deal architect

Revitalizing the ERP franchise

We saw Mission: Impossible - The Final Reckoning last weekend. Lots of flashbacks to the previous 7 movie episodes. And made me relive some of the amazing stunts Tom Cruise has performed across them including a deep, freezing dive into a sunken Russian submarine, crawling up the Burj Khalifa, a HALO jump onto the Louvre, riding on the outside of planes, spectacular motorcycle rides, knife fights and many more.

Lots of people don’t know about the trajectory of that franchise. It was a TV series from 1966-73, revived from 1988-90, and has reached stratospheric levels with the last few movie episodes.

Cruise is 62 now and yet his stunts keep getting more and more ambitious. He recently shared 20 photos on X of some of the MI stunts. Not just as Ethan Hunt or as Maverick in the Top Gun series or in other action films, as he says in this interview his goals are endless. He even wants to make musicals. He is a one-human industry. There is a joy, a passion in him which is infectious.    

I also like to track long-running music franchises. I bought a copy of the Netflix documentary on the Eagles a few years ago and have watched it quite often. The two principals, Glenn Frey and Don Henley had an amazing education in modern music. Glenn had Motown and Bob Seger influences in Detroit, Don went to music school and was exposed to country, blues and other genre growing up in Texas. Their California influencers included Linda Ronstadt, Jackson Browne, David Geffen and others. The UK producer, Glyn Jones introduced them to many legendary British bands during their time in London. The band broke up in the late 70s, but continued to be wildly popular as staples of music radio in the 80s, and due to the solo efforts of the two principals. The band has mutated multiple times since, and continues to appeal most recently with a highly applauded series of shows at the Sphere in Vegas.    

I also have a copy of Crossfire Hurricane – the documentary on the Rolling Stones. The Stones were never my favorite band. I am more of a fan of Led Zeppelin, Eagles and others.

However, it is always amazing to see the bad boys of the music industry continue to thrive almost six decades after they were formed. As they like to say. the Beatles were the clean-cut “white hats”. The world needed a mirror opposite.

They continue to evolve from their early roots as a blues band. This article describes the various twists and turns which has been the key to their longevity and their ability to attract the next generation of fans – over and over.

We similarly have a long running franchise in the tech sector. It is that around ERP software.

It started life as a custom application called Materials Requirement Planning (MRP). In the 1960s, J.I. Case, an industrial manufacturer worked with IBM to develop what is considered the first MRP solution. The scope expanded to the manufacturing shop floor and the acronym changed to MRPII – with M now for Manufacturing. Buyers were also now buying the functionality as packaged software, not custom building it.

My first exposure to that came as a PwC consultant when I implemented mainframe and AS/400 based packages in the 80s for customers around the world. In the 1990s, Gartner coined the term ERP for Enterprise Resource Planning and pushed the industry to expand the functional scope to include headquarters, distribution and other business processes. I had a front-row view when I joined Gartner in 95. It turned into a rocket ship as customers looked at ERP as a Y2K solution, as a replacement to consolidate countless departmental applications, as a shared service solution across borders and countless other use cases.

When I left Gartner in 2000, I was confident ERP would continue to evolve and thrive. At Deal Architect, I have advised clients about ERP strategies, taken many to delivery centers in India, E. Europe and elsewhere. Have written several books and blogs where ERP case studies are prominent.

In some ways ERP has evolved nicely.  In many other ways, it has disappointed. I have talked to CIOs who say they went out of their way to avoid a second ERP project showing up on their resume. I have seen young IT professionals avoid anything to do with maintaining ERP applications or upgrading to the next release.  Definitely nowhere as popular as Ethan Hunt or Mick Jagger.

What happened?

Lack of vendor ambition

I have said often the enterprise software industry has been blessed with fantastic technical architects – Hasso Plattner, Larry Ellison, Jan Baan, Parker Harris, others at Workday and more. ERP has moved from mainframe to client/server to cloud now to agentic architectures. Each of those shifts has been a heavy lift. But ERP is not just about IT infrastructure – it is much more about business functionality. You need just as smart business architects who can vividly re-imagine business processes. That has been lacking.

If Jeff Bezos had called Larry Ellison in 2010 and said “Larry, Amazon is building massive fulfillment centers. What software can Oracle sell us?”,  Larry would likely have said, ”Here's our warehouse management system. I'll put it in the cloud for you.” Hasso would have responded similarly and additionally reasoned that HANA would handle the loads better than the Oracle database. Neither would have likely re-imagined the warehouse saying “These are going to be huge centers. We don't want to exhaust employees walking too much, carrying back-breaking loads. How about we bring the work to the worker?”

