08.08.2025

ForwardThinking 🤔

How mature is your AI effort?

It turns out that in spite of the hype, most companies are not all in on AI yet. While there’s a huge dose of FOMO out there, the Economist reported this week that just 10% of firms are using AI in a meaningful way (whatever that means).

I’ve seen figures as high as 25%, but you get the idea: you may think everyone’s using AI, but apparently they’re not. They just like to talk a good game with 45% of the S&P 500 execs talking about AI in earnings calls in the first quarter this year according to that same Economist article. The gap between these two statistics is hard to miss.

That’s why it may seem, ahem, premature to be discussing your company’s AI maturity, but it’s a concept I came across in a couple of different places this week. In both instances, the implication was clear: the more mature your approach to AI, the more successful you are likely to be as an organization (even if most companies aren’t close to getting there yet).

Splunk defines a maturity model as a way to “measure an organization's capabilities over time: They provide a structured framework to assess progress, identify gaps and guide improvements in areas like IT, security or business operations.” In this case, it’s determining where your organization sits on the AI maturity spectrum. That requires some sort of framework.

Where are you now?

Consultants typically help companies understand how to get from Point A to Point B when a new technology comes along. Change is never easy inside large organizations and AI requires a level of change we may never have experienced previously. That’s why finding a way to measure is helpful.

Brian Solis, head of global innovation at ServiceNow and author of eight books, says it’s not so much the complexity of the technology itself, but the age-old organizational problems that have held companies back through every technological and business shift.

“Leadership has consistently tried to put all of these massive shifts and events into the box of ‘business as usual’ and AI cannot work in the box of business as usual. It really requires actual transformation, mostly from a data standpoint, which then requires shifts from a silo and operational standpoint,” Solis told FastForward in a recent interview. That means breaking down information silos and finding ways to put AI to work to improve your company’s operations.

Black and white picture of sign on gritty city sidewalk. Sign says: All Businesses Open As Usual.

Photo by Randy Laybourne on Unsplash

Ray Wang, founder and principal analyst at Constellation Research, published a blog post this week that is a good attempt at building a maturity model for AI. In Wang’s estimation, AI Luddites are at one end of the spectrum with AI Exponentials at the other end. It’s worth a look as a mental model as you try to assess your own company’s AI efforts.

AI Luddite is pretty clear. You don’t believe in the technology and you have no intention of adopting it, while AI Exponentials represent an ideal state that includes having a mature data model, being fully engaged with agentic AI and having what he calls an “excessive” investment in AI. The end goal here is to create an AI-fueled company with a strong digital workforce, the minimum number of human employees and the highest amount of revenue.

But fear not because you don’t go from one extreme to the other in one fell swoop. He identifies three in-between states: AI Followers, AI Literate and AI-first. To be clear, I don’t see any company anywhere close to being exponential. Right now the agentic piece is still very much a work in progress, as we have discussed frequently in this space.

So once again we find ourselves talking about aspiration versus current reality. With just 10% of companies meaningfully using AI today, that leaves a long way to go to reach Wang’s AI ideal.

Looking at the current state

To get a sense of the current state of affairs, I looked at the ServiceNow Enterprise AI Maturity Index report published in conjunction with Oxford Economics in May. The purpose of this report is to measure where public sector and private enterprise organizations stand on their AI journeys.

This is the second such report, so ServiceNow has a body of data to compare against, and what they found was surprising. Instead of the maturity index rising, they found it dropped off pretty sharply. They attributed this to the speed of change, making it difficult for large organizations to react to the pace and falling behind instead of moving ahead.

Orange streaks of light moving through city suggesting speed.

Photo by Jean Gerber on Unsplash

It’s not like it’s slowing down. As we saw this week, the target just keeps moving. This week’s announcements alone included an update to Claude, the release of OpenAI’s open-weight models and the long-awaited release of ChatGPT-5. That’s just one week. It’s no wonder large companies can’t keep up with it. None of us can.

What’s holding companies back? Well, besides the pace of change, ServiceNow identified measurement, governance and talent problems. As for the latter, when you have Meta offering mega compensation packages to anyone with AI expertise, what are mere mortal companies to do when it comes to attracting the finite supply of workers who can help accelerate the move up the AI maturity curve.

None of this is to say that this is impossible. The report found that in spite of these big obstacles, companies they call “pace setters,” defined as “an elite cohort of companies that are leading the pack across all dimensions of AI maturity,” are seeing much better results than typical enterprises, as the chart below illustrates:

Graph showing how AI Pacesetters dramatically outperform average AI users.

