管理層發言
Good afternoon. And thank you for joining us for CI&T's second quarter of 2026 earnings call. I am Eduardo Galvao, Director of Investor Relations. Joining me today to discuss our quarterly results are Cesar Nivaldo Gon, our founder and CEO; Bruno Guicardi, founder and president for North America and Europe; and Stanley Rodrigues, our CFO. Before we begin, I would like to remind you that our remarks today will be 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 will discuss certain non-GAAP financial measures. We believe these provide a more comprehensive view of our underlying operational 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 will host a Q&A session. To participate, please submit your question via email to investors@cint.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 it. With that, I am pleased to hand the floor over to our founder and CEO, Cesar Nivaldo Gon.
Thank you, Eduardo, and good afternoon, everyone. Global AI spend is projected to hit $2.6 trillion this year, up 47% year over year. And yet, according to MIT, roughly 95% of AI initiatives still show no measurable business return. That gap is where I want to begin today. We published two papers this quarter that get at why. The first argues that most companies are optimizing the wrong variable: chasing incremental task efficiency instead of asking where AI can return 10x rather than 10%. The second paper calls it organizational hallucination: the confident belief that a company is transforming when it is actually just experimenting. In both cases, the constraint was never the technology; it is the organization's capacity to absorb it. That gap is exactly where CI&T plays. And it is why we build our business around two things: AI deployment—installing real capability inside a client's core—and AI monetization—capturing together with our clients the productivity gains and business impact that AI deployment creates through value-based commercial models rather than headcount. Everything you will hear from us today—the robust and sustained revenue growth, the increase in our sales investments to foster momentum given the AI opportunity—is the same thesis playing out inside our own business. This quarter's numbers reflect that opportunity and the deliberate choices we are making to capture it. We delivered record revenue of $142.8 million, up 21.9% organically and above our guidance. Growth was broadly based across geographies, industry verticals, and client cohorts. Growth was increasingly fueled by new client wins and by initial engagements scaling into large partnerships. Our new commercial models are also letting us capture a greater share of the value we create. That shows up directly in adjusted gross margin, which expanded from 30.6% in the first quarter to 32.4% as these models scale. In the first six months of 2026, 30% of new engagements were under new value-based pricing models, and we project this gross margin expansion to accelerate in the coming quarters. Our adjusted EBITDA margin in the quarter was 13.3%, reflecting our deliberate choice: 2026 is a transition year where we invest in our commercial engine to turn this AI deployment opportunity into durable, profitable growth in 2027 and beyond. As a result, our commercial pipeline is now 40% larger than in the first half of 2025. In short, our top line shows the demand is solid, our gross margin shows monetization is working, and this year's commercial investment is what lets us compound that advantage going forward. The second quarter of 2026 marks our seventh consecutive quarter of double-digit organic growth. At a time when parts of our industry are consolidating through acquisitions, buying growth rather than building it, we have extended this streak without a single M&A deal—prioritizing our capital allocation toward our own transformation, and our sales effort and growth engines. This embodies one of our cultural tenets: we play the infinite game. We are not optimizing for a single quarter. We are building a company designed to keep compounding for decades. This consistency reflects a structural shift in client demand and CI&T's ability to capture it. The case studies that follow show how this AI deployment momentum is translating into tangible business outcomes. MRV, Latin America's largest homebuilder, now has rebuilt how it speaks to the world: five brands, one voice. CI&T with Adobe laid the foundation for a new digital experience and then handed over the keys. The marketing team runs it alone now—no scaffolding, no outside help left on-site. Three months in, the traffic tripled: 117% more people at the door. MRV brought the ground, CI&T brought the blueprint; the windows opened on their own. At Alpargatas, home of Havaianas and Rothy's, we spent a full day at our HQ in Campinas, Brazil, mapping the digital terrain together—from AI-powered commerce to the journeys that turn products into seamless experiences. Some journeys are better walked with the right partner. At Alpargatas and CI&T, I think brands that will be successful in this new age are brands that have solid foundations—found data foundations and awareness of customer behavior—because it is not necessarily about speed to market, but how fast you can learn, pivot, and build experiences that really matter and resonate for the customer. The agentic enterprise is not coming; it is here. And it just got a bold ally. CI&T just joined Anthropic's Claude Partner Network. Our engineers are now certified in Claude; Claude code is wired into CI&T Flow. Thirty years across the globe, now setting the standard for how the world's largest enterprises deploy AI. Options generated, outcomes decided. This is the power of a partnership of a global AI deployment partner boosted by Anthropic. We got your back. A quick recap of our quarter: first, we launched Organizational Hallucination, a new pocketbook by the CI&T team with chapters from myself and Silvio Mera on why companies invest in AI and keep solving the wrong problems. Then came The Wrong Math of AI, a paper from the CI&T team, our CFO Stanley Rodrigues, and cofounder Bruno. Its bottom line: AI will not transform your organization; you will. Business Complexity Points (BCP) went open source—ten years of work with Itau, one of Latin America's largest banks, now free for everyone on GitHub. The Retail Tech Report, Agentic Edition, landed next, led by Melissa Minkow, our global director of retail strategy, on how agentic AI is already reshaping retail. We were around the clock at the biggest stages in AI and innovation this quarter. We introduced the ESG Consumer Index 2026, a sharp read on what people now expect brands to prove. And one milestone stands out: we are the first software company in Latin America with SBTi-validated net-zero targets—climate action measured by science, not marketing. Beyond that, a partnership with Mistral, a pioneer in open-weight AI, to power the next generation of agentic enterprises. