Disclosure: City Profit sells startup models and inbound subscriptions to IT companies and other businesses, so we have an interest. The forecast section is our opinion, not a fact: it rests on the public data below and on patterns from more than 200 projects. Every number carries its source and date at the end of the article.
Developers who use or plan to use AI toolsStack Overflow survey, 2025
Developers who trust the accuracy of AI tools (46% distrust it)Stack Overflow survey, 2025
B2B software buyers who start research in an AI chatbotG2, March 2026
Crunchbase counted $425 billion of venture funding in 2025, up 30% on 2024. About $212 billion went to AI, close to half of all venture dollars, up from 34% in 2024. Five companies raised more than $5 billion each and took roughly 20% of the total. For an IT company outside the AI headline this means two things: investors ask harder questions about moat and unit economics, and a large share of capital competes for the same few themes.
AI share of global venture funding
Crunchbase, full-year 2025 data. Bar scale: 0 to 60%.
The Stack Overflow 2025 survey found that 84% of developers use or plan to use AI tools, up from 76%, yet more of them distrust the accuracy of the output (46%) than trust it (33%). About two thirds name solutions that are almost right, but not quite, as their main frustration. A randomized METR trial with 16 experienced open-source developers and 246 tasks found that tasks took 19% longer with AI tools, while the developers believed they had been about 20% faster. Veracode separately found security flaws in 45% of AI-generated coding tasks.
The perception gap in the METR trial
Early-2025 tools, experienced developers, mature open-source projects. Results may differ in other settings. Bar scale: 0 to 25%.
The lesson for a IT company: output grew, and the bottleneck moved to review, testing, security and accountability. IT companies that measure those stages ship better software than IT companies that count generated lines.
Gartner predicted in June 2025 that over 40% of agentic AI projects will be canceled by the end of 2027, citing cost, unclear business value and weak risk controls, and warned about agent washing, where vendors relabel chatbots and automation scripts as agents. The same forecast expects at least 15% of day-to-day work decisions to be made autonomously by 2028. Both can be true: a real shift with a high failure rate on the way.
Pew Research found that people clicked a regular result in 8% of visits with a Google AI summary, against 15% without one. Ahrefs reported in December 2025 that an AI Overview correlated with a 58% lower click-through rate for the top-ranking page (a vendor analysis, so methods vary). Content that only answers a simple question gets summarized and skipped. Content with original data, a clear opinion and proof gets quoted.
In February 2026 software stocks sold off sharply after agentic tools showed they could do work that used to need many licensed users. Commentators began calling it the SaaSpocalypse. The fear is simple: if agents do the work of ten or fifty people, a vendor that charges per human seat loses revenue. Vendors answered with hybrid models, such as seats plus usage, flat enterprise agreements and per-resolution fees, as reported for Salesforce and Zendesk. Many IT companies still report solid results, so read this as a repricing, not a collapse.
Layoffs.fyi counted about 122,600 tech layoffs in 2025, and trackers reported more than 100,000 already by the end of May 2026 (totals differ by tracker). Entry-level roles shrank at IT companies, while demand grew for people who can review, integrate, secure and sell.
G2 surveyed 1,076 B2B software buyers in March 2026. 51% start their research in an AI chatbot instead of a search engine, and 69% chose a different vendor than they planned after chatbot guidance. If assistants shortlist vendors, your documentation, comparison pages, case studies and reviews work as your first sales team.
National Bank of Ukraine data shows computer services exports of $6.66 billion in 2025 (up 3.3%) and $3.34 billion in the first half of 2026 (up 4.1%). Computer services made up 41.4% of Ukraine’s service exports, and the United States was the largest buyer at 36.9%. The figures exclude defense tech. Nominal growth is modest, so growth now comes from higher-value work and products, not from adding headcount.
2025 and 2026 at a glance
Sources and dates are listed at the end of the article.
A forecast is an opinion. Our forecast for IT companies rests on the data above and on patterns from more than 200 projects. Each shift in IT company strategy comes with a signal you can watch, so you can correct course early.
Agents shrink the number of humans who need a license, so every IT company selling software will keep adding usage, credits and per-outcome fees next to a base subscription. Buyers will ask what they pay for each result. Signal to watch: how many of your top 20 competitors publish usage-based or hybrid pricing.
With AI, an IT company of five strong people can ship what used to need fifteen. Routine coding gets cheaper, while architecture, domain knowledge, security review and delivery accountability get more valuable. IT service companies will sell fixed scope and outcomes instead of hours. Signal to watch: the share of your tenders that ask for fixed price and measurable acceptance criteria.
When many IT companies can build a decent product in weeks, attention is the limit. Owned channels win: an email list, a content library AI assistants can quote, communities and partners. Signal to watch: the share of your pipeline you can trace to inbound, and the number of AI answers that mention your brand.
Features copy quickly. Named case studies with numbers, security evidence, transparent pricing and an honest roadmap are slow to copy. Generated content floods every channel, so verified proof gets more weight. Signal to watch: how often prospects ask for references and security questionnaires before the first call.
