I recently attended the annual HR Technology Conference in Las Vegas. There were 400-500 HR vendors in attendance pitching their wares in the exposition hall and countless more roaming the halls looking to buttonhole anyone they could. I spent most of my time in the Press Room in 30-minute meetings with various HR software vendors.
I did 21+ such meetings over three days. In statistical terms, that isn’t statistically significant, but it does provide an adequate cross-section of HR vendors, their plans, new product announcements and their vision for the market.
This year’s stories (plus what I noted in the exposition hall interactions) really left me wanting. I didn’t walk away with one breakthrough, a-ha moment. Here’s what I encountered instead.
Have you heard the good word of AI, Brother Brian?
AI almost hit a 100% pitch rate amongst vendors. It was at the level of a religious fervor. AI was all over vendor signage, presentations, etc. but I expected this. I’ve been hearing about machine learning, agentic, algorithmic and generative AI unceasingly for 2 ½ years. You couldn’t swing a dead cat at this show without hitting someone yammering on about AI.
Okay, so the AI gabfest was overly hyperbolic, so what!?
The so what is that the conversations were almost all about the current/new/impending vendor use of a new technology, agentic AI, and rarely about how the technology will be used to create (outsized or any) value for customers. It would be like listening to a sales spiel for new cars and all the sales people want to talk about are the wrenches they intend to use to tighten up the bolts/fasteners on the vehicle. In and of itself, these ‘announcements’ are not newsworthy. They are astonishingly vague and non-specific. You can’t even connect many of these to a specific value proposition. These AI announcements are the no-calorie, diet cola versions of software innovations.
The disappointment
The outsized hype regarding AI creates an equal expectation that vendors would provide new solutions with outsized vision, capabilities and innovation. Well, that didn’t happen either. In fact, it was pitiful how little innovation actually appeared.
One vendor touted how they used AI to integrate two best of breed HR applications together. I’m sorry but that’s not some amazing AI story or even a great value creation tidbit. No, it’s an ‘integration’ that had a bit of help from an AI tool. But, this same ‘solution’ could have been done years/decades earlier with any number of integration technologies available on the market.
At the macro level, these AI-powered ‘innovations’ were not transformative, novel, new or capable of creating much value. Instead, they were incremental improvements to broken integrations, disjointed processes, out-of-date applications and obsolete or irrelevant workflows. They were derivative extensions of existing application software. They were the tasseled handgrips on a little kid’s bicycle – nice to look at but adding no value. AI, as being delivered today, is not a technology that delivers breakthroughs but a technology that spiffs up older application software. AI is merely window dressing.
To established software vendors, AI is a tool to be used judiciously, in baby steps, and this is most important, as a tool of incremental improvements. It’s stunning to see vendors preach ad nauseum about the BIG “T” transformative powers of AI only to see them deliver a job description generator or some other less than noteworthy/low value creating tool.
The agentic opportunity
AI agents are like pieces of a large process. Using a travel example, one agent could help book your flight, another books a car service, a third books the hotel, a fourth loads all of the particulars into your office productivity software calendar. A final agent can then trigger all of the appropriate reminders for you. Additional agents could then monitor the flights as to their on-time status and autonomously rebook as needed.
If those are all agents, agentic AI is this clever, programmatic glue that can connect all of these agent pieces into a viable workflow. That glue is ‘smart’ too as it can complete different tasks based on changing needs, new data, different users, etc. This glue is the ‘orchestration’ within agentic AI and it’s what makes agentic AI different from older process automation tools. Another characteristic of these agentic workflows is that the software learns and recommends based on its ever-growing data store/knowledge base.
Today, customers (and vendors) might be tempted to just connect different process components together and call that a smart workflow. That’s what a marketer might call it but it needs the orchestration logic and power to really be an AI enabled activity.
The best vendors will use these new capabilities wisely. Instead of re-paving old cowpaths, they need to really look at current workflows, controls, process steps and data to see if these could be improved or totally reinvented. Using agentic AI in a basic process automation use case is expensive and wasteful. Using agentic AI to dramatically improve operations is not.
Finally, the feel and operation of agentic AI powered processes is different from old application workflows. Menus are mostly gone. Data is served up to a user at the point of need. The upgraded software anticipates next user or process actions, reviews relevant data or recommendations and lets the AI tools complete the business event in the background. If you or a vendor are still using menu-driven applications, keying in a lot of data, etc., you are so old school!
The biggest failure
In between the conversations at the HR Tech show about how great AI is, more than half of the vendors I met with acknowledged how broken Talent Acquisition (aka Recruiting) is today. And, what’s more interesting is that citizen-AI tools are the proximate causal factor for many of these break points.
