Where This Is Going¶
SOLTECH in 2028. Written as if it has already happened.
The thing we used to sell became abundant¶
Since 1998, SOLTECH has sold something genuinely scarce: the ability to take a business problem from a non-technical executive and turn it into working software.
Software engineering was a specialty skill. Engineers were hard to find and expensive to assemble into a team. And like most engineering disciplines, success depended less on individual brilliance than on a well-managed process and the kind of judgment that only comes from having done the work many times before. Technology kept evolving, and it was never easily understood by anyone outside it. That knowledge gap is what created the need for specially educated engineers who had to do the work themselves. The gap was the product.
That gap is what AI is chipping away at. Not the ability to describe what you want, which was never the hard part. AI removed the barrier between describing something and realizing it as working software, and realizing it was precisely the thing only experienced engineers could do.
The mechanical act of producing software has collapsed in cost and will keep collapsing. Our clients know it. They arrive having already built something themselves, or having priced the work with an AI-assisted shop offshore, or simply expecting the number to be a fraction of what it was. This is not a negotiating posture. They are right.
A firm whose product is hours of software construction is selling a commodity that gets cheaper every quarter. That firm has a future, and the future is smaller every year.
What did not become abundant¶
Technical leadership. This is the one that matters most, and it is the one AI comes nowhere near. No combination of a capable business person, a vibe-coding tool and a frontier model provides the leadership, the safety net, or the true technical ownership that comes from having a senior and authoritative person behind the wheel of a technology endeavor. Someone who has seen this go wrong before, who knows which decisions are expensive to reverse, who can say no, and who is accountable for the outcome rather than for a deliverable. That role has never been automatable and it is not becoming so.
Complexity that will not yield to a prompt. AI plus modern hosting will carry a determined person a long way, further every month. It will not carry a product owner through a genuinely complex enterprise environment: systems that already exist, real data, integration, compliance, scale, security, and the constraints of a business that is already running. Those problems still require technologists. A subscription to an AI tool is not a substitute for someone who has solved them before.
Making it real. There is now a large and growing population of organizations holding software that partly works. It demos. It does not survive contact with real users, real load, real regulation, or the person who has to run it at 2am. The distance between "it works" and "it runs the business" has never been wider, and almost nobody sells across it.
The honest limit¶
Technical leadership is exactly why a client chose us over an offshore vendor in the past, and it is still why they call. But the premium it commands is concentrated, not continuous.
While uncertainty is high, leadership is worth almost any price. Once the blueprint is in place and the unknowns have been resolved into a clear implementation plan, most of that value has been delivered, and the pull toward the lowest-cost way to execute the plan becomes very strong. Offshore vendors were already commoditizing that second phase on larger and longer projects. AI is doing it faster, more cheaply and more successfully.
The anchor moved¶
The scarce part of the work sits at the front of an engagement. The revenue did not.
Implementation projects were the anchor. A larger share of our revenue came from them than from anything else. They were the big, long engagements that carried the fixed cost of the business, and growth meant winning more of them. Everything we built was built to make that happen: the size and shape of the engineering organization, how architects spend their time, what marketing chases, how we sell, and what we measure.
When your largest source of revenue shrinks suddenly and substantially, you do not get to re-examine one part of that structure at a time. The whole foundation comes into question at once, because all of it was sized against the same assumption.
So the work in front of us is not adjusting the implementation business. It is deciding what our new anchor is, what the company organizes itself around now, and then reorganizing to match.
Two consequences follow immediately, and both are already visible.
A large pool of software engineers becomes a liability rather than an asset if we cannot create a meaningful pipeline of work to keep them busy. That is not hypothetical. It is why we have been covering with contractors rather than hiring, and why we have not backfilled departures.
And if more of the real delivery work now sits in the architect channel, architects become the constraint on growth. They were never sized for that. They were sized as a short ingest process at the front of the funnel: convert opportunities, set direction, and keep the implementation engine fed. If they are becoming the implementation engine, that is a different job at a different scale, and it requires significant realignment rather than incremental hiring.
Who this makes us right for¶
The collapse in build cost does not affect every buyer the same way, and being clear about who we are for matters more now than it did when everyone needed us.
Startups and entrepreneurs are becoming less viable as clients. They can build their own interfaces and iterate toward a complete vision without us. The ones willing to put in some deeper learning can vibe-code their way to a mostly functional application. They typically do not fully understand what they have built, and they come to us for last-leg delivery. That is real work, but it is low-revenue work for buyers who are by nature scrappy and want to pay as little as possible. It is not a business we can build on.
There is a variant worth naming: the entrepreneur who is not technical, is reasonably funded, and has no interest in the sweat equity. They want a partner who knows how to use AI and passes the savings on. That is a legitimate buyer, but the price expectation is set before the first conversation.
The need for last-leg delivery and application finalization is genuinely increasing, and it will keep increasing. The problem is the mix. Too much of that demand comes from people who are excited about an idea, have no real funding, and have no concept of what rolling out and running a system actually involves. The same wave produces a far better opportunity one step downstream, which is covered below.
