CI&T Q3 2026 Earnings Call Transcript

CI&T (NYSE:CINT) reported third-quarter financial results on Wednesday. The transcript from the company’s third-quarter earnings call has been provided below.

This transcript is brought to you by Benzinga APIs. For real-time access to our entire catalog, please visit https://www.benzinga.com/apis/ for a consultation.

View the webcast at https://www.youtube.com/live/AsDaI4eU2Mo

Summary

CI&T reported record revenue of $146.3 million for Q3, a 14.9% year-over-year increase, driven by double-digit organic growth and AI monetization.

The company highlighted a shift towards value-based commercial models, with 40% of new engagements signed under these models, improving gross margins by 420 basis points sequentially.

CI&T’s pipeline is 50% larger than a year ago, indicating strong demand, though macroeconomic factors like election cycles and currency fluctuations introduce some forecast uncertainty.

The company’s strategic focus includes expanding AI-enabled services and AI transformation offerings, as well as partnerships with companies like Anthropic and SpaceX.

Management emphasized that revenue per AI builder increased by 10% year over year, demonstrating improved productivity and efficiency from AI deployments.

CI&T expects Q4 revenue between $142.3 million and $152.0 million, with full-year guidance raised to $568 million to $578 million, reflecting 16-18% organic growth.

The company’s capital allocation priorities include continued investment in AI offerings, share repurchases, debt reduction, and selective M&A opportunities.

Full Transcript

Eduardo Galvan (Director of Investor Relations)

Good afternoon and thank you for joining us for CI&T 3Q26 earnings call. I am Eduardo Galvan, Director of Investor Relations. Joining me today to discuss our quarterly results are Cesar Gon, our Founder and CEO, Bruno Guicardi, Founder and President, and Stanley Rodrigues, our CFO. Before we begin, I would like to remind you that our remarks today will include forward-looking statements. These statements, including our business outlook, are based on management’s current expectations and are subject to risks and uncertainties that could cause actual results to differ materially. We caution you not to place undue reliance on these forward-looking statements as they are valid only as of the date when made.

Additionally, we’ll discuss certain non-GAAP financial measures. We believe these provide a more comprehensive view of our underlying operating performance. For a full reconciliation of these measures to the most directly comparable GAAP metrics, please refer to the tables in our earnings release. Today’s session is being recorded and all participants are currently in a listen-only mode. Following our presentation, we’ll host a Q&A session. To participate, please submit your question via email to investorsint.com.

The full presentation deck is available on our Investor Relations website and a replay of this call will be posted shortly after we conclude. With that, I’m pleased to hand the floor over to our Founder and CEO, Cesar Gon. Cesar, please.

Cesar Nivaldo Gon (Chief Executive Officer)

Thank you, Eduardo, and good afternoon everyone. This quarter we delivered record revenue. Our 80th consecutive quarter of double-digit organic growth. In a year of investment and transformation, that growth is now reaching our margins. Adjusted gross margin up 420 basis points and adjusted EBITDA margin up 220 basis points sequentially. Investors often ask us: why is CI&T growing when most of our industry is not? Four reasons working together. First, founder-led. We run the company through autonomous growth teams close to our clients with no layers of bureaucracy. That’s how we move fast. Second, deep client context. Our largest clients have trusted us for decades, and that trust gives us what AI needs to work in production: deep knowledge of their business and operating model. It’s hard to replicate.

Third, we are a rare scaled AI player and nimble. All our offerings are built around AI deployment, delivered by 6,800 AI builders with no legacy service lines to protect. Fourth, early to AI, multimodal by design. We launched CI&T Flow in the first half of 2023, then reskilled our teams and redesigned our career tracks around AI. So the way we develop talent matches the way we deliver. Flow orchestrates the leading models, so our economics do not depend on which model wins.

You will see these four reasons in our numbers: broadest growth, commercial models that pay us for what we deliver rather than for hours, and operating profit growing again quarter over quarter. CI&T puts AI into production at enterprise scale with measurable business outcomes. We delivered record revenue of $146.3 million, up 14.9% year over year, all organic, ahead of guidance. The growth is broad-based. Clients outside our top 10 grew 17.1%, faster than the 12.2% from our largest accounts.