Amazon re-imagined it – see my write up from a tour of a center here. They have hundreds of Kiva robots. They have thousands of sensors everywhere. They have packing stations with all kinds of cutting and printing devices.  They have miles of conveyor belts, so people don't have to walk or lift much. They designed a “chaotic” storage paradigm that no human needed to master. They completely redesigned the warehouse.  

In its early evolution, ERP moved from Materials to Manufacturing to Enterprise Planning. In the last two decades, it should have evolved to Global Planning, it should have absorbed CRM, SCM and other categories and it should have gone beyond Manufacturing into the broader Services economy. Services as in Healthcare, Insurance, servitized Industrials with unique books of record for patient records, claims, digital twins of assets and more.  It should have gone global with localization for new taxes, customs, payment protocols and more around the world.

We barely scratched the surface. Then we embarrassed customers to just accept the limited vision and not make customizations with advice like “Keep the core clean”. So customers ring-fenced their ERP solutions with 10, 50, 100 best of breed solutions.  So much for the promise of an integrated ERP suite.

Not expanding the functional footprint also meant their developers and their new AI agents do not have access to operational or global domain expertise or unique data. Their AI use cases still focus on SG&A processes (ironic given ERP ‘s operational roots)  and are weak relative to the GPU and LLM resources that are being thrown at them. It is the equivalent of Ethan Hunt repeating the stunts he performed in MI episode 1, and expecting audiences to continue to cheer and pay for expensive popcorn in aging movie theaters.

Distorted economics and project failures

If you prompt your favorite GenAI tool about ERP issues you will undoubtedly see cost overruns and project failures listed. You can find hundreds of articles on them. My books have plenty of footnotes which point to them.

In an industry which prides itself on Moore’s Law and economics tied to innovation and productivity, ERP has become known for “lock-in” economics. Even after 25 years. your annual maintenance or subscription costs often remain the same or are higher with inflationary increases. The sector has used lawfare against competitors and even customers with an entitlement mindset about those economics.  

Another major cost element in ERP projects is migration costs. There is plenty of discussion around greenfield, brownfield, rapid implementation etc. but few have addressed the elephant in the room. ERP migrations are still heavily dependent on human labor, especially those staffed by SIs. In the last couple of decades, the industry has done hundreds of thousands of these migrations. Each has artifacts around data conversion, acceptance testing, end user training, some have code conversation, global rollout templates. Feed them into an LLM and have it generate the first draft from the next project. Have humans work on the deltas for this new project. But most SIs will have humans start from scratch.

McKinsey has brought many of its knowledge assets into its Lilli tool whose goal is “Build a generative AI platform powered by our firm’s knowledge to shorten the time to insights for our teams and the time to impact for our clients, and rewire the way we work.” . Every ERP vendor should push its SI partners to show similar. And factories to automate data and code conversions, specialize with testing tools, create training modules. This will reduce cost. Importantly, it should improve project predictability when you are not as dependent on new, often inexperienced, staff.

In the last few years, we have piled on hyperscaler costs on top of software and SI costs. If that followed Moore's Law, you would applaud that. But given the lock-in mindset in the sector, the jury is still out. Any wonder that less than 20% of last-gen ERP sites have migrated to the cloud versions and are not chomping at the bit for the new AI agents? 

Analysts - not raising the bar for vendors  

I was part of a group of analysts Gartner started hiring in the mid-90s to expand their business applications coverage. It launched the Administrative Application Services (AAS), Customer Relational Management (CRM), Integrated Logistics Services (ILS), and several vertical ones starting in 1995. Its coverage of these areas started to get richer and soon there were all kinds of Magic Quadrants and events focused on the categories.

I presumed that would continue when I left Gartner in 2000. I assumed they would have Magic Quadrants for ERP-I (Insurance), ERP- R (Retail) and push vendors to showcase vertical books of record. I presumed they would launch MQs for ASEAN countries, EMEA etc. Not just Gartner or Forrester – I presumed that specialized vertical or regional analyst firms would cover those opportunities. Has not happened and vendors continue to score well on analyst reports based mostly on their horizontal functionality and English speaking country capability.

