Enterprise hits and misses - the modern data stack is messy, content marketing needs to change, and developers put AI to the test

Lead story - The modern stack has a messy data problem

Judging by the platitudes of data management vendors, I thought we had this data quality thing figured out. Not so fast, says Neil Raden, after attending a surprisingly vigorous panel:

Usually, these panels are dull, especially when the panelists are all founders or executives of software companies. This was different.

At issue? The so-called "Modern Data Stack," (MDS) and what it takes to achieve it:

One particularly memorable moment came when George Fraser clarified Fivetran's role in the MDS. As a data replication company, Fivetran focuses on replicating data into the desired destination without getting involved in any of the workflows intended by the user... However, as the MDS evolves, the "messy problem of data integration" will undoubtedly remain a challenge for all players in the industry.

Data management vendors are increasingly sophisticated, but: the messy data problem hasn't gone away. Neil quotes Fraser:

People need to realize that the sources produce very unclean data. And if you need to send the data to a relational database that supports updates and things like that, the data you will be looking at will be very ugly.

AI raises this problem to an fever pitch:

Timely and virtually frictionless access to data is a critical requirement for the expanding need for data science and AI/ML. The precious time of skilled practitioners is often spent managing data instead of building models.

What did this panel of luminaries propose to do about it?

Some panelists believed that federated data (e.g. distributed cloud data), is alleviating the messy data problem; others weren't so sure. Phrases like "stupid idea" definitely make for a memorable panel. Neil says the panel did agree on one thing: data lakes are not the answer. Bottom line: solutions to the dirty data problem are emerging, but it's not a one-size-fits-all situation.

Ergo, organizations in pursuit of an MDS can't ignore the data problem:

There are many solutions to messy data, but the MDS still needs to address it. That is Fraser's point.

Also see: Neil's Generative AI needs better context - you can't talk about Large Language Models without reverse neural networks and backpropagation.

diginomica takes our generative AI position: this week, we issued a rare team blog: Generative AI and diginomica - where we stand.

The diginomica editorial team will continue to document the undoubted potential, as well as the pitfalls, of all forms of AI on the enterprise. We have always felt that AI can aid in transformational projects - but only if the ethical questions of data privacy, bias and governance are thoroughly addressed.

However, we are drawing a firm line in the sand when it comes to the use of generative AI in our own content creation. Other media outlets enthusiastically proclaim their plans to explore the production of content via LLMs. Our readers are rightly wondering, given our in-depth coverage of this technology, if we would ever use these tools to write diginomica content.

The answer is a resounding no!

Vendor analysis, diginomica style. Here's my three top choices from our vendor coverage:

• Planful Perform 2023 - CEO Grant Halloran on the real impact of AI for finance - and why generative AI hype doesn't deserve a press release - "Wait - no generative AI hype festival at Planful Perform this year? No LLM-is-cool press release? As Planful Perform kicked off, I talked with CEO Grant Halloran about what really matters to customers - including Planful's own AI progress." - Jon
• Zendesk Relate - Zendesk AI packages up AI models for CX business users - "At its annual Relate conference, Zendesk launches new AI capabilities including some based on GPT technology - but LLMs aren't the answer to every AI requirement." - Phil
• Appian - where AI meets RPA and Process Mining - "The recent Appian World show was a masterclass on everything process technology but the insights regarding AI were big news, too." - Brian

Tableau Conference '23 coverage - seems like this annual convergence of data geeks and visualization enthusiasts is becoming more important. AI can help us tell stories with data but without the right tools/data in the right hands, we can't frame the story. Stuart's on the case:

So how can Tableau help as analytics "democratize" across the user base? Stuart quotes Tableau CPO Francois Ajenstat:

We all know that data is still too hard, not accessible to everybody. There are still too many people that don't have access to data to make better decisions. Only 30% of people in organizations use data to make decisions.

A few more vendor picks, without the quotables:

Jon's grab bag -  Madeline reports on Intel's supply chain diversity mandate  in How Intel builds diversity into its supply chain with $2.2 billion spend. Chris tried to make sense of the conflicting-reports-conundrum in UK ceding control to machines, say reports. Human leadership essential, says another.

And finally - and I do mean finally - my diginomica dbook on reaching the informed enterprise buyer is finally out. I posted about why I wrote it, and how I shifted from provoking marketers to providing a new framework (the "what are you going to about it"): Reaching the informed B2B buyer - our new publication explains why marketers are getting it wrong, and what to do about it:

B2B influencer tactics are stale. Sponsored research rings hollower with every white paper. Instead: earn your own influence. Elevate internal experts. Empower them to create their own content, and forge relationships with those who truly influence enterprise buyers. Top-down targeting? Or bottom-up? For B2B content strategy,we need both.

Best of the enterprise web

My top seven

• Building Trustworthy AI - Bruce Scheier with a different take on overcoming AI trust obstacles.
• What every CEO should know about generative AI - pretty goofy moving-glob graphic by McKinsey, but worthwhile piece: "It's out-of-the-box accessibility makes generative AI different from all AI that came before it. Users don’t need a degree in machine learning to interact with or derive value from it; nearly anyone who can ask questions can use it."
• Tech professionals need to 'get out of the data' and talk to people again - Joe McKendrick's puts the challenge outto IT. He quotes the CEO of UserTesting: "So, tech professionals face a few challenges in regards to customer experience, mainly they are not interacting with customers on a daily basis, if at all. They don't have that level of insight, which can be critical."
• S&OP: Should Not Be a Hammer Seeking a Nail - Lora Cecere is en fuego: "The good news is that we can attend conferences again. The bad news is that most discussions are based on opinions, not facts."
• The State of Visibility Into Performance of Digital Services  - interesting frame from Digital Enterprise Journal's research:
• Developers Put AI Bots to the Test of Writing Code - "With so many AI-powered assistants waiting to be explored, we’ve now entered the phase where excited users try their own homegrown experiments — and share the results online."

One more from The New Stack:

Whiffs

When you find yourself on video, be ready to wing it:

A classic "how not to introduce yourself" example from the open season sales cesspool business professional's home base, LinkedIn:

See you next time... If you find an #ensw piece that qualifies for hits and misses - in a good or bad way - let me know in the comments as Clive (almost) always does. Most Enterprise hits and misses articles are selected from my curated @jonerpnewsfeed.

</div