So our momentum does not depend on a few clients. What changes this quarter is how growth turns into margin. Adjusted gross profit margin reached 36.6%, up 420 basis points from the second quarter and 270 basis points year over year. The main driver is AI monetization. In the third quarter, 40% of our new engagements were signed under value-based commercial models, up from 30% in the first half. More and more we are paid for what we deliver, not for the hours we bill.

Operating profit follows. Adjusted EBITDA grew 19.1% sequentially, with margin expanding from 13.3% to 15.5%, even as we step up investment in our commercial engine. That was a deliberate choice for 2026: new offerings, new verticals, and scaling Agentic SDLC across our clients. It’s already showing in our pipeline, now 50% larger than a year ago. Stanley will walk you through the details later in the call, including the currency effect. Two years of double-digit organic growth at constant currency.

More important than a streak is what produces it. First, what we sell. We put AI into production in two ways: services with AI built in like agent software development and AI transformation; modernizing legacy systems and data; and redesigning how a business runs with agents. Second, who we grow with. Existing clients are expanding, and new logos are ramping across industries and regions. The best proof: in production. Here are a few examples with numbers attached.

Kraft Heinz feeds millions of tables. But its sales team was hungry for more clarity. So CI&T built Data Mate, a copilot where machine learning turns raw data into the next best move. Active users grew 371%. 527% engagement. 88% of sales executives adopted it, the people who drive a third of all revenue. And in 2026, it took first place in AI at the Involve Awards. Because great sales, like great recipes, come down to the right ingredient at the right moment.

CI&T and Kraft Heinz together. The market keeps asking if AI actually works. We stopped answering in slides. We answer in production. More than 20 engagements, 15 industries across three continents, live at enterprise scale inside 40-year-old monoliths and greenfield builds alike. 84% lower cost per unit in mobility. 90% less human effort in legal tech. 14 days to 15 minutes in automotive configuration. 20 times faster in energy. All reinvented with Agentic SDLC. It’s a structural change of delivery.

We turn access into capability and capability into tenfold speed. The biggest numbers speak louder than the loudest market buzzwords. Explore more on our. Seven days just to finish the financial close. Teams, systems and data sources all waiting on each other. One calendar to tame with us. Getnet, a global top ten acquirer, planted Project Sequoia, governed data at the roots. One control layer for the whole close. AI agents validate and document every step. Seven days became two, up to 60% less operational effort. And the people, they step in only for the exceptions. Now expanding to seven countries.

Getnet and CI&T. Deep roots growing across borders. CALDIC is a global supplier of specialty ingredients to life sciences and industry in Latin America. It grew by acquisition. Its data grew the same way—fast, just not in one place. Two ERPs, scattered platforms, a month-end close done by hand. We built a data factory on Databricks, governance in the formula. Close time dropped at least 30%. Finance stopped consolidating and started analyzing. Chemists call that a solution. So do CFOs.

A quick recap of our quarter. First, our own stages at the New York Stock Exchange. We hosted Impact AI with Anthropic, AWS, Mars and Sonos to talk about what it really takes to make AI work at scale. Then a world first: the Lean AI Summit with Lean Institute Brazil, Ford, Itaú, Johnson & Johnson and others shared AI-first cases built with CI&T and Box 1824. Next came the AI Deployment Room, 30 plus C-levels on one question: how to deploy AI at scale.

The answers came from people already doing it. With Fundação Dom Cabral, we launched the second edition of the AI Lighthouse Awards, recognizing companies whose AI maturity shows up in actual results. Now the stages we stepped onto. At Big Data London, we moved the conversation from what AI can do to how to make it work inside the business. At Databricks World Tour in Singapore and São Paulo, we show data and AI running in production alongside great partners. And in a webinar with Craig LeClair, VP and principal analyst at Forrester, our co-founder Bruno Guicardi explored agents and the future of work, why most never escape proof-of-concept purgatory, and how a few trailblazers are leading the way out.

That’s our quarter. See more on our website. Bruno will now show you what this means in productivity: revenue growing faster than our team. Bruno, I hand it over to you.