Again, if you are selling tools for IT infrastructure that is less of an issue – but when it comes to business applications it is critical. In the end, it is caveat emptor so it is a bit of a cop out to mostly blame analysts, but they should have been howling that after two decades of cloud ERP, only 25% of the grid of applications by industry and country are adequately covered by ERP vendors. They should have developed TCO models for ERP projects and pushed for Moore’s Law improvements. Yes, I should have also howled louder.

So, is an ERP revival likely?

I live in hope we can shake up the ERP sector. Having spent 4 decades looking at every aspect of the sector, I would love to see more customers proudly talk about their ERP software and more young professionals clamor to work on it. For its costs to come down nicely and for outcomes to improve. It will take a new generation of PE firms, VCs and executives and analysts to revitalize it.   

Part of me says I already saw the golden days of ERP when I was at Gartner in the 90s.As I wrote here, back then BOPS (Baan, Oracle, PeopleSoft and SAP) were Wall Street darlings. Now ERP shops barely earn 10% of the market cap of the Magnificent Seven stocks. One exception is Microsoft but its ERP revenues are tiny. Its Azure cloud, its OpenAI relationship and its Office franchise explain its high valuation.

The optimist in me says the old ERP dog still has a few tricks left. Actually it is still a young one compared to Keith Richards of the Stones, now in his 80s.  The running joke about his longevity  is “Only cockroaches and Keith will survive an apocalypse”

Christopher McQuarrie who has directed the last few MI episodes said he had to stop Tom Cruise from doing the 20^th retake of a stunt with “You’re done. Do not anger the gods.”

It would be nice to see ERP vendors show some of that drive and ambition. Please, please tempt the innovation and budget gods!

As the Eagles famously sang “Take it to the Limit…one more time”

Enterprise AI – a post-SAP Sapphire Update

Bob Evans of Cloud Wars and I recorded a Burning Platform episode about Agentic AI just prior to SAP Sapphire in Orlando a week ago. I have embedded the video below and I am updating here with some thoughts from the Orlando event

Bob and I talked about the war of words between software CEOs even as AI agents released so far are fairly narrow in their reach. Bob commented it’s time for vendors to not just nitpick at each other but to lift us up, and talk more about the impact they can have for their customers. Bob also wants the sector to become more ambitious in its agentic definitions and said “All my life, I've had this idea that all limitations are self-imposed. We choose to say, I can't do this, I can't do that, I'm too big. I'm not fat. You know, whatever, we can always make excuses.”

I added “To me, the excitement about agents is you can closely tie them to enterprise value. If you can mirror agents to individual occupations and roles, and you can have a junior employee punch above her weight using an agent, or if an agent can allow humans to massively scale outcomes, you can far easily quantify productivity payback.”

I also made the point we need to go beyond Agentic AI to what Jensen Huang of NVIDIA terms Physical AI to include robotics, UAVs etc. That would allow them to assist in industrial, farming, construction and other physical world settings.

SAP did fine at Orlando. None of their executives, at least publicly, took jabs at the competition. ServiceNow only showed up in one of their demos (and unless I missed it, without any snarky comments). One of their partners wrote in a blog about the competition “SAP doesn’t need to tear down a sales tower, a workflow company, or an HCM+Fin vendor to prove its place.” Fairly harmless. And SAP showed off a Neura cognitive humanoid robot in one of its demos.

To its credit, SAP also announced several what I would call agents differentiated from competition which leverage its broad portfolio of applications. They include Quote Creation, Maintenance Planning, Catalog Optimization and Field Service Dispatching among others along with others in Finance and HR areas.

The key word for now is announcements. SAP threw out promises of 30% to 70% improvements in various presentations. We need to see customers talk about their experience with these agents and validate the payback.

However, when it comes to really putting a distance from its competitors, SAP could have done a lot better and focused on agents in several verticals it is strong in.

Look at the total addressable market. The Bureau of Labor Statistics tracks over 800+ occupations and for each it identifies skills that can be mapped to human faculties. And, btw, the BLS only updates its SOC table every decade or so – the last time it did so was in 2018. Many more new occupations since.

As I pointed out to Bob, today in 2025, there is already a vendor offering a much broader range of such agents. In our latest book, The AI Analyst, there is an SV automation vendor called Polestar which has an expansive definition of its verticals as each of the occupations the BLS tracks. Their solutions automate limbs, eyesight and other human faculties, not just cognitive skills. Polestar has a slick video which show its products on shop floors, hospitals, labs, delivery trucks, and warehouses. One of its most successful marketing channels is its relationship with Hollywood studios – many of its bots and agents show up in movies and TV series.