Specifically, talent acquisition is struggling because:
- Too many resumes/applications are being submitted to employers simply because citizen-AI tools make it easy for jobseekers to mass apply to hundreds or thousands of positions each week. (Volume Problem)
- Too many resumes/applications have been ‘perfected’ by citizen-AI tools. Jobseekers use generative-AI tools to do more than polish their resumes. A fast-food cashier is now a “Treasury Transactions Expert with Extensive Front-Line Experience and Numerical Precision Skills”. Great jobseekers without ‘perfected’ applications will never get ranked well by easily fooled Applicant Tracking Systems (ATS). (Real Skills vs. Puffed Up Experience)
- Some malcontents are using the credentials of legitimate people to gain employment, often for remote work. (Fraud)
- The recruiting process suffers from a garbage in, garbage out (GIGO) problem. Vendors don’t know how to effectively eject irrelevant (e.g., a person from Belarus applying for an in-office role in Akron, Ohio), fraudulent, inauthentic, or otherwise inappropriate applicants from the process. (Dirty Data)
- Recruiters overly rely on ATS technology to match keywords and are highly vulnerable to applicants that use citizen-AI tools. Those AI tools make an applicant’s mediocre work experience and professed skills score well and magically align with the job description. (Misdirection, Erroneous Candidate Ranking)
- Etc.
Yes, HR software vendors are absolutely aware of the myriad problems in Talent Acquisition but the solutions they have developed are immature, one-dimensional, incomplete and amateurish. Worse, none of them offered up a compelling long-term vision that would actually help HR teams actually win the war for talent – a war they are losing badly right now. These vendors are bringing pocket knives to a gunfight.
Another related talent acquisition problem that vendors admit is happening is the destruction of the value derived from job board placements. HR departments are spending small fortunes on placing job postings on these boards and getting inundated with irrelevant, fraudulent and otherwise useless applicants. These firms are paying more than ever for worsening results. No vendor I spoke with had any suggestion to this business problem.
Different vendors had some ideas as to how to deal with some of this dysfunctionality but no one was comprehensive. Specifically, no one was interested in stepping back to review the entirety of Talent Acquisition functionality in a fresh new way. Instead, all I heard were of small point fixes to aspects of the Talent Acquisition problems. Other fixes were often couched in aspirational what-if statements like “I suppose we could look at the IP address of the applicant and see if it is even close to where the job is located” or “Maybe we could integrate a credit or payroll database into our talent acquisition process to see if the applicant is already employed at 1-2 other firms in faraway locales before moving them into the Talent Acquisition workflow”. These conversations were full of half-hearted partial solution ideas that would be bolted onto an already bad, dysfunctional process.
I heard vendors say:
- “We’ve actually thought about that”
- “That’s actually a good idea!”
- “I think we could actually do that”
- And more
For the vendors reading this, here’s a million-dollar clue for you at no charge. When you think about radically reimagining Talent Acquisition, look at the whole process not just the bits you already deal with. Most vendors and at least one industry analyst seem to think that Talent Acquisition begins once applications or resumes end up in the ATS. WRONG! The majority of problems that businesses need help with originate in all of the data problems, candidate acquisition, and other steps before or during the ATS processing steps. You will have no impact by reimagining the parts of the process that aren’t broken or are only slightly off kilter. Expand the scope and massively improve the outcomes for software buyers.
Talent Acquisition, the process and technology, is broken. The needed changes must reflect a new view/vision of talent acquisition that actually aligns with the modern AI and citizen-AI empowered world. Bolting things together or making a few small changes to existing workflows and processes is a non-starter. It’s time for a Big “T” transformation of Talent Acquisition.
Why are vendors stuck on incremental changes and innovation?
I have several theories as to why this is happening. Herewith are my best guesses:
- The history of the software company has been built over decades of incremental changes. They just don’t do bold, transformational change. Incremental change is what is in their DNA. It’s the only way they know how to innovate (i.e., strictly at the margins and in small, digestible pieces).
- Their knowledge of AI is not that great after all. Hey, three years ago, they hadn’t even seen an AI application.
- Delivering radically rethought and reinvented processes via Agentic-AI that are also error proof, highly accurate and repeatable is not in their wheelhouse. It’s new and risky work. If they build something, they want to take baby-steps to minimize risk to their firm and their customers.
- They need a customer or two to tell them what to build. Their core competency is taking customer orders, not anticipating customer needs, or, not building cutting edge solutions well in advance of customer requests. Real, transformative solution ideation is not their core competency.
- Doing original/primary research is something they don’t do or they do it poorly. They think fast copying the competition is akin to radical reimagination. It isn’t. (Please note that some speakers and vendors also spoke of how AI is redefining work. But few could point to actual designs that have been rolled out or what specifically these entail.)
- They have already invested a lot in the old ways of doing something and won’t change unless absolutely forced to. Vendors have historically relied on their existing customers to take an eternity to upgrade to new software versions, turn on new capabilities, etc. They are counting on customers to drag their feet. As a result, speed has not been a priority for many of the larger software firms.
- They don’t really deliver competitive advantage software capabilities to customers as competitive parity will suffice. They’ve already concluded if they don’t need to completely reimagine key business processes (e.g., procure to pay), as long as the new solution is similar to another vendor’s technology. They think customers will be happy with look-a-like, work-a-like products.