Bruno Guicardi Neto (President, North America and Europe Operations)

Thank you, Ceci. Good afternoon everyone. The clearest measure of our new model is revenue per AI builder. On a last 12 month basis it reached $82.5k, up 10% year over year, an acceleration of 7% last quarter. What drives it is simple. We closed the third quarter with about the same number of AI builders as a year ago, while revenue grew 14.9%. As our people now deliver more with AI, our revenue no longer moves one to one with headcount. With our new commercial models, we are paid for what we deliver, not for the hours we bill.

And that’s what lifts our gross margin. And the team is stable. Voluntary attrition was 9.9%, among the lowest in our history. Let me give you an example of one of our offerings that turns AI into measurable decision intelligence. Our clients have built strong data and analytics foundations. Analytics tells them what will probably happen. Decision intelligence takes the next step. It turns that prediction into a decision made at the right time. It closes the loop. A model detects the signal, business rules shape the recommendation, a person approves and the action runs directly into the client’s operational systems.

Every outcome feeds the next decision and people stay in command. Clients start with a human approving every decision in the system. More autonomy, only as trust builds with every step auditable. To deliver outcomes like this, clients need the best model for each problem, not a model their provider needs to sell. That’s why we are multi‑model by design and why the model builders want to work with us. Through the Google Cloud Partner network, we reached the preferred tier in the services track. We are joined with a Certified Delivery Team, cloud use cases delivered for clients, and public customer references. We have more than 50 major projects underway with cloud across software engineering, customer service, and user experience for global clients such as Shredo and Vanguard, and we use it ourselves.

CI&T Flow runs extensively within CI&T Flow, so we bring clients lessons learned from our own integration and scale. In October we also became a launch partner in the SpaceX Superintelligence Network, adding GROK to the models orchestrated by CI&T Flow alongside Mistral and the hyperscalers. Even the companies that build the models need partners to make AI work inside enterprise. That’s the layer where CI&T operates. With that, I will hand it over to Stanley to talk through how our delivery engine shows up in our numbers.

Stanley Rodrigues (Chief Financial Officer)

Thank you, Bruno, and good afternoon everyone. Revenue reached a record $146.3 million, up 14.9% year over year, entirely organic and ahead of our outlook of at least $145.7 million at constant currency. Revenue grew 11.2% for the first nine months. Revenue reached $425.7 million, up 19.8% or 13.6% at constant currency. Let me show you where that growth came from. By region, Latin America led at 20%, followed by new markets at 17.6%. North America grew 8.7% year over year and its momentum is building. Revenue in North America grew 3.9% from the second quarter, the fastest sequential growth among our three regions. By industry, financial services, our largest vertical, grew 26.1% and technology and telecom 40.2%. And by client, revenue from clients outside our top 10 grew 17.1%, ahead of the 12.2% from our 10 largest accounts.

That is the third consecutive quarter in which our broader base grows faster than our largest clients. We are adding new clients and our offerings are reaching more of the accounts we already serve. Our growth does not depend on a few large relationships. Last quarter we said AI monetization would expand our gross margin. This quarter it did. Adjusted gross profit reached $53.5 million, up 24% year over year and 15.8% sequentially, growing well ahead of revenue. Adjusted gross margin reached 36.6%, up 270 basis points year over year, our first year‑over‑year expansion of 2026, even with the stronger Brazilian real and the payroll tax resumption posing headwinds. There is a concern in our industry that AI productivity challenges revenue when paid by the hour, but it is not how we are increasingly being paid.

In our value‑based models, we are paid for output and outcomes. So instead of a discount, productivity becomes margin. That is why 40% of the new engagements we signed this quarter are value‑based, up from 30% in the first half, and CI&T Flow lets us deliver more with the same team. Our cost of service grew 2.9%, below revenue growth of 14.9%. AI makes us more productive. Our commercial model lets us and our clients share the value. Last quarter we said we expected our adjusted EBITDA margin to improve sequentially. It did. Adjusted EBITDA reached $22.6 million, up 19.1% from the second quarter, with margin expanding from 13.3% to 15.5%. Year over year, adjusted EBITDA was 3.9% lower because of the stronger Brazilian real. On an FX‑neutral basis it would have been $24.3 million, up 3.4% from the third quarter of 2025 with a 17.2% margin excluding currency. Our EBITDA is already growing while we keep investing in growth.