Polestar’s CEO, Barry Roman became convinced about a focus on micro-verticals when he asked his team to bring in his father, a retired fireman, as an adviser while working with a fire and rescue customer. His father commented “Barry, all they have bought from you is some crew scheduling software. Give me a team for a week and I will show them technology a modern fireman can use and you could be selling in addition” That opened their eyes to wearable sensors, augmented reality, thermal vision glasses to allow firemen to see through smoke and fire, drones with sonic fire extinguishers to create acoustic boundaries to prevent wildfires from spreading, using AI to monitor imaging feeds from fires and so much more.

Barry never fails to point out to his sales team how easy he had made it for retailers to contract their warehouse bots. “Not everyone is an Amazon and needs large distribution centers. We saw a growth in smaller warehouses located close to customer demand. I told Swanson (the CFO) to offer our bots as-a-service, and deliver as few as five as a start. Then seasonally we would deliver five or 10 or 15 more bots, as their customer demand spiked. Customers loved this flexibility, and it was one of our most successful product introductions.”

Our book has many more agents, copilots and other automation. It is a fiction mystery where, an Industry Analyst, Patrick Brennan, who used to work for Polestar and helped its explosive vertical growth, works with the FBI in the search and investigation in the disappearance of Barry.

Even in the non-fiction, real world. customers are being similarly ambitious. Jamie Dimon of JPMorgan Chase was recently quoted (thanks to Larry Dignan of Constellation for the pointer) as saying they will have nearly 800 AI use cases by the end of the year including those that support hedging and equity trading floors. He says “it will affect every job, every application, every database and it will make people highly more efficient”

We are in the early innings of agentic rollout. And if enterprise vendor CEOs just nitpick on each other and not starting thinking more ambitiously, they are leaving the market open to Jensen and Elon  Musk with his robotaxis, Starlink satellites and Optimus humanoid robots), to CEOs like Jamie and yes even the budding Barry Romans.

And, us analysts need to become more like Patrick Brennan and encourage vendors to think more ambitiously. There will be room for billions of agents across thousands of categories. Time to open our aperture much wider.

SAP Sapphire – “Change isn’t going anywhere”

I have told this to SAP executives a few times. They should not let analysts or competitors narrowly define their “category”. They are a category unto themselves.  They have hundreds of thousands of customers across the world. Across countless verticals. They also have the widest range of partners. However, they tend to be comfortable in ERP and back office world.

So, I was excited to see CEO Christian Klein’s kickoff keynote at Sapphire this week billed as “Change isn’t going anywhere—but the right strategy can help you overcome anything.”

Given all the wild changes around the world already this year, my expectations of “anything” were high.

The first customer to present in his keynote was NBCU – a huge, impressive media and entertainment company. Nothing wrong with them going first especially since Universal has a massive new theme park opening not far from the conference location this week.

Far better from my pov would have been a retailer showing off how SAP has helped them rapidly redo pricing and profitability scenarios in the chaotic tariff regime unleashed by the Trump administration.  

Or an Apple which is re-balancing its iPhone production from China to possibly India and even the US. No SAP competitor could boast a more impressive supply chain scenario which CEO Tim Cook engineered decades ago.

Instead, SAP had a more tactical Joule AI supply chain use case and had Standard Chartered talk about AI in their back office . I would preferred a use case built on SAP’s global database of country regs and business practices they have localized for over 60 countries. And perhaps have a global bank talk about their customers as they look at Vietnam v Mexico v Country Z as they redo their China +1 strategies.

Talking about another much needed AI use case - E&Y CEO Janet Truncale got a 15 minute slot, but did not talk at all about how her consultants are using LLMs, data and code conversion factories and other automation. They are essential to deliver to Christian’s promise of faster, more streamlined projects. SI project costs are one of the biggest gating factors for 60% of ECC customers who have not made the move to S/4HANA.            

