In the third article in our series, let’s consider the final steps before putting pen to paper for signature. In the first two posts, we’ve looked at what elements need to be codified before your team starts looking at AI options, followed up by a guide about how to sort through vendors, and which ones to take seriously. If you haven’t had a chance to review these please do so, they contain important markers that help define our field of vision.
We will get to deployment and success in subsequent articles, but before we do there is one important step. The actual purchase and planning phases are vital to ensure that your project launches on-time, with the right resources and is anchored with the right expectations.
#1 Making the Purchase
Getting procurement or finance approval for AI projects can be an unexpected variable that hinders an on-time deployment. AI is new, and if it’s the first such project for your business, it’s untested. There may be implicit biases or skepticism of the technology you need to overcome. Here are a few things you can do to move this process forward:
- Uncover your business buying process. Who needs to sign off? What do they need in order to approve? Tailor the business case accordingly.
- Ensure the ROI is reasonably calculated. Project the value of automated work, and what new valuable work can be created to get to this figure.
- Take the time to formulate a coherent business case, folding in the ROI calculation. Find out what the executive sponsor, procurement and finance need in a business case to sign off. Work with your vendor here, they will have experience getting AI signed off with other companies.
- Leverage your executive sponsor. They will have sway in the buying process and can influence other approvers, especially if you involve them early.
The other part needed to complete the purchase is the technical win. You’ll have to run a technical due diligence of your infrastructure, systems and available data to sufficiently plan for AI onboarding, and importantly, to assure the CIO/CTO that there will be minimal (or ideally, no) disruption to their systems. This process will help you earmark resources at the outset, which sets expectations and ensures your costs are in-line with the project scope - nothing derails AI projects like misaligned understanding of requirements and investment. Technical diligence may take one-two weeks but it is time well spent to ensure the project can deliver to business expectations.
Armed with business, procurement/finance and technical wins, you’re off to the races and can begin implementation.
#2 Planning for Success
Any vendor worth their salt will launch a planning, implementation and onboarding process post-purchase. However, it’s critical to run your own exercise pre-purchase so that you define the resource requirements, key milestones, and success metrics as part of the business approval process. You’ll want to run this process during the business approval stage.
Planning will commit stakeholders to delivering resources and prevent executive mission creep. Here are some key questions to resolve in the process:
- Who is the leader? This is someone who will work across the main business lines and take the project from kick-off through to onboarding and adoption. They will also need to speak the language of the end users as well as executives, and manage the vendor. AI literacy helps, but is not required - either way the leader will define the future of AI adoption for the business.
- Who is providing air cover? Oftentimes the executive sponsoring the purchase is not the executive responsible for the project. Work out who this person is, and ensure they will receive sufficient and consistent information to educate and enable the management team. Keeping them onside ensures the project stays on track and helps expand it long term. That is, after all, why you’re bringing AI onboard.
- What’s the communication strategy? Define how and when to communicate progress - is it a weekly or monthly cadence? Who should be included, just management, the business unit, all impacted teams? And if successful is there an external communications opportunity? Successful AI projects generate good headlines and attract business. Base this decision on internal expectations and the level of interest or scrutiny the projects receive. Well defined communication will help attain buy-in and ultimately inform the success of the project.
- What’s the plan? The nature of the project will define the plan. Pilots will work differently to multi-year agreements, success-based projects will have a different rhythm to time-anchored projects. Either way, set down milestones for reporting and the results that should be achieved at each stage, so that your team, and your vendor can get you there. As projects unfold, unknown variables tend to emerge, a set timeline with priorities is a good way of guarding against them, and ensuring you get value from your vendor.
With these elements, you will be far better positioned to start the project on-time, with appropriate resources in place, and with the correct expectations set. Now you’re ready to bring your vendor onboard.
In the coming articles we’ll address deployment and driving success in AI projects, and while these are the obvious topics after deciding on your AI provider, don’t overlook these critical steps in getting your project firing on all cylinders.