Let me offer this proof point as to the bankruptcy of their innovation, ideation and research capabilities:
- Only one of the more than one dozen vendors who spoke of the considerable Talent Acquisition problems out there could produce a slide that mentioned a couple of the challenges. They didn’t have any stellar solutions and nothing that could be called reimagined or reinvented. They were suggesting a point solution for the existing Talent Acquisitions tools out there. It’s a start but it was far from complete.
- Several other vendors could talk to one of the talent acquisition challenges but their ideas were limited in vision, completeness, etc.
- NO vendor offered up a comprehensive and viable solution. Further to this point, although some executives spoke of the need to move some new capabilities BEFORE the ATS, no one could produce anything close to a solutions roadmap and what it would do to the existing Talent Acquisition workflows. If there is no roadmap, there is no solution (or vision).
Vendors are still building a home brick by brick when the market now needs a recreational vehicle.
Don’t forget the apologists
We shouldn’t overlook the software apologists out there. These are the vendors who try to get you to believe that while they believe in these Big ‘T’ transformational technologies like AI, they just know their customers can’t handle this much radical change in short order. No, these vendors act as an on-ramp that gently permits customers to ease into the world of AI at a more measured pace.
Existing software customers, while they may be accustomed to such positioning from their ERP and other application vendors, should be offended/insulted by such stories. These vendors are basically saying that customers can’t handle the new technologies and the process, business and other changes these can produce.
Vendors spin these yarns as:
- Slow, small upgrades make it harder for a competitor to get a foothold into their account. If your firm is implementing a few AI point solutions instead of radically reworking an entire process (e.g., order to fulfillment), then you will likely wait and use little AI bolt-ons from your existing vendor.
- Slowly replacing parts of a process with AI capabilities means the vendor can charge for both the old functionality and new AI capabilities simultaneously. If the old process gets replaced, the vendor is only left with one income stream instead of two.
- Significant numbers of their customers are slow to adopt new technologies. Laggards and Late Majority software users/buyers will drag their feet on the purchase and deployment of AI-powered technology. They’ll want it bulletproofed and cheap. Timeliness is rarely one of their top buyer values.
- Partner firms (e.g., integrators) might love to talk a big transformation game but the bulk of their staff simply configure tables, cleanse data and integrate systems. Partner firms know they can’t really do Big ‘T’ transformation and/or can’t do it at scale. They’ll need time to develop new IP, skills, etc. to really roll out radically new ways of work.
My take
When I flew out of Vegas Thursday night, I was unsettled. While I wasn’t thrilled that my flight was delayed, I was really bothered by the things I heard in the briefings and what I saw on the show floor. On the flight home, I used old-school pen and paper and wrote out the draft to this article.
Once back at my office, I read Allyn Bailey’s great LinkedIn piece re: the HR Technology Conference and the products featured there (note: Allyn’s with SAP SmartRecruiters). Apparently, I wasn’t the only one who left there troubled. I very highly recommend you check out her short, packed article. She noted:
I’m coming out of hashtag#HRTechConf with one big realization, we are in trouble, my friends. I could write a hype post about the shiny things, but what’s really on my mind is the gap I saw in us as analysts, buyers, and users.
Bailey detailed 10 rails that any new solution will need today. You’ll need to read Allyn’s piece for those. But here’s a taste of what else got noted:
If your mental model is pre-AI, you’re thinking in legacy code. The rails have been rebuilt. Data flows, system connections, and decision-making inside the machine are fundamentally different. Keep buying and building like it’s 2019 SaaS, and you’re laying train tracks for a world that now flies.
She’s right.
Last week’s experiences were a wakeup call although I doubt the show producers intended that result. While I had been hearing so many vendors for so long tout the power that AI was going to deliver, the lack of progress and imagination was quite disappointing. I saw no before/after process or workflow documentation. I saw no original research. I saw no imagination. There are big problems to be solved but vendors weren’t showing anything. And, it didn’t sound like they had even anything remotely transformational on tap.
HR solution vendors aren’t the only software companies that need to do better in their AI efforts. Financial accounting vendors could do more to advance planning, forecasting and better accounting staff utilization given the upcoming shortfalls expected in CPA holders.
If you’re a software vendor who is still charging pre-AI prices and delivering pre-AI era solutions with no real innovation, then do us all a favor. Cut your prices, admit that you don’t know how to create breakthrough processes, and, turn your firm into a lifestyle operation for the founders. You’ve already lost the relevancy game. It’s time to milk that old cash cow for all she’s worth.
If you think your products can command a premium price due to their AI capabilities, then prepare to prove it. Show us your great research, the before/after workflows, your before and after process workflow metrics, the value proposition your solution generates, your easy-to-understand AI pricing, your security/controls and more. But also show us how long this solution will last before bad actors use citizen AI tools to circumvent your solution. Can your firm innovate and develop appropriate countermeasures faster than the speed of innovation?
I have my doubts.
link