This expansion came even as we stepped up investment in our sales engine. Selling expenses were 14% of revenue, up from 11.9% in the second quarter. That reflects the structural expansion of our commercial organization for new offerings and verticals and the targeted investment to scale our agentic SDLC initiative, which we concluded this quarter. From here we expect selling expenses as a percentage of revenue to decline from third‑quarter levels.

Those investments are already paying off. Our pipeline is 50% larger than a year ago and we are adding new clients at the same time. General and administrative expenses fell to 2.2% of revenue from 11.3% in the second quarter and 11% a year ago, delivering operating leverage across the rest of our structure. Moving to the bottom line, adjusted profit was $11.5 million, up 2% year over year, and adjusted diluted earnings per share was $0.09, up 4.5%. Earnings per share grew faster than profit because of our capital allocation. We repurchased $2.9 million in shares this quarter and $9.1 million year to date, and our weighted average diluted share count is 2.4% lower than a year ago. We buy back our own shares because we believe in the value we are building.

Cash generation remains strong. Operating cash flow was $22.4 million in the quarter, nearly 100% of adjusted EBITDA. With that, I will turn the call back to Cesar for our business outlook.

Cesar Nivaldo Gon (Chief Executive Officer)

Thanks, Stanley. Fourth quarter we expect revenue in the range of $142.3 million to $152.0 million. At the midpoint, that is a 9.6% growth in US dollars or 9.3% at constant currency, against our strongest comparison base of 2025. The range reflects Brazil and the United States, our two largest markets, coming out of election cycles. Clients are setting new budgets and currencies are more volatile. Demand is strong. Our pipeline is 50% larger than a year ago.

The question is when demand converts, not whether it does. For the full year, we are raising the low end of our revenue outlook to $568 million, with a range up to $578 million, or 16 to 18% organic growth. Our revised outlook includes a positive FX impact of approximately 450 basis points, and we continue to expect adjusted EBITDA margin in the range of 15 to 17%. This quarter showed that our model works. We said gross margin would expand, and it expanded 420 basis points. We said EBITDA margin would improve, and it improved 220 basis points. Revenue is growing faster than our team and our broad client base is growing faster than our largest accounts. We close 2026 with a strong model and a large pipeline, well positioned for 2027. We put AI into production for large enterprises, and increasingly our clients pay us for outcomes, not hours.

I want to thank our clients for their trust and every CI&Ter for the energy and talent behind these results. With that, we are ready to begin the Q&A session. Thank you.

OPERATOR

All right, we’ll now begin the Q&A session. I’ll announce each participant’s name. Once you hear your name, please unmute your line and ask your question. Then when you’re done, please mute your line. The first question comes from Gates Schwartzman from TD Cowen. Hi, Gates.

Gates Schwartzman (Analyst at TD Cowen)

Hey, thank you for taking my questions, on for Brian Bergen today. So you guys had mentioned the 50% growth in the pipeline this quarter. That’s an uptick from 40% last quarter. Can you help us parse out what are some of the underlying trends between what’s core pipeline growth in agentic SDLC versus maybe more of your traditional digital engagements? Additionally, you’ve seen some good growth here in the top client cohort. Just curious, through both of these—pipeline and top client growth—do you have any visibility now into next year, any sort of early considerations in terms of growth levels? Thank you.

Cesar Nivaldo Gon (Chief Executive Officer)

Thanks, Gates. I can start with your questions first. We disclosed that in this third quarter 40% of our new sales were already new value‑based commercial models, and an additional data point is for new clients this number goes to 75%. So for every single new sale for new clients, we are being able to basically convert to the new value‑based commercial models. That is the first big news in terms of the demand and our ability to convert the pricing model.

In terms of demand type, we see basically three. We are using this model of three types of demand. The first is AI‑enabled services, basically agentic SDLC. That is probably half of what we are selling now. The second is AI transformation that is a package that includes modernization, data including ontology, decision intelligence, and so on. This is basically going ahead and preparing the company for AI transformation. And then we have recently launched what I consider the crown jewel of our offerings that is the GenAI reinvention, the business reinvention with AI.

