companies utilize several KPIs catered to each aspect of their organization (e.g. financial, sales, customer relations, process performance, and marketing KPIs.) It’s a good idea to familiarize yourself with both your business’s important KPIs, and any initiatives already underway to achieve them. As a rule of thumb, your AI solution should aim to supplement an existing initiative in service of a specific metric. If possible, try to produce realistic figures that outline how your AI project will progress a given KPI. For example, an introductory sales pitch may sound something like, “We forecast that this AI initiative will enable our business to increase the number of resolved customer ticket requests from 500 per day to 1,500 per day.”
While claims of this sort may serve to gain the ear of executives, it’s also necessary to explain what specific efficiencies your AI project will create. Returning to the previous example, you might add, “While our human support staff can currently handle approximately 500 ticket requests per day, an AI chatbot, capable of automatically answering frequently asked questions, can free up our human support staff to focus on more pressing or involved support requests.” In this case, the core business problem and AI solution are clearly defined.
Second only to the content of the proposal is the language you use to deliver it. Try not to get bogged down in the minutiae of the technologies involved in your AI project. In early discussions it’s more important to get an executive engaged with the high-level business implications of your AI project. Although there’s a time and place for finer, tech-oriented details, you should primarily focus on what AI will do to positively affect KPIs, and how this will be accomplished.
2. Present Your Project as Proof of Future Innovation
Your AI project should not be framed as a one-time solution to a one-time problem. The best AI initiatives are emblematic of something bigger: a business’s commitment to innovation. Whether you’re an incumbent business looking to stay on top, or a startup looking to disrupt an existing market and capitalize on a growing customer base, innovation is key to the growth trajectory of your business. Persistent innovation allows companies to adapt to shifting markets, differentiate themselves from their competitors, and increase productivity.
If there’s a lesson to be learned from the economic strain of recent years, it’s that
no growing business can afford to be complacent. We live in a less consistent world, where supply and demand are subject to rapid fluctuations, and where existing business models struggle to cope with unpredictability.
As a result, the businesses that stand a better chance of remaining secure, profitable, and capable of growth, are those that continue to innovate.
Take, for example, any number of successful companies (Nestlé, Adidas, and Dyson to name a few) that have looked to AI innovation to help transform their business models in recent years. Some of the more promising initiatives include using AI to forecast demand, adjust supply chains, mitigate labor shortages, and increase e-commerce revenue through personalization (1)(2)(3).
Additionally, timely updates to AI-driven features, both at a system and product level, can help to future-proof your business. For business-to-business (b2b) operations, this could mean utilizing AI and predictive analytics to generate more leads, capitalize on trends, or discover actionable customer insights. Alternatively, for business-to-consumer operations, this could mean using AI to increase personalization, or improve user experience (UX) design. In both cases, realizing AI’s potential to deliver innovative products and services can help businesses to maintain and expand brand loyalty.
What’s more, the idea of innovation can be a selling point in itself. For this reason, many businesses want to be perceived as innovators in their given field. Whether or not a business is truly innovating is evidenced by their ability to implement novel solutions to complex problems. Successfully deploying AI solutions that makes progress on KPIs can serve as a valuable proof point. It signals to the customer that you can back up your claims of innovation, and to the executive that your business is capable of developing impactful AI solutions.
As a final point, you should maintain that, if your project succeeds, it will serve to break ground for the consistent deployment of AI-powered features into your business. Evidence suggests that continuing to grow your business’s AI capabilities is directly translatable to tangible financial gains. For example, a 2020 report released by IBM revealed that while adopting AI at a pilot phase increased revenue gains by 4-5% on average, continued maturation and AI optimization can increase revenue gains by as much as 10-12% (4).
3. Emphasize AI’s Positive Effect on Valuation
In addition to the near-term value of KPI progress, executives want to increase their business’s overall valuation. Although there are no formal rules that dictate how much a specific initiative will bolster your company’s valuation, there are a couple of key characteristics that investors and venture capitalists (VC) look for: namely, the promise of strong growth and the ability to create defensible positions.
Innovation plays a significant role in promoting growth. As previously mentioned, technological innovation - specifically AI - can increase productivity by introducing newer, more efficient processes. Increasing productivity enables businesses to meet higher demand, and demand may be driven upwards by innovative product design and improved user experience. Moreover, automating repetitive or cumbersome tasks with AI can reduce the time and labor required to run your business. When operating expenses are reduced, profit margins widen, and each dollar saved may be reallocated towards expansionary efforts. In these ways,
Increasingly expansive AI-powered innovation can help businesses to fulfill their growth potential, and raise valuation.
The creation of defensible positions refers to a business’s ability to maintain or increase their market share in the face of intense competition. While historically this may have been accomplished by some form of economic moat - think strong brand identity, unbeatable prices, or exclusive technology - when it comes to AI, the discussion tends to be framed around data moats. A data moat creates a competitive advantage for a business through proprietary data and models. If your dataset is valuable and difficult to replicate, you may be better equipped to provide unique services and features to your customers. An AI solution capable of extracting useful insights from a strong proprietary dataset can create a defensible position, which in turn may increase valuation.
