This text was originally published in Brazilian Portuguese
I believe we are reaching the end of a cycle for Professional Services companies that should dictate much of the work of consultancies and software companies in the coming years. The current market situation seems challenging, not only due to the supply of innovative and higher value-added work, but also due to the prospect of using artificial intelligence to replace development work.
As we explored what the technology market would be like at Lambda3 , back in 2020 and 2021, we realized that the main challenge for the future would be sustaining a business model based on differentiation following the model proposed in the Harvard Business Review:
If you are not familiar with this model, I highly recommend reading the article by Ashish Nada and Das Narayandas , which goes into depth in describing the foundations for a professional services business, and what to do to differentiate yourself.
In a market of growth and easy access to capital , this strategy is quite efficient, and allows companies to focus on customer profiles that really want a differentiated service, even with a higher ticket.
To take advantage of the capital flow, for more than a decade companies hired more than necessary to execute projects, and our market ended up super-specializing roles: button UX, front-end Dev who only programs with react, NoSQL DBA, etc. Salary competition was also very common until 2021, as many companies had a target headcount hired.
We took a different, longer and more challenging path, which I believe worked out well, and eventually our company was acquired . It was an incredible challenge and may be the source of other blog posts.
As investment money has dried up with the big tech crisis in 2022 , also due to rising interest rates in the US, all market expectations of growth and speculation have been replaced by overwhelming pressure for short-term results. This movement can be seen in the way many of the layoffs that have been happening in recent years :
In times of technology crisis, the companies that suffer first are those that offer specialized labor, whether through development squads, labor allocation or closed-scope projects. When companies suffer budget cuts, the first to suffer are suppliers: delayed (or postponed) payments, reduction of teams and possible breaches of contract.
Gergely Orosz, the Pragmatic Engineer, has a long article describing this movement , and its impact on the tech market. I highly recommend reading it if you are interested in this trend.
Looking for a way out
In this context, professional technology services companies all found themselves in a serious bind: ensuring business profitability based on headcount and hours was no longer enough with the pressure to reduce margins. For this model of super tight margins, there is only one work reality that works: bodyshop with low margins (10% or less), which eliminates the possibility of selling squads and teams with higher ticket prices. Escaping this reality is a difficult path. It is as if the entire market, at this moment, were leveling down to meet the expected results - natural, but still problematic.
I have always been critical of the bodyshop model and the challenges that running a business focused on this brings, but there is no denying that it has an important appeal and a lot of room for growth in this market: the supply of cheap labor and technology has its benefits, despite the higher operating costs and much lower margins. That is why every company seeks to avoid this model: in the long term, it is not the most sustainable.
What happened in 2023 and 2024 was that all the big companies moved towards other service offerings: managed services, digital products or closed recurring revenue packages. Just browse the websites of several of the big consulting firms to understand the movement that has already been happening since 2023: greater emphasis on recurring ticket services, digital products or a greater offer of standardized services that generate less complexity in negotiation. These companies underwent consolidation of operations and reduction of staff, in a natural movement to reduce costs.
And more than that: all companies reduced overall investment costs in training and benefits, increasing workload: the short-term cut that will affect everyone in the coming years. Consulting firms are the first to suffer in this market reality, and the result was clear in several of our competitors and much larger companies that either consolidated or downsized, and this also ended up generating a cascade of layoffs. Thoughtworks, in fact, went back to being a private company , a clear move to focus on this market transition thinking in the long term, which is impossible to do if you are a listed company.
AI is the "perfect" solution for those who buy IT services
When you combine this scenario with the perspective of Artificial Intelligence, promoted by tool sellers, hunger and the desire to eat come together: a market expecting greater operational efficiency, and a solution looking for a problem, with a tacit killer argument for companies: the possibility, even if remote, of replacing professionals.
This intention is increasingly mature, and I have already discussed it a few times on this blog . For me, it is the real reason for so much speculative money in investing in AI solutions. Regardless of the contrary evidence we already have about the effectiveness of these tools .
The reality of teams working on projects will face a dark scenario. I believe that this is how this cycle of evolution in the Professional Services market will end. From now on, it will only be downhill. This model of operation has its days numbered - and our expectations for the type of work that will come are not so positive:
The simple perception of productivity gains generates pressure for more deliveries
Pressure for margins forces a reduction in quality: whether in terms of people or delivery
More individualized deliveries have a perception of greater team efficiency and greater execution capacity
Teams are split and reduced to reach the previous baseline, which now requires fewer people, even though the codebase is being degraded to the point of no return
More stressed teams produce worse code, accelerating the degradation of the codebase. What was bad gets even worse.
Maintenance issues and code complexity further reduce the delivery capacity of teams and will be the next hiring driver, who will need to rewrite products to compensate for the state of previous systems.
To truly understand this emerging future, I decided to take a deeper step into the technical world of artificial intelligence, through a master's degree to prepare myself for the next cycle that will impact our market, not because the solutions are only transformative, but because the interest in using these tools is being sold to us as a technological panacea like has rarely happened in the world.