Mid-sized companies remain our market. They have busy executives who do not have the time to become part-time coders. They have real problems that matter to the business. They expect to pay less for software than they did three years ago, and they are right to. And crucially, they have no internal IT team to lean on to figure it out for them. That combination, a real problem, no internal capacity, and an executive whose time is worth far more than the learning curve, is exactly the buyer an architect-led firm serves best.
Enterprise-grade complexity does not only live in enterprises. A mid-sized company with twenty years of accumulated systems, real compliance exposure and no architect of its own has exactly the kind of problem AI does not solve on its own. That is where we are worth what we charge.
The firm we are becoming¶
We sell answers and outcomes, not effort. Clients buy a decision they can act on, or a result that works, at a price we quote up front. Feature Billing was the first step. It broke the link between our revenue and our hours, which is the link that was going to kill us. What follows is pricing that is not derived from a labor estimate at all.
Architects are the delivery unit, not the front end of one. A senior person owns the client relationship, the problem and the outcome, from the first conversation to the last. They are amplified by AI and supported by specialists pulled in where real depth is required. They are not supervising a handoff. They are doing the work at a level of leverage that was impossible three years ago.
We win on speed and volume, not on size. The old model needed a handful of large projects a year. The new one runs many fast engagements, each delivering something real in weeks. That changes what a good quarter looks like, what we measure, and what we optimize for. Responsiveness becomes a weapon: when a client asks a serious question, we answer it before anyone else has scheduled the discovery call.
We are smaller and more senior, and that is the point. The classic consulting pyramid ran a few seniors over many juniors and made its margin on the markup. AI removed the base of that pyramid, everywhere, for everyone. Rather than defending a structure the market no longer pays for, we build the firm that structure was hiding: senior-heavy, heavily leveraged, with fewer people producing more value each. Fewer people is not the strategy. It is the consequence of the strategy working.
We are the firm that runs it, not only the firm that builds it. Almost anyone can now produce an application. Far fewer can operate one. Keeping a live system healthy, secure and current, and being ready when it breaks at an inconvenient hour, is a different discipline from writing it, and it is not a discipline AI makes easy. Today our support practice is as-needed and sold as an afterthought to development work. It becomes a front-line offering: recurring, contracted, and embedded in the client's everyday infrastructure. That matters well beyond the revenue, because everyday infrastructure is increasingly where the AI work actually is, not in custom development projects. And we come at it from the opposite direction to a traditional managed service provider. They grew up keeping servers running and learned about software later. We grew up building software and now applying AI. That makes us a different breed of MSP, and a mid-sized company with no IT team of its own is exactly who needs that breed.
We staff for a steady flow, not for a pipeline of large projects. Average project team size falls. The base load becomes a steady stream of small engagements, and our long-term resourcing aligns to that flow rather than to the handful of large wins we hope for. We still pursue the large projects, and we still compete for them against low-cost vendors. But when we win one it arrives as a sharp spike in demand rather than a predictable ramp. A firm built only around a fixed in-house team cannot absorb that shape. A firm built around a managed resource network can.
We sell delivery options, not a delivery team. For most of our history the product was our in-house team of US developers. That worked for zero-to-app implementation and almost nowhere else. Very few clients ever stayed at onshore rates for long in ongoing product development or support. They stayed while we got them out of trouble and into a stable place, and then moved to something they could sustain: an internal team of their own, or an offshore or nearshore vendor with operating costs they could live with year after year. We treated that as losing a client. It was the client behaving rationally against a menu with one item on it. Going forward we carry a broader resource pool and the management structure to run it, and we sell implementation at more than one cost point. Judgment, architecture, accountability and the client relationship stay senior and onshore. Implementation capacity flexes.
We are the vendor clients keep, not the vendor they graduate from. With demand generation in flux, the cheapest revenue available to us is revenue we already have. Every client we hand off to a cheaper vendor at the moment they stabilize is a lead we then have to go replace. Building a network of clients we can sell to continuously, across advisory, development, support and infrastructure, is worth more than any improvement we could make to the top of the funnel.
Being US-based stops being a handicap, because we stop selling it as the product. For fifteen years our biggest competitive problem was that someone offshore would quote half our number for the same scope. Two things change that. When the labor content of building software falls far enough, the offshore cost advantage compresses with it. More importantly, we stop putting onshore labor at the center of the offer. What stays onshore is what clients have always actually paid a premium for: senior judgment, proximity, accountability, and someone who understands the business. What flexes is implementation capacity, priced honestly against what that work is now worth.
What does not change¶
We are the firm that actually ships. We put senior people on real problems and we stand behind the result. Clients stay with us because of the individuals they work with and because we tell them the truth. None of that is up for reinvention. Everything on this site is about protecting it and building an economic engine around it that works in the market we now live in.
What we have to decide¶
The picture above is not a plan. It is a direction, and there are many ways to move in it, some of which are cheap and reversible and some of which are neither.
The rest of this site breaks the direction into specific Strategic Objectives, organized by area of the business. Some can be started next month at almost no cost. A few are real commitments. The most useful thing this leadership team can do is tell the difference, start the cheap ones now, and spend our collective judgment on the handful that actually require it.