Systems integration is getting disrupted. The best systems integrators (SIs) are telling me that selling projects is harder than ever.
Customers aren’t sure if implementing legacy software is a good idea. With the pace of innovation, the system could be obsolete by the time it is live.
Then there is the actual work of implementing. AI dramatically accelerates implementation work if you’re using it right. That’s not great if you charge by the hour.
Firms are scrambling to redefine themselves. Most firms now believe that without major changes they will die.
So lets go back to what SIs even are. An SI is a group of people who have expertise. SAP SIs know SAP really well, and they build expertise on certain industries to differentiate themselves.
They are defined by their expertise in:
- Systems
- Industries
- Regions (language, regulations, physical presence)
SIs should focus all of their energy on making those strengths more powerful with AI.
Systems expertise means knowing which knowledge bases to tap for product knowledge. It means knowing how to customize the system for a gap that’s been filed on several projects. It means asking the right questions at the right gates and following the right methodology.
Industry expertise is hard to build. It requires a unique mix of systems expertise for a certain industry. Knowing what the system’s gaps are for that industry, and having already thought about solutions and accelerators to fill the gaps is a huge asset. Previously this was a lot of manual work and hard to package. With AI, it probably takes the form of questions, validations, and guidance which are all package-able.
Regional expertise is knowing what the Brazilian tax gaps are in SAP, and knowing how to fill those gaps because you’ve already done it. It’s also being able to run a workshop in Portuguese and making the outputs useful for your english speaking colleagues.
The success of projects is subjective. Many will say that if a project is double over-budget, it has not necessarily failed. So “fixing” a project with Tech is hard given the different incentives from both sides. The SI makes more money if the project costs more. The customer wants to go live faster and cheaper, while realizing the value for which they set out to do the project in the first place.
If we focus on making the SIs shine on the things that make them unique, will they want to sell the enabling software to their customers? Does it need to be their own? And is that any different to SIs that built their methodologies on top of Jira?
SIs have never been good at building products. They are good at implementing and customizing products they don’t own. AI has made it easier to build software. Some will think that means SIs can now be good at building products. I don’t think so.
SIs should focus on making their systems, industry, and regional expertise compound. That’s possible because AI can help you scale things that depend on unstructured data, but it requires great products. Not one built by the guy who had some time on the bench between two projects.
I need to explore how the Palantir model challenges these ideas. I’m excited to do that soon…