That is the most transformative offering and we are getting very good traction in these first two quarters of launching this new offering. So the trend is companies are combining preparing, enabling the services with agentic SDLC, a lot of traction with our AI transformation offerings, and initial traction for the new business reinvention that probably will be more relevant in 2027. You asked regarding visibility: we have a very strong pipeline, 50% larger than the same period last year, so we are foreseeing a very, very good year ahead. But right now we are focused on our Q4, and we’ll be back here in the next call to give you details about what we are forecasting for 2027.

Stanley Rodrigues (Chief Financial Officer)

I think, if I can complement, you asked about how much of bookings we already have for next year, right? So remember, most of our growth comes from the portfolio clients, right? Usually less than 10% of the year comes from new logos and land activity. So we do have visibility into most of what we are going to do in 2027.

Gates Schwartzman (Analyst at TD Cowen)

Understood. One more on my end. So a lot of moving parts with gross margin this quarter. Can you maybe help unpack some of those moving parts—you’ve gone over FX. Just saying, we know that some of those investments also in the agentic SDLC stopped this quarter. Can you help unpack maybe how we should expect SG&A and gross margin to trend in the coming quarters, especially now as you take more and more engineers to 10x productivity, to 20x productivity, etc.?

Stanley Rodrigues (Chief Financial Officer)

Gates, well, thank you for the question. Gross margin. We have two major factors there. First one is FX, as you mentioned. The other one is the deliberate investment we are making towards sales and agentic SDLC. So let’s unpack this. The FX, simple, 170 bps. If you compare the year-over-year margins, we dropped 300 bps. So out of those 300, 170 is with regard to FX. The other 130, that’s derived mainly from the investment. And we have within that number we have a bunch of things.

For example, first, the movements we are making towards adding value-based commercial models has been added. You may see, for example, revenue per headcount improving by 10%. You see, for example, we grew 14.9% in sales while we grew only 1.8% in headcount in that period. So also we have built in there, we have a payroll tax resumption in Brazil, which is a headwind. Of course, the sales and agentic SDLC by themselves represents 600 bps. We have G&A operating leverage helping as well there.

So all combined into those 130 bps. There, sequentially, we have a gross margin improving by 19.1%, which is 220 bps, and 106 bps, if you would, by FX-neutral. So that trend will continue. So you mentioned what we should expect, right, with regard to margins. So you should expect incremental presence in our mix of those new models, new commercial models, and also slowly that migration from traditional time-and-material into those new models that will reflect in the margins, also the operating leverage on both the G&A and also sales.

You should expect that thing happening towards the future. So mainly those elements. Understood. Thank you. Thank you.

OPERATOR

Thank you, Gates. The next question comes from Gustavo Farias from UBS. Hi, Gustavo.

Gustavo Farias (Analyst at UBS)

Hi everyone. Thanks for taking my questions. So, two things on my side. So first of all, if you could comment about partnerships, the several partnerships that you guys have, Anthropic, Mistral, and all the others, how that has been performing so far, how much of it, if any, has already contributed in this quarter. We should expect it to kick in further in the future. And the second one, if you could please update us on your capital allocation priorities for 2027.

Thank you.

Cesar Nivaldo Gon (Chief Executive Officer)

We can take the first one; Stanley can take the second one. Partnerships has been important for us, Gustavo. Today it’s an important channel. Like, all partners combined, they represent around 20% of kind of the—we measure this as we call partner attachment—like deals that have a partner, either that we brought the deal to the partner, we went together, or the partner brought it to us. So 20% in 2026, it’s already with partners. So it’s a big increase comparatively to the kind of the last years.

And I think we’re just going to grow in 2027. Those partnerships with the frontier labs are more recent and they’re very exciting in terms of types of opportunities that they are bringing. Not very relevant for the revenue in Q3 yet, but we hope that will be very relevant for the business generation in 2027.

Stanley Rodrigues (Chief Financial Officer)

Gustavo, can you repeat the question, please?

Gustavo Farias (Analyst at UBS)

Of course. Just an update on your capital allocation priorities for 2027.

Stanley Rodrigues (Chief Financial Officer)

Okay, thank you. We, of course, first priority is to continue this trajectory on investments towards the combination of AI new offerings, which is key for our future. We have been very successful in everything that we do and our bets, so very assertive. So that’s the main priority. We have the opportunity to continue our share buyback program, which we will. We continue to invest. Of course, growth itself demands capital. As we’re growing, this is kind of automatically happening.

