Midagon Company logo
Back to blog

Are you ready for Artificial Intelligence in sourcing & procurement? [Part 1]

Are you familiarized with all the ways in which Artificial Intelligence (AI) can benefit the area of sourcing and procurement? Have you taken all the necessary measures to fully utilize the possibilities that are available? If you think there might be something left to do, here are some thoughts on how you could take the initial steps.

Erik Kolppanen

Erik Kolppanen

Five steps to taking advantage of AI opportunities in Sourcing & Procurement

The approach to understanding AI opportunities and their value-add for your organization, as well as implementing AI technologies could look like the following:

  1. Understand - Maturity check and brainstorm: Do a maturity check, current state study for identifying business opportunities (analyze current manual tasks, main business value drivers, main pain points) and list capability requirements.

  2. Feasibility check: Scan various AI technologies and ideas, get broad AI training, carry out technical feasibility and business value check for identified ideas, capabilities or opportunities. Trial or do proof of concept (PoC) in a small scale.

  3. Plan: Align strategies, roadmap and manage portfolio - Funnel business opportunities and capabilities based on expected business benefits and take it as an input for aligning the AI, business and IT strategies. Roadmap the opportunities/capabilities for supporting the strategy goals and channel those to the capability portfolio management – cluster, plan and prepare proposals for capabilities initiation, decide in-house or outsource approach.

  4. Implement: Project / program management (preferably Agile) – Make the investment decision, resource and execute the project and develop and deploy capabilities for business use.

  5. Realization - Business value follow-up: Check how the business was impacted, how the business case was realized and what was learned.

You would need a professional - who is familiar with the best practices, available AI technologies and has a realistic view of its applicability and capabilities - to start mapping your processes’ current state and maturity. Here, brainstorming would be a good approach to take. Some typical questions could be asked: Would you like to automate any tasks? What are your main business value drivers and how do you impact them? What are your pain points and how do you eliminate them?

Based on the outcome, develop an AI-based opportunity and capability requirement list with a high level of possible benefits. You should discuss the AI possibilities and ways to proceed with your leadership team. Perhaps the way to go is to hire an AI Leader for carrying out all the preparation and coordination related tasks and for kicking-off and managing the AI related activities within your organisation.

Next, plan the way to proceed with a business and technical feasibility check. Based on your opportunity and capability requirement list, you could even select some preliminary technologies or platforms to try out the idea on a small scale. Many aspects of feasibility should be validated. This includes functionality, attainable benefits, scalability, user friendliness, security, integration and costs. This would require a small trial budget reservation for either IT or business. Since most of the solutions relevant for trialling and PoC are cloud-based, it would be most important to check now if the planned approach is consistent with your organization's information and IT security policies. It is also important to know if the data business owners would be willing to support the approach by providing access to their data. When preparing for trialling and PoC, at the latest, you should engage, hire or outsource data analyst and scientist resources.

While funneling the opportunities to the roadmaps, capabilities could be categorized into three major categories:

Align the opportunities with AI, business and IT strategies

Based on the value for the business (initial business case and feasibility checks/outcomes from trials) and how easily they can be implemented, the opportunities should be aligned with AI, business and IT strategies.

Once the opportunities and capability requirements have been identified, the capabilities should be planned into the corresponding portfolio, clustered based on division principal agreed to within the company (based on process ownership, organization responsibilities, implementation responsibility etc.). Efficient portfolio management would ensure that the best possible implementation plan is taken for all opportunities for maximizing the business benefits with scarce available resources for the time span concerned. Within each portfolio, the opportunities should be placed into the roadmap and planned for accordingly.

Based on the agreed portfolio roadmaps, plans and approvals for the relevant projects and programs would be kicked-off. The projects of this type are usually agile type of projects/agile release trains/product backlogs, where small development increments within sprints would be implemented. However, if there is a platform type of establishment as a key enabler, the traditional waterfall type of project approach might be more sensible in the beginning.

Once the implementations or development incrementals are ready and deployed, it is time to review what has been achieved and check if it brought the planned and desired benefits. The initial business case should be reviewed and corrected with the achieved outcomes and the lessons learned. Typically, this could happen three to six months after the deployment, at the latest.

Finally, the experience and knowledge gained from such projects should be channelled back to the opportunity requirement lists and roadmaps for identifying and processing new ideas or eliminating ideas which are not feasible and documenting why not.

In my next blog we will give some concrete examples of how AI technologies could be applied within sourcing and procurement.  

More posts