Same thing with customers who were profiled in Thomas Saueressig’s keynote. I would have liked a nuanced approach to the questions they were posed. So, I would have liked Mars to talk about how the new health regime in the US will affect their additives and flavorings and the impact on recipes SAP helps manage. Phoenix Global on likely expansion of steel making in the US and related impact on the industry’s Capex and Asset management systems. And not just about S/4HANA, but I would have liked Mercedes to talk about the maturing of autonomous vehicles and how their US manufacturing is likely to evolve and the impact on local Tier 2 and 3 suppliers. And how the Catena-X Automotive network that SAP has been influential in is adapting to the billions that Japanese, Korean and other companies have committed to invest here in the last couple of months.  

I would have loved to hear from a couple of HR executives about the changing tone on DEI. And sustainability execs as the world refocuses on oil and gas, nuclear, hydrogen and other fuels.

There were other items which made it into the keynotes but did not get enough time or hit the mark.

I was pleased to see a Neura cognitive humanoid robot in a maintenance use case but SAP should have been all over it and spent a lot more time given they are strong in many industrial verticals which need what Jensen Huang of NVIDIA calls “physical AI”.  SAP has been also talking about Industry 4.0 for years. Elon Musk says his Optimus humanoid robots are going to be Tesla’s future.  

Same thing with SAP’s new partner, Palantir. CEO Alex Karp deserved way more than the minute of time he got. He is a “big brain” and could have talked about his new book which is about the “future of the West”. He could also have talked about public sector verticals, the potential for DOGE-like capabilities SAP could deliver as every country faces budget crunches, especially NATO countries as they increase their defense spending. Finally, it would have been nice to hear from Karp about being part of the recent Trump entourage in the Middle East and winning trillion investment dollar commitments.

Of course, I realize I am myself just a “category of one”. Sapphire caters to a much wider range of audiences. And I only watched the keynotes. Lots more content is presented in the many breakouts and booths at Sapphire.

However, I will continue to push for more ambition from SAP. Its customers need it and the world needs it as it rapidly changes.

BTW – you can watch both keynotes by registering here

Burning Platform: Agentic AI – the war of words and valuations

For the 133rd episode of Burning Platform, we host Bob Evans of Cloud Wars, a highly respected, long-time observer of enterprise technology trends

We discuss comments from Satya Nadella of Microsoft that we have moved to the age of digital agents and they will cause the end of SaaS and broadly other business applications, which he says essentially only provide skin deep services front-ending databases – create, replace, update, delete (CRUD). Satya can speak with credibility at Wall Street which values Microsoft at 10X that of enterprise vendors like Salesforce and SAP. In turn, Marc Benioff of Salesforce has taken shots at Microsoft’s Copilot AI offering being too dependent on OpenAI and the Consumer Web and called it underwhelming in the Enterprise, especially compared to his own agents.  

Bob also points out that Bill McDermott of ServiceNow, without naming specific competitors has warned of the “imminent “collapse of the software-industrial complex” whose limitations pose existential threats to businesses that have poured tens or even hundreds of millions of dollars into those outdated technologies.”

We discuss that Aneel Bhusri of Workday emphasized during their recent Innovation Summit that we will need to focus a lot more on Agentic TCO and ROI.    

We discuss that architectural shifts in the enterprise tend to be additive, rather than “kill” incumbents – so we have ended up with a diverse portfolio of mainframe, client/server and cloud technologies and now will need to also add Agentic to the mix.

We then discuss the excitement around AI, the cautious uptake of agents at customers, Jensen Huang of NVIDIA already calling for the next wave of “Physical AI” and the need for agents to reflect massive global and industry changes being unleashed by the Trump administration.    

Plenty more in this wide-ranging 45 minute conversation with this veteran industry analyst.  

Sapphire 2025 – a conversation with a different breed of SAP partner, Mindsprint

In prep for Sapphire 2025 in Orlando next week, I recorded a long conversation with Amit Jain, SVP and Global Head Enterprise Application Services at Mindsprint about Digital Transformations in SAP World.

Mindsprint, headquartered in Singapore, has a strong global presence across the US, UK, Middle East, India, Australia and Africa. And as he describes with their corporate heritage as part of the Olam Group, a $40 billion food and agri-business MNC, they are deep into many unique verticals few other SAP partners can match

Here are 5 segments from the conversation – each about 3 to 5 minutes

Introductions

Amit provides more background on Mindsprint and Olam, and provides a little more color on what he defines as Digital Transformation.

“Clean Core’ and Digital Ambitions

Here we discuss how customers need not constrain their Digital Transformations too much even as they balance SAP’s guidance for a “Clean Core”

The changing Data landscape

Here we discuss SAP’s more recent announcements around Business Data Cloud and GenAI announcements and previous predictive and other AI efforts.