And we’ve been paying down debt as well. So I would say those four items would be the—and we are very selective towards M&A, for example. We are not very proactive in terms of thesis. But we do see opportunities here and there. And if we see any opportunity, we may tackle those opportunities. Right.

Gustavo Farias (Analyst at UBS)

All right, very clear. Thank you.

OPERATOR

Thanks, Gustavo. Our next question comes from Steven from Wedbush. Hi, Steve. Please go ahead.

Steven (Analyst at Wedbush Securities)

All right, good evening, everybody. Thank you for taking the questions. I want to start off with the comments around the pipeline because you mentioned that the pipeline is actually doing quite well; it’s up 50% year over year. Can you talk about the breakdown between the new and existing clients of this pipeline? Is majority still coming from existing, expanding their deals as you mentioned on the call, or is there a lot of new logo traction? And then specifically when it comes to converting that pipeline, I mean, it seems like there’s a lot of positive momentum there, but the organic growth is actually not heading in the right direction—it actually decelerated in the quarter—but we’re still seeing a lot of positive momentum in the pipeline. So is there any issue with actually getting deals to convert from the pipeline to revenue?

Cesar Nivaldo Gon (Chief Executive Officer)

Thanks, Steven. Basically we are seeing, of course, as a result of our increase in our very relevant increase in our sales pipeline—50%—it’s basically continued to follow. Seventy percent of this pipeline is from our current client base and 30% are related to potential new clients. So it’s a larger pipeline, but following the same blend-and-expand pattern as always. Regarding the conversion, I mentioned during my explanation that we just got out of two election cycles in Brazil and the U.S., so we kind of add some caution in forecasting, and the visible effect really affects volatility a lot. As you know, we have a representative number in revenues in Brazilian real. So that explains the range we are forecasting Q4. But the pipeline is there. The conversion rate is in a max of our history. So it’s going to convert—it’s just a matter of when it’s going to convert.

And we are kind of adding this layer of conservative because of this recent macro news. But I think that we are very, very, very assertive that all this effort is paying off and increasing our pipeline, and we will continue to have an amazingly good conversion rate. So I’m sure we are in the right direction, and combining this with expanding our gross margin—that is the original way: AI deployment for growth and AI monetization to increase our margins.

And both things are important.

Steven (Analyst at Wedbush Securities)

Understood. And if I can just squeeze one more in, because, Stanley, I think from one of the answers to your question that the employee headcount actually increased 2% year over year, which is a slower growth rate than we have seen over the past couple of quarters. Was that intentionally done? Was there anything that impacted you from hiring more sales capacity to accelerate that growth even further? Can you give us more color on what to expect moving forward?

Stanley Rodrigues (Chief Financial Officer)

Yeah, we may say it’s intentional as a consequence of everything that we are doing with AI, the way we are delivering everything that we are producing in terms of services. So it is a natural consequence as we improve. As I said, the value-based commercial models, they require less of that relationship that used to be one-to-one, right. Anything that would grow in revenue would grow in headcount. This is changing. So it’s intentional from the business perspective, and this is a natural consequence.

Steven (Analyst at Wedbush Securities)

Understood. Thank you.

OPERATOR

Right, thanks, Steve. The next question comes from Luke Morrison from Canaccord. Luke, please go ahead.

Luke Morrison (Analyst at Canaccord Genuity)

Hey guys, good to see you. Thanks for taking the questions here. So I want to ask one that [is about] the relationship between sort of revenue growth from here and your gross margin. So I’m wondering, as you’re moving existing customers onto your value-based models, what happens to the size of that relationship? Does the client end up spending about the same with you and just getting more out of it? Are they giving you more budget because you’re doing things more efficiently and so they’re taking spend that would otherwise go to competitors and spending more with you, or do the savings you’re delivering them come out of whatever is left in their budget? Basically just trying to understand the extent that the higher margin comes with durable growth going forward.