If a large share of agent projects is canceled by 2027, surviving IT companies will be those with a measured return for the customer and a narrow scope. Signal to watch: renewal and expansion rates of AI features, not sign-ups.
| IT company type | 2027 outlook | 2028 outlook | What to do now |
|---|---|---|---|
| Outsourcing and services | Pressure on hourly rates for routine coding. Demand holds for architecture, integration, data and security. | Fixed scope and outcome contracts become normal in tenders. | Productize two or three offers with fixed price and scope. Publish proof from delivered projects. |
| B2B SaaS | Seat pricing under pressure. Hybrid pricing tests. Churn to AI-native substitutes. | Winners own workflow depth, data and a distribution channel. | Narrow the niche, test usage-based tiers, build email and GEO assets. |
| AI-native apps | Shake-out as model providers add the same features. | Survivors have proprietary data, community or contracts. | Own data and a niche. Measure customer ROI from day one. |
| Developer tools and infrastructure | Fast adoption, but a trust gap on quality and security. | Consolidation around products with proof and integrations. | Publish benchmarks, security evidence and teaching content. |
Our assessment of IT company outlooks, not a measured result. Re-check it every quarter against the signals above.
Most weak IT company strategies start from a build plan. Strong ones start from a startup model that answers who exactly buys, why they switch from the current solution, what it costs to win them and when the company breaks even. At City Profit this is the first step, and we call the package THE MODEL. Seven parts do the work.
The startup model for an IT product, step by step
Niche and Core Offer. A promise such as software for dental clinics that cuts no-shows in 30 days beats a platform for everyone. A narrow offer helps an IT company rank, get quoted and defend its position.
Pain research. We collect the exact words customers use for the problem, the alternatives they tried and the objections they raise. AI helps to cluster the material fast, and a person decides what it means.
Competitors and uniqueness. POSCO scores real competitors on five criteria to justify the price and to find the open position. The offer then adds something that is very hard to repeat: proprietary data, an integration nobody else has, a process, a community or a distribution deal.
Insured by experience. The model is checked against conversion rates, price ranges and channel ceilings from more than 200 projects, so the plan does not depend on optimistic guesses. If you want to learn to build such a model yourself, mdopen.pro teaches the method. For a written business plan and investor package, see bizonplanner.com.
A product ladder (ABCR) for an IT company, with examples
Every rung gives a visitor who is not ready for the core product a reason to stay in your list.
On top of the model comes the inbound system: articles that AI assistants can quote, comparison pages, an info product that grows your list, and a monthly email. That is what our Digital System 1.0 subscription runs: one new or redesigned funnel based on CRO, one extra SEO and LLM article, one extra info product, reporting, research, copy, media, automations, lead search for your email list, newsletters and 2 hours of your revisions every week. We do not run paid ads and do not negotiate with your clients. If a site is part of the plan, read why an AI-generated website carries risks before you choose a builder. For internal operations, see our comparisons of project management tools and of B2B partner search options.
Enter your own numbers. The calculator turns traffic into customers, then shows acquisition cost, lifetime value, payback and the month you become cash-positive. The defaults are illustrative, not benchmarks.
Calculator: CAC, LTV, payback and break-even
A simple monthly model with constant traffic and conversion. Real results vary, so run three scenarios: pessimistic, expected and optimistic.
Checklist: is your IT company ready for 2027?
Tick every statement that is true for your IT company today. The answers stay in your browser.
Readiness score: 0 / 12
Nobody can promise an answer, and the evidence is mixed. Adoption is high, but trust is lower and a 2025 trial found experienced developers slower with AI tools. The likely shift is in the work: less routine coding, more review, architecture, security and ownership of outcomes.
Not always. A better first step is to productize two or three services with fixed scope and price, publish proof from delivered projects and build an inbound channel. If demand repeats on one narrow problem, a product becomes a logical next rung of the ladder.
Expect hybrid models: a base subscription plus usage, credits or per-outcome fees, because agents reduce the number of human seats. Test pricing on a narrow segment first, and track revenue per customer and churn for each tier.
GEO means being quoted and recommended by AI assistants. For a vendor it means clear documentation, comparison pages, original data, named case studies, schema markup and consistent facts about your product across the web, so that an assistant has reliable material to cite.
Define the customer outcome before you build, for example hours saved, tickets resolved or errors prevented. Measure it for a pilot group, publish the numbers as a case study and track renewal and expansion rates of the feature, not sign-ups.
THE MODEL covers the niche and Core Offer, marketing strategy and competitor scoring (POSCO), a revenue model and forecast, product strategy with funnel mapping (ABCR) and an investor-ready pitch deck. It is checked against statistics from more than 200 projects.
Start from the calculator and from your unit economics. Many teams begin with a monthly budget that covers one strong article, one info product, an email and reporting, then raise it only when the cost to acquire a customer and the payback period stay healthy.
National Bank data shows steady, modest nominal growth: $6.66 billion of computer services exports in 2025 and $3.34 billion in the first half of 2026. The figures exclude defense tech. Growth now depends on higher-value work and products more than on adding people.
Order a free audit of your IT company, project, product or idea. In 30 minutes we review your niche, offer, funnel and search visibility and show what to fix first for 2027.
Order a free auditSee THE MODEL and Digital System 