Graph courtesy of ServiceNow

As ServiceNow pointed out, even the most advanced enterprises are still in the early stages of AI transformation, which means you still have time to figure this out. You should, however, begin assessing where you currently stand in order to set clear goals for the future. Because if AI is the game changer it’s purported to be, doing nothing is not a viable option.

~Ron

What's new on the blog 📰

Ericsson’s CTO sees agents completely transforming enterprise software

Even at a 150-year-old company like Ericsson, AI isn't a new thing for them. It's something they've been working on for a long time. I spoke to chief technology officer Erik Ekudden about how they are incorporating AI into their network products. When it comes to agents, Ekudden believes agents could upend the entire SaaS business model.

“It’s almost—well, the question has to be asked—how many of these classical source systems do you want to retain, given that you can actually create so much more value by an agentic framework on top of the data that you have."

Read the full article >>

Thomson Reuters unveils AI agent for legal research as lawyers start to trust the tech

Lawyers aren’t known for moving fast, but when it comes to AI, that might be changing. Thomson Reuters, the company behind legal research service Westlaw, just launched a new AI agent built to mirror how a competent legal researcher works.

Read the full article >>

Statue of Justice blindfolded holding scales.

Photo by Possessed Photography on Unsplash

Box CEO Aaron Levie sees a bright agentic future in the enterprise

I've been having fascinating chats with Box CEO Aaron Levie about the latest technology trends for almost 15 years now. Early on when his focus was on pushing Box technology in the enterprise, he also had to evangelize SaaS as a concept.

Today his focus is shifting to agents, and we talked about why he sees agents changing the very nature of enterprise software in the future, and how this is very different from the AI chat experience.

“It really is going from a world where AI is this kind of read-only experience where you ask a question and it gives you an answer to where it can go off and do real work."

Read the full article >>

Photo by Possessed Photography on Unsplash

New Linux Foundation project takes aim at multi-vendor agent interoperability

When Cisco donated the AGNTCY project to the Linux Foundation this month, it could represent the first step towards agents from multiple vendors working together, but the absence of key players like Amazon and Microsoft could be problematic as the project tries to become a standard.

Read the full article >>

Why Oracle’s AI chief builds AI internally first

Oracle's head of AI and machine learning, Roger Barga, has been involved with enterprise computing for almost three decades with long stints at Microsoft and AWS before joining Oracle in 2022. Throughout his career, he has found the best way to build software for customers is to pressure test it internally before he releases it.

"Throughout my career, I have built services that first served our internal needs. You have a very active audience willing to give you feedback and engage with you very early on."

Read the full article >>

Photo courtesy of Oracle

News of the Week 📣

In a first, Some OpenAI Models Land on Amazon Bedrock

Amid a flurry of announcements from OpenAI this week including the long-awaited release of ChatGPT-5 and the first new open-weight models in years, the most surprising news might actually have been a new partnership — with Amazon, of all companies. In a first for the platform, OpenAI’s newly released open-weight models will now be available on Amazon Bedrock, AWS’s model management service.

The move is significant because OpenAI's models have been conspicuously absent from Bedrock, likely due to the company's deep partnership with Microsoft. While Bedrock offers support for models from providers like Anthropic, Cohere and Meta, the lack of OpenAI options was a major gap for AWS customers.

Two racks full of black weights with red numbers on them.

Photo by Brett Jordan on Unsplash

Open-weight is often used interchangeably with open source, but they are not the same. With open-weights developers can see the model's training parameters, known as weights. They can then download the model and run it locally or on their own infrastructure. Open source software typically allows wide access to the code base, while open-weight tends to be much more restrictive.

This doesn't mean you'll have access to OpenAI's proprietary models like GPT-5 on the Bedrock, at least not yet, but for the many enterprise teams working with AWS, having access to any OpenAI models directly through Bedrock has to be a welcome development.

SAP Acquires smartrecruiters to beef up HR software offerings

SAP announced this week that it was upping its recruiting software game with the purchase of SmartRecruiters, a SaaS recruiting platform founded in 2010. The purchase price was not disclosed.

The company plans to fold SmartRecruiters into its SuccessFactors platform, a company it acquired in 2011 for $3.4 billion.

Holger Mueller, an analyst at Constellation Research, says a move like this is a long time coming and the company's approach to human capital management (HCM), the software category where these companies fit, was in need of some updating.

"SAP has underinvested in its HCM offering for the last several years and now needs to catch up. Re-thinking talent management from the recruiting side is a good design approach, but will have to result in complete re-design of talent management for the AI era," Mueller told FastForward.

