Amaresh Tripathy
Co-Founder and Managing Partner

The enterprise data stack is going through an inflection point and the battle lines are being redrawn- which happens every decade or so. This time, to no surprise, the AI agenda in Enterprise going to determine the winners and losers.

And the data platform battleground this time is about about the enabling the building of AI applications. To do that, the trick is to have an answer around the following four areas:

  1. semantic layer: understand the nuances of business data across the enterprise,
  2. app building low code no code toolset: connect them to tools to build applications, 
  3. simplicity of orchestration: have workflow and interoperability which makes the end of the folks making applications and managing the ecosystem easier 
  4. AI capability: with powerful enough AI frameworks that get the job done

Snowflake, Databricks, Microsoft and Salesforce are most aggressive in taking positions to be the data platform that allows the end to end capabilities to build out the AI applications. All have made big moves over last twelve months on building / partnering on their AI capability (Cortex AI from Snowflake, AI playground and DBRX from Databricks, Einstein from Salesforce and OpenAI partnership from Microsoft). Beyond that:

01. Databricks is anchoring on data cataloging space with its Unity catalog making it easier to understand the data context. Application toolset strategy is less clear

02. Microsoft is anchoring on Fabric which has a good enough simplicity orchestration play and also decent app building toolset. Semantic layer approach is its weakness

03. Salesforce has everything for the customer applications - but most of customer applications also need data outside its data cloud which is going to be a challenge it has to overcome as enterprises will not have a lot of incentive to do that. So the richness of what applications can be developed may become a challenge

04. Snowflake in some way is under siege from all the players, but has all the elements in place. And it is able to maintain its simplicity of orchestration because it is a walled garden. It has to figure out how to trade off experience / economics and expand to app building with business users

On the enterprise side, data teams and application teams are rarely under the same leader nor are they tightly connected.

So if the AI applications becomes the battleground then there will be new winners and losers in the Data Platform world and potentially changes to the enterprise IT op models.

AI impact to software engineering jobs

Anecdotes are great but what does the data say?
Let us see some interesting insights that come out of analyzing 20M job postings over 16 month...<br>

Read More
Three things to get right when scaling your data analytics operation

I wrote earlier this year about how the heightened pressure and need for quick, accurate decision-making during the pandemic had helped to boost...

Read More