Exciting AI and automation advances

We discuss SAP’ s Joule agents and the broader world of Agentic AI and physical automation in the form of robotics, drones, UAVs etc

Other SAP vectors

Here Amit talks about global, multi-country SAP projects especially in emerging markets, customer adoption of RISE and Mindsprint’s work alongside various hyperscalers

I realize Sapphire is a whirlwind for most attendees, but if you are coming to Orlando for the event you should definitely try to meet Amit and other members of the Mindsprint team. And get to know them better through other channels.

Due Diligence in the AI Services Gold Rush

Much has written about the California Gold Rush in the mid-1800s. Many miners had a miserable time and made little money, whereas those who sold them jeans, shovels, meals, banking and other services did a lot better

Stephen Klein, CEO of Curiouser.ai draws a similar parallel in a LinkedIn post.

“While startups bleed cash and foundational model companies lose billions on every product, five consulting firms: Accenture, BCG, Deloitte, McKinsey, and PwC, have quietly built the fastest-growing revenue engine in the GenAI ecosystem.

In fact they are the ONLY ones making money.”

And he goes on to describe the consulting firm version of picks and shovels:

• Strategy frameworks
• Integration blueprints
• Responsible AI protocols
• Change management programs

Let’s be fair. These firms are not taking advantage of desperate miners who are a long way from home. They sell to corporations who have plenty of sharp analysts, seasoned IT and procurement folks. These firms are the tip of the spear and often bring early knowledge of what works and doesn’t in a fast emerging sector.

However, it is still hugely important to do your due diligence on how these firms themselves are applying LLMs, copilots, Agentic AI and other automation to their own operations. The reality is one of the biggest costs of migrating to the cloud (the platform for many AI solutions) is that of systems integrators like Deloitte. These firms have done literally thousands of such migrations and have countless artifacts related to data conversion, code conversion, end user training, unit testing and other phases of such conversions. Are they using LLMs to harvest that intelligence from previous projects and only need to focus on deltas unique to your project? Are they using conversion and other factories and specific tools to automate various phases? Do they have digital agents who can turbocharge their own consultant productivity and can they show metrics that they are showing on today's projects?

Do the cobbler’s children have the best fitting and looking shoes?

Most of these firms are also much more familiar with white collar, back office processes and have large ERP and CRM practices in the Western world. Clearly useful but when it comes to AI payback and ROI that only delivers to a tiny portion of modern economies. And they will likely deliver even less to the re-industrialized workforce and robotics, UAVs, sensors and other automation coming to US and other factories, logistics, farms, construction and other sectors as a result of global and vertical re-alignment coming especially with Trump's "tariff turmoil"

Stephen cautions the “success stories” we hear from these firms come from:

• Reports authored by the same consulting firms selling the services
• Surveys funded by GenAI vendors
• Case studies with no external verification

My firm, Deal Architect has helped clients run RFPs, negotiate contracts and do due diligence on service providers in a wide range of emerging markets – when they first started selling offshore services in the early 2000s, when they first started offering cloud data centers, when they first started bragging about their “SaaS” expertise. It has continually evolved market intelligence I did at Gartner helping clients evaluate SIs on their large ERP projects in the run-up to Y2K. Each wave of services called for different skills, pricing and service levels but a common thread is firms tend to “spray paint” on top of older capabilities.

I distinctly remember sitting with the CIO of a Fortune 500 at a bar of a Bangalore hotel. Jet lagged and wide awake way past midnight he looked at me “What the fxxx are we doing here? Can’t you find us a vendor close to home?” We changed course and did. Very differently,  an exec of a software vendor looking to diversify his development centers told me (also in India) “It’s such a wide world. I hope you are also planning to take me to meet vendors in E. Europe and S. America.” We did.  Infrastructure services vendors would promise cloud capabilities but take us to visit data centers from the 1990s and offer none of the scale and unit pricing that hyperscalers (they weren’t called that back then) were starting to offer. We advised several clients to stay put with their existing, fully depreciated infrastructure for a couple of more years.

The new generation of AI services are very different but require a similar sense of curiosity and  cynicism. Hopefully, without the jet lag.