Photo courtesy of SAP

He sees SmartRecruiters as a good fit for SAP because they have a lot of enterprise customers in common, but that SAP still has to prove that it can make all of this software work well together. "SAP now has to show it can integrate SmartRecruiters in a seamless way as it competes with organic, in-house talent acquisition offerings from Oracle and Workday," he said.

SmartRecruiters raised a total of $225 million, per Crunchbase. The most recent raise was $110 million on a hefty $1.5 billion valuation in July 2021, at a time when many companies were valued highly. Regardless, the transaction is expected to close in the fourth quarter this year, pending regulatory approval.

Genesys gets $1.5 billion funding infusion from salesforce, servicenow

Genesys, an enterprise cloud customer service vendor, recently announced funding from two competing companies: Salesforce and ServiceNow. Each one contributed $750 million for a $1.5 billion cash infusion for the company.

Genesys has been around the block a couple of times, having been founded back in 1990, before the web, before the cloud, a long time ago. It went public in 1997 before being sold to Alcatel Lucent for $1.5 billion in 1999. Today the majority owners are Hellman & Friedman and Permira, a pair of private equity firms.

So what do two rival companies fighting for modern AI business in the enterprise see in a legacy customer service company that makes them both want to invest in it? Brent Leary, founder and principal analyst at CRM Essentials, says the two companies see value in the Genesys data.

People working in customer service center with head sets on.

Photo by Charanjeet Dhiman on Unsplash

"I think it just shows the importance of all the data coming out of all those contact center interactions. Whether those interactions are initiated and handled via self-service or escalated to an agent, there's tons of data that can be used for [model] training purposes," Leary told FastForward.

"And because a majority of customer interactions come via the contact center that data is incredibly important as it can be used not only to improve support but also product development, product marketing, sales and CX," he said.

It's worth noting that Salesforce Ventures, the venture arm of the CRM giant invested another $580 million in Genesys in 2021.

My adventures in AI 🤖

Photo by Andy Kelly on Unsplash

I was using Gemini as my editor this week. I asked it to do a final review of my document before I moved it into WordPress for publishing. It generated a table of suggestions as I asked, and the one that really got me was a suggestion that I switch something I had written as two sentences into a single sentence.

It's funny because I had broken it into two sentences in an earlier draft, per Gemini's suggestion (which I agreed with). So which is it? One sentence or two? You just can't win sometimes with these LLMs, apparently.

This post originally appeared on my LinkedIn.

What I'm reading 📚

Person reading book, but all you can see is their hands holding the book. They are sitting on a hardwood floor steeped in shadow, creating an interesting look.

Photo by Blaz Photo on Unsplash

RedMonk

The Cyber Resilience Act: A Five Alarm Fire

Stephen O'Grady

Wall Street Journal

Thanks for Your $1 Billion Job Offer, Mark Zuckerberg. I’m Gonna Pass.

Berber Jin and Keach Hagey

InfoWorld

Bridging the trust gap in AI-driven development

Matt Asay

Paul Kedrosky Blog

Honey, AI Capex is Eating the Economy

Paul Kedrosky

Look who's talking 👄

"A great reminder that letting startups grow into independently successful businesses, rather than be bought up by existing giants, can generate enormous value. A win for employees, investors, innovation, and the public."

- Former FTC Chair Lina Khan on X talking about the value of allowing Figma to go public, rather than getting absorbed by Adobe.

Please subscribe to the newsletter and feel free to share.

3350 Virginia St, Suite 262, Miami, FL 33133

ForwardThinking 🤔

How mature is your AI effort?

Where are you now?

Looking at the current state

What's new on the blog 📰

Ericsson’s CTO sees agents completely transforming enterprise software

Thomson Reuters unveils AI agent for legal research as lawyers start to trust the tech

Box CEO Aaron Levie sees a bright agentic future in the enterprise

New Linux Foundation project takes aim at multi-vendor agent interoperability

Why Oracle’s AI chief builds AI internally first

News of the Week 📣

In a first, Some OpenAI Models Land on Amazon Bedrock

SAP Acquires smartrecruiters to beef up HR software offerings

Genesys gets $1.5 billion funding infusion from salesforce, servicenow

My adventures in AI 🤖

What I'm reading 📚

RedMonk

The Cyber Resilience Act: A Five Alarm Fire

Wall Street Journal

Thanks for Your $1 Billion Job Offer, Mark Zuckerberg. I’m Gonna Pass.

InfoWorld

Bridging the trust gap in AI-driven development

Paul Kedrosky Blog

Honey, AI Capex is Eating the Economy

Look who's talking 👄

Subscribe

FastForward #32: How mature is your AI effort?

You may also like...