Labor Substitution (and other worries)

Recently there has been a drumbeat of articles both by mainstream media and by individuals proclaiming the “death” of consulting. “The Consulting Crash is Coming” writes Joe Nocera. But there are also the more moderate voices like Joulie Gindi (MSc) who writes “AI will change consulting. But it will not “Kill” or replace it.” I am more in Joulie’s camp than in Joe’s but I do think that massive changes are on the horizon - and not just for consulting but really these lessons should be applied to every business in every industry.

Consulting is exposed first because its core inputs (analysis, slides, playbooks) are easiest to automate which will force the industry toward evaluating innovations such as outcome based pricing, an increase in software as a component of both the means and delivered results, and continuous or “always‑on” as opposed to episodic project services. I’ve been inspired by Phil Fersht and David Cushman in the writing that they (and other analysts) at HFS Research have done on both the software–services blur and the idea of the “human premium.” I highly recommend reading their reports and not just relying on my interpretation below. But hopefully this brief analysis will inspire you to go deeper. I believe that every leadership team should assume their own value chain will follow the same arc that consulting is following right now and should be preparing for their own market reconfiguration.

I’ve outlined 7 themes that I think apply to every business, although I primarily describe them in the context of software and services. Read them as building on each other as opposed to being discreet, the trends and implications are intertwined. At the end I summarize what to expect for the consulting and software industries but again I think you should consider the implications for your industry as well.

Outcomes > inputs: The consulting and software industries have always been tasked to justify themselves based on what is accomplished through their services/products. But now fees and licenses are shifting toward contracts that pay for outcomes, e.g. measurable business results. While clients/customers have tried to evaluate the “return on investment” (ROI) of software and services spend, mechanisms for agreeing on measurement often eludes the procurement and contracting process. Now vendors are confronting the challenge that AI can do a task in minutes which a services professional used to do in hours, and so it is no longer hours by a particular level of employee (partner, manager consultant, etc) that can be held up as a measure of value. Similarly for a software company that is selling per seat licenses where the employee count is plummeting even as the volume of activity is increasing, it is no longer seat counts which can measure value. So it is becoming better for both the vendor and customer to assess value through actual business outcomes. Yes it will still be hard, but motivations are aligning and having more automated systems in place is improving visibility and thus measurability of these outcomes.
IP over labor: Perhaps scariest for those of us who have reached a senior level in our careers is that the non‑linear scale of technology is going to replace staffing pyramids. Our business mindset since the early 1920s has been the “multidivisional form” or M-form organization, originally pioneered by DuPont and General Motors. And the evolutionary precursor of the pyramid management structure first introduced by the Erie Railroad in the mid-1800s. Managing the scale of human participation and the geographic distribution of larger and larger industrial companies required centralized decision making, functional specialization, and hierarchical management layers. And business models evolved around these organizational limitations. A typical consulting firm for example can pay high wages to the top professionals on the economics of a large number of junior staff billing at multiples of their compensation. Throughout industries this logic prevails in organizational structure, operating models, and business models and this will all be challenged as technology replaces human activities
Productization of services & servitization of software: HFS has a pithy way of saying this, “The software-service blur has finally triggered a genuine orbital intersection.” The “planets” of software companies and services companies crashing into one another makes for a great visceral feeling of doom but what might be more realistic as we think about the implication of this theme across all industries is to see it as complementarity as opposed to collision. Consulting is packaging repeatable IP into platforms while software companies are embedding “execution” (automation, agents, prebuilt expertise). The reinvention of both services and software is a product of increasing IP over labor and will support the renegotiation of payment for outcomes between suppliers and customers.
Smaller teams, higher leverage: A corollary to these first three themes is that AI and automation will displace junior work, resulting in smaller, autonomous, and more effective teams. In addition to the business model challenges of migrating to outcome based pricing and realigning organizational structure, this will also require us to rethink how proficiency in disciplines is gained. We will have to develop new models for learning and development to move junior talent into increasingly productive roles.
Talent model convergence: The definitions of consultant roles vs developer roles will increasingly overlap as the artifacts of consulting become more functional (working software instead of powerpoint) and the iterative work of developers becomes more consultative (rapid and continuous iteration based on user feedback). Clear definition will certainly exist at either end of the scale but the middle will become blurry. Already product managers can build prototypes and every employee can automate simple tasks. Tools will become more powerful (Lovable is an interesting example) and individuals who can imagine better software to get their work done will be able to make that better software with little effort in becoming proficient at software development skills, at least at the smallest scale of satisfying their own individual user requirements.
From projects to “always‑on” operating outcomes: This will in turn spur a much more complex set of requirements for the most sophisticated software systems, which will need to be continuous, telemetry‑rich services that update themselves and prove value in production environments. Software development experts need not worry that their talents will become unnecessary but should worry that the requirements for their talents will become much more complex. The systems which provide these always-on operating outcomes will also displace episodic services work which used to be needed to sift through and analyze historical data to make improvement recommendations, another example of where services and software complement or displace each other.
Speed-to-value and modularity as the new standard: “Months to impact” is now a losing pitch. In fact increasingly a winning proposal comes from the vendor who walks into the pitch meeting with working software, not slide ware. Companies will increasingly need to adopt a “minimum viable product” (MVP) mindset and commit to iteratively learning and improving through the adoption cycle instead of having a “first time right” (FTR) bias. Software systems and organizational architectures which promote composition will win over monoliths in an accelerating business cadence because they can provide islands of surety and stability for certain functions which then support extensions into experimental domains for other functions.

So what does this all mean for services and software?

Short term – Ecosystem orchestration will beat single‑vendor plays: A different way to say this is that customers will prefer purchasing an integrated outcome. So in the near term there will be a lot more “frenemy” co‑selling even as categories blur in every industry. For consulting this will mean strategy + platform + hyperscaler + implementation. Each company will however be in an uneasy relationship with the others, producing an increasingly unstable configuration. Over time each company will have to evaluate the advantage of their ecosystem of vendors and the likelihood that each will remain focused on their contribution vs. the challenge of an emerging category of companies which integrate components into a single offering.

Long term – AI‑native platforms as category shapers: The challenger to ecosystems will be in platforms that fuse software and embedded expertise to deliver outcomes quickly thereby reducing the need for separate vendors and projects. In this future state, a “consulting firm” that just sells advice without proprietary technology will struggle, as will a “software firm” that just sells code without ensuring adoption and impact. The models will converge into firms that do both with reduced friction and increased accountability for achieving business outcomes.

The Result: Massive market reconfigurations as industries collide.