I believe that AI is being pushed into every industry worldwide as a way to offset the tens of billions of dollars invested in infrastructure and energy , to the point where BigTechs are putting the planet at risk to fund the amount of energy needed to run increasingly complex machine learning models.
For us technologists to be prepared for what is to come, the only way is to be a source of expert and skeptical knowledge regarding the promises of Artificial Intelligence, to support decision-making, not only at the business level, about the use and regulation of these tools.
Software Delivery Services - the future?
There is a lot of speculation about where the IT Professional Services market is headed next, and it is clear that the shovel-building machine is acting to create even more excitement about AI.
A new, transformational concept is the idea of using AI agents to perform atomized (micro) and relatively independent work that would be contracted out instead of a freelancer or consultant - yes, more dream than reality. In this idea, there is a shift towards usage-based business models in Software Delivered Services (SES) companies, or Software Delivery Services , particularly as it intersects with the traditional professional services sector. By leveraging technology to automate labor-intensive tasks, some companies are offering cost-effective alternatives to established professional services firms while still meeting client needs.
In this concept, teams using AI agents can fill gaps in an automated way, for jobs that are small enough that the entire sales management, onboarding, and delivery process is completely different. In essence, successful SDS companies will embody the “full-stack startup” concept introduced by Chris Dixon .
Examples of such companies include Sierra, Lighthouse, Crescendo, Micro1, and Exec.com. Each of these companies offers a unique service that leverages AI to meet specific industry needs, charging based on unit or performance.
The success of this model will depend on the ability of these companies to find work that can be automated as much as possible, breaking the traditional project cycle in Professional Services.
I remain skeptical, but I'm interested in learning about some of these trends.
How AI can transform the market
The Internet of Bugs channel is perhaps the most interesting on the topic of software development today. Carl Brown is experienced and skeptical about all the hype, and has been a reference that I like to follow. No memes, no click bait, no nonsense to make the video longer: pure content.
Does software development tend towards mediocrity?
On Twitter, amid the turmoil of last year, I described my frustration with the technology market. Here I share some of my experience in 20 years of Professional Services.
Does all software development tend toward mediocrity? Yes, and there is very little we can do about it - we have lost the fight for a technology ecosystem that values software quality, delivery consistency, and sustainable pace. More than that, there are no adequate mechanisms in organizations that can achieve this in a lasting way.
Despite the commendable examples of those who are trying hard to maintain a sustainable pace and some kind of mastery in their day-to-day work, the main part of our software industry is average, mediocre, incompetent and has no intention of improving: it is easier to change companies, or give up on launching a product, than to prepare for a real evolution that will bear fruit in the long term: this is the TikTok industry of IT. No one stays long enough in the company to understand the impact of the bad decisions we are making: not executives, not employees, not entrepreneurs. All that is left is the vicious cycle of picking up the pieces of increasingly fragmented technologies until we can no longer take it, with professionals who are increasingly stressed and less experienced to meet the ever-growing demand.
It only takes a few months for an entire technology team to be replaced and for all decision history to be lost - We've gone from a world where people retired at work to one where no one lasts two years at a company . And general incentives only make the situation worse: Executives measured by quarterly results make absurd decisions to ensure current numbers, even if it generates future losses. Salary rules prevent employees from being evaluated above a certain percentage per year, creating a scenario where it's better to leave the company for a raise than to wait for a promotion.
The very pace at which teams are formed and disbanded works against the development of sustainable products - in technology and in business. Responding quickly to change is not the same thing as delivering anything at any cost, no matter how much your manager makes it seem that way in your day-to-day work. With managers increasingly distant from technology and at the same time increasingly dependent on the results of what is done with technology, there is no way of knowing the obvious: almost every company is an IT time bomb.
I really think the future of the profession is quite bleak: increasingly pressured by short-term visions and the search for immediate results, to the total detriment of the survival of companies' technology parks. Nowadays, it is easier to declare bankruptcy and move on to the next greenfield startup . Even with the advancement of the discussion on code maintainability, quality metrics and process efficiency in various fields, the set of incentives for executives who touch technology is so perverse that there is no chance, under any circumstances, that a motivated team will remain that way for long.
In this context, artificial intelligence comes to make the situation worse. As Giovanni Bassi said: managers and executives who don't understand much about technology now expect the possibility of using a tool whose greatest potential is to replace the developer on the team, while knowingly adding more complexity and technical debt to the code, further aggravating the problem.
Technology vendors, in turn, overestimate the capabilities of their tools, in order to differentiate themselves in the already polarized and hype-driven market: 2021 web3, 2022 data science, 2023 gpt, 2024 AI for everything.
I have been tired, for quite some time now, of this model that I fought so hard against. I don't want to sound nihilistic in this text, but I often wonder if there is a way out.
We continue...
Comments