Cesar Nivaldo Gon (Chief Executive Officer)

Sure. Thanks for your question, Mike. Basically, the model for existing engagement was kind of deploy agentic SDLC and getting—replacing some competitors or low-performance competitors, I would say. So basically we try to keep the same volume but getting more client share and expanding our margins and sharing part of the benefits [with] clients. I think this was the main pattern of keeping the revenue, increase our margins, and also help our clients to get efficiency around the deployment of agentic SDLC.

But we did across the board, we’re still working on getting more client share by replacing non-AI or low-AI-performance competitors. And this is basically the strategy of evolving the value-based models but not destroying or reducing the revenue level that we have. Look, if I can complement.

Bruno Guicardi Neto (President, North America and Europe Operations)

If you think about how this industry works, it’s like 80% of what our clients spend is just to keep the light on, and just enhancements and kind of keep compliance going. So when you have the chance to actually do more and actually innovate and partner with business areas, and actually do things actually they’ve been in office for years—so start, we call it playing on attack, and stop just playing defense—that’s what our smartest clients are doing.

They’re just taking advantage that all things can do it. You can do things faster and cheaper, and then do more, and kind of stop playing defense and doing more, capturing in some cases just client—also capturing some client pocket share. You’re right. And also, very risky, pre-AI—for example, legacy modernization was considered a very risky business before AI. Now it’s not just fast and cheaper, it’s riskless because you have AI to understand the legacy, update that. So a lot of very complex things now are less risky and the equation of return on investments are better. So AI is kind of a positive force in terms of leveraging demand in such large organizations.

Luke Morrison (Analyst at Canaccord Genuity)

Yeah, helpful. Thank you. And then just maybe a follow up. The model providers over the last few months, they’ve announced huge investments in various services-led initiatives and ventures. As you see that happening, as those model providers build out their own deployment teams and partner initiatives, how do you see the economics and the share of business splitting between the labs and firms like yours?

Cesar Nivaldo Gon (Chief Executive Officer)

Thank you for these questions because this is probably the most amazing news for us. It proves that AI will increase AI deployment offering by any scale that no one will be able to fulfill. So they are worried about the lack of AI deployment companies. And I think we have a scale AI deployment now in a number that is very rare in the industry. So of course the economics will converge. I don’t believe the fee levels these labs are promoting are sustainable.

But the main concept [is] that to deploy AI in large organizations is, first, it will demand a lot of specialized services. Second, it’s a long-term journey. We have 10 years ahead of demand around AI deployment. And this is very, very good news for us, as we foresaw this opportunity almost four years ago.

Luke Morrison (Analyst at Canaccord Genuity)

Excellent. Thank you.

OPERATOR

Thanks, Luke. The next question comes from Maria Clara. Hi, Clara.

Maria Clara (Analyst)

Hi guys. Thanks for the opportunity. I would like to explore a little bit more on how you see the pipeline evolving per geography, especially in Brazil. Can you please elaborate how clients’ willingness on investing in AI have been evolving, especially taking into consideration the macro landscape? And a second question related to engagement levels. Since you are noticing an improving productivity within the same clients, do you have any operational KPIs such as churn that could be shared with us so that we gain comfort that your retention rates keep improving?

Thank you.

Cesar Nivaldo Gon (Chief Executive Officer)

Thanks for the questions. The first, I think it’s too early. We have a very strong pipeline in Brazil as in the US. Too early to measure the effect of the recent auction. Let’s see. I think somehow these investments are inevitable. Companies, for competitive reasons, cannot avoid capturing the efficiency. And then the differentiation is around the artificial intelligence. I think we move from the pilot moment in the industry to the production moment, and that will increase the competitive barrier in different industries and verticals.

And that will lead to more demand around that. In the short term, maybe some delays. But so far we don’t have any concrete results, only the effects of the new level for the Brazilian to US dollar ratio. The second question is regarding churn. You mean client churn? Yeah. Our rotation is—we normally disclose the revenue retention rate annually. So next, in the next quarter, we will provide the number for 2026. And we don’t see any meaningful variation. I think we continue with a very strong retention. That’s why the majority of our growth continued to happen in our portfolio in the short term, and new clients totally contribute for year two and year three as we expand.

So the land-and-expand engine continues to work, and we don’t see any relevant difference around that. Thank you. Thanks.

Eduardo Galvan (Director of Investor Relations)

Clara. Next question comes from Pune Cheng from J.P. Morgan. Hi, Pune.