TCO and ROI in an Agentic AI World

I first heard the TCO (Total Cost of Ownership) acronym when I joined Gartner in 1995. Gartner had formalized the term for IT usage in the late 80s. However, its specificity applied much more to PCs and other personal productivity items. The term as applied to enterprise tech turned out to be more elusive. Trying to estimate TCO for enterprise projects and assets which deliver value over 20+ years and which have a wide range of hardware, software, systems integration, application management, upgrades and other cost elements is not for the faint of heart. Yet, without reliable TCO numbers, the ROI calcs are by definition soft.

So, one of the nice nuggets from the Workday innovation Summit a couple of weeks ago was to hear co-founder, Aneel Bhusri take a walk down memory lane and talk about “measurable business value."

“Going back to the early days of cloud, we didn't sell cloud and say, “Hey, this is cool new technology with the new user interface. Buy it.” It was basically you can make your business more agile, better data, and you can do it at half the cost, and we grew 50% during the financial collapse of 2008-2009. Today it's no different. All technologies need to have a really, really strong TCO and ROI model.”  

“You have to go to a customer and prove that we're going to save you money. You're going to get all these capabilities, but it's also going to be cost beneficial. If you can't have that second piece, it goes away. I mean, AWS took off because it's, frankly, cheaper than running your own data centers. It's a great platform. There's 100 reasons why it's a better platform, but if it was more expensive than what's out there, it'd be, it'd be a hard one to sell. I mean that. I think that almost every piece of technology comes down to a payback, TCO or and when you say about TCO, it can be productivity, but that productivity has to be translated into cost, or you're never going to get those things through the CFO”

I would argue that one of the main reasons legacy ERP, CRM and other vendors have seen only 20 to 25% of their customer bases migrate to their newer architectures is because they do not  fully developed TCO and ROI models. Most vendors have “value realization” teams who can help a customer develop nice slideware for board presentations but they often lack rigor, especially about migration and on-going maintenance costs for even 5 years, forget 20 - especially when it comes to validating numbers their partners present. And when it comes to outcomes to anchor ROI around, the metrics are often too generic, not industry or geography specific.

We may face the same dilemma as we try to sell agentic solutions. There are new cost elements – GPUs, hyperscalers, data acquisition costs, data science talent, new testing and validation processes among them that few customers know how to measure with much specificity.

The ROI may in some ways be easier to correlate to outcomes if agents mimic current job roles. If agents allow you to scale up volumes without new human effort, if they allow you to reduce some of the tasks human employees currently perform that productivity is relatively easy to quantify. As I found in research for my book Silicon Collar a few years ago automation tends to target “dull, dirty and dangerous” tasks. You can qualify savings from the reduced human time for the tasks. There is, however, a nuance as I described in the book. Automation transforms roles. Bank tellers did not completely disappear as ATMs targeted their cash handling tasks. Their jobs became more of salesperson and customer service.

In some scenarios, the payback is from increased revenue. More from the book - UPC scanners and related self-checkout did not completely eliminate grocery clerks. CPG companies found UPC codes allowed them to come up with a lot more mass customization and SKU codes  and grocery volumes grew significantly. There was positive revenue payback for grocery chains from their investment.

There is however one big constraint when it comes to today’s enterprise software. Our definition of agent is not ambitious as that of Jensen Huang of NVIDIA. He is thinking beyond agentic to physical AI.

Actually, even Jensen is not as ambitious as Barry Roman. Barry as in the CEO of Polestar in our recent fiction thriller, The AI Analyst.  Polestar’s “agents” include robots, drones, UAVs, sensors and other “physical AI” which allows them to target firemen, shop floor workers, pilots, doctors and many of the 800 occupations the BLS tracks via its SOC methodology. And it has plenty of worker skills, attributes etc. they use to identify which automation – machine vision, robotic limbs, conveyor belts can be applied to each task.

Today’s agents that enterprise software vendors are showcasing mostly focus on white collar, back office workers. Clearly useful but when it comes to payback and ROI that only delivers to a tiny portion of modern economies. And they will deliver even less to the re-industrialized workforce coming to US and other factories, farms, construction and other sectors as a result of global and vertical re-alignment coming with Trump's "tariff turmoil"

We are still in the “soft” phase of calculating TCO and ROI numbers for AI. We are focused on “trust”, “ethics” and often on short lived, “ephemeral” agents. Aneel is right – we need to become much more quantitative in our enterprise AI pitches. But I would also like us to sign up for a stretch goal for our product development -and become much more ambitious in our definition and reach of “agents”.