4. De-Risk Your Project
De-risking is the process of gauging the benefits and feasibility of an AI solution before moving forward with a heavy investment. While de-risking is a necessary step in the development cycle of every AI project, it takes on redoubled importance in the context of a project proposal. In the absence of a thoroughly de-risked proof of concept, or previously established trust between you and your superiors, there is little reason for an executive to believe that your project will deliver on its promises.
The first phase of the de-risking process, also referred to as “kicking the tires”, is the time to sit down and ask yourself a series of big picture questions in order to identify fundamental failures and avoid falling victim to sunk costs. For starters you might ask:
Is the project technically feasible?
If so, will it bring value to the business?
If the project is rendering a product or service to consumers, is it desirable to them?
If the answer to any of these questions is “no”, it might be best to return to the ideation phase. However, if you can confidently answer in the affirmative, it would be sensible to test your intuition and develop a proof of concept.
Early on, you should try to determine the degree of perfection your model needs to exhibit to positively impact your business. You might reframe the question as: how accurate and efficient does my model need to be to perform as well as, or even better than, a human in the same position? Alternatively: how well does my model need to perform to be worth the investment? Whether or not you can achieve your preferred level of accuracy will be determined by the quality of your data, and the efficacy of your preliminary models.
It’s worth noting that some organizations are more strict in regard to data access, and you may need to make an initial proposal before you get access to the data necessary to de-risk your project. Assuming you can get access to data, you can start by inspecting a small sample to ensure it’s sufficient for the task at hand. In some cases, you may need to invest in ground truth data for training and testing your models, and it’s worth considering the rate, cost and approach to efficiently collect it. If your data is sufficient, you can begin labeling, training a rudimentary model, and getting some early predictors of efficacy.
For instance, if the ideal outcome is to successfully identify fraudulent insurance claims, at the end of the de-risking process you should be able to make a statistically credible claim along the lines of, “this model can successfully predict fraudulent insurance claims with approximately 83% accuracy.” Note: you may need to go beyond a small sub sample to a larger sample depending on the level of risk you want to address.
Once you complete the de-risking process you should be confident that you understand the risks of your AI project and can explain them. Ideally, you’d be able to support the assertions that your project will perform as intended with a reasonable level of proficiency, and increase in accuracy when given access to more data.
5. Calculate the Costs Involved
There will come a time when stakeholders want to put a number to their investment. However, as anyone who has been involved in large-scale AI integration will tell you, the multiplicity of factors influencing the cost makes it extremely difficult to provide an accurate estimate - especially for a new project. Luckily, if you’ve already taken the time to de-risk your AI project, you should have a better idea of the resources you’ll need to complete it.
One thing to consider from the outset is whether an off-the-shelf AI solution could work for your business, or if you require a more intensive, bespoke solution. If you opt for the latter, the cost tends to increase in proportion with the time and manpower required to pursue the project. If in the process of de-risking you discover that you’re going to need to invest in a significant amount of ground truth data, you can expect the time spent gathering and preprocessing to add to your estimate. Similarly, if you’re aiming for a high level of model accuracy and stability, you should expect to spend more time and money on the additional iterations it’ll take to get you there.
Lastly, take stock of the resources you have on hand. For instance, do you have the requisite skills in-house, or will you need to outsource to professionals? How large of a team will you need? Do you have access to on-premise servers for training and deploying your model, or will you need to purchase access to cloud servers?
While all these factors certainly add up, there are a number of considerations to keep in mind when presenting your proposal. By this point, you should be able to make the case that your AI project will deepen your firm’s innovation capability and move the needle on your KPIs. Additionally, if it’s your first project, any investment you make into infrastructure will serve to lay the groundwork for future innovation. While initial investment costs may be high, refusing to innovate can result in your business being left behind, which may ultimately cost you more than an investment into AI.
Trying to secure executive buy-in for your next AI project? Schedule a free consultation with one of our expert data scientists: HERE
SOURCES:
Alcik, N., Dilda, V., Görner, S., Mori, L., Rebuffel, P., Reiter, S., Samek, R. (2021, April 30). Succeeding in the AI supply-chain revolution. Mckinsey. https://www.mckinsey.com/industries/metals-and-mining/our-insights/succeeding-in-the-ai-supply-chain-revolution
Gow, G. (2022, August 28). The Labor Shortage Is Killing American Manufacturing. Here’s How AI Can Bring It Back To Life. Forbes. https://www.forbes.com/sites/glenngow/2022/08/28/the-labor-shortage-is-killing-american-manufacturing-heres-how-ai-can-bring-it-back-to-life/?sh=718bd11e7374
Transforming through digitalization. (n.d.) Nestlé. Retrieved October 27, 2022 from https://www.nestle.com/investors/annual-report/digitalization
IBM. (2020 November). The Business Value of AI. https://www.ibm.com/downloads/cas/ZENVBND4