Pune Cheng (Analyst at J.P. Morgan)

Hey, thanks for taking my question. So nice improvement in gross margins on a sequential basis. So my question is, how high your gross margins can go as the new models continue to improve in mix? In other words, how should we think about incremental margins at GM level on some of the new work you do?

Cesar Nivaldo Gon (Chief Executive Officer)

Thank you, Pune. I think the equation, our equation for gross margin improvement, is less related to the capabilities of the models. I think we have a multi-model model where we can combine the best of breed for each kind of problem we are solving in terms of capabilities and costs. I think the equation is more related to our ability to continue to convert the commercial models. This is the main driver for improving our gross margin. And there’s part of this equation is IP.

As more intellectual property we have in our offerings, it clearly increases our ability to speed up the introduction of value-based models. So I think there is a lot of room to continue improving our gross margin, and this is, I think, what we are by design—executing a company that can expand revenue and margins not in a linear way with headcount. And this is already happening as our numbers are showing. So I think there’s still a lot of room for gross margin improvement based on commercial models and intellectual property delivered.

Pune Cheng (Analyst at J.P. Morgan)

No, I’m sorry, I should have clarified. When I said models I meant like this new—

Cesar Nivaldo Gon (Chief Executive Officer)

Delivery. Okay, okay. But you’re referring to the LLMs.

Pune Cheng (Analyst at J.P. Morgan)

Yeah, yeah. No, but is there a way to think about incremental margins? Like the new work that’s coming in under these models, delivery models—how much of GM is for those IP-based work or for some of these agentic SDLC and all those.

Cesar Nivaldo Gon (Chief Executive Officer)

Yeah, the current number is, in average, we have 8 to 10 percentage points above the old traditional models. So it’s very relevant, and it will continue to be. Of course, it’s an equation of differentiation. Yeah, no—

Pune Cheng (Analyst at J.P. Morgan)

That’s amazing. That’s good to know. And then your fourth quarter guidance—the range was wider than usual, like it’s give or take about 10 million. So what were the reasons for that? And then what can drive acceleration from 4Q levels into next year? You talk about some of those pilot projects and all those that can come through. Elections are behind you in Brazil and midterms in the US, so how should we think about 2027 growth as it relates to where you’re going to exit in 4Q of this year?

Cesar Nivaldo Gon (Chief Executive Officer)

Okay. For Q4, the low end is delays, macro uncertainty, and FX volatility. And the high end is pipeline conversion. It’s the best pipeline in our history, and we are sure we’re gonna convert a good chunk of it, but not sure if it’s gonna happen within 2026. And so even though the prospect for next year is very good, as we believe, we are kind of moving the pieces of our strategy. We have more offerings coming out of our differentiation strategy with more IP-based offerings.

We have data evolving in an amazing way as a big part of our strategy. We have partnerships, as Bruno mentioned, evolving—it’s around 20% of our sales and it will increase for sure, especially with this new partnership with Anthropic and other frontier models. So a lot of things are moving in the right direction. Of course, we cannot anticipate our outlook for 2027. It will be the main topic of our next meeting here, for sure.

Pune Cheng (Analyst at J.P. Morgan)

Okay, thank you.

Eduardo Galvan (Director of Investor Relations)

Thank you, Pune. Thanks, Pune. That concludes our Q&A session. I’ll now invite Cesar Gon to proceed with his closing remarks.

Cesar Nivaldo Gon (Chief Executive Officer)

Amazing. Thank you. Stanley, Eduardo, thank you all for joining us today. To all CI&Ters, thanks for your hard work. And to our clients, thank you for choosing CI&T as our partner for AI deployment. As a final note, we said AI would show up in our margins. It did. Same thing: more revenue, better economics. So we are building a company that scales revenue faster than headcount and grows more profitable as our new commercial models mature. That is the main message here. Stay well. See you soon. Bye-bye.

Disclaimer: This transcript is provided for informational purposes only. While we strive for accuracy, there may be errors or omissions in this automated transcription. For official company statements and financial information, please refer to the company’s SEC filings and official press releases. Corporate participants’ and analysts’ statements reflect their views as of the date of this call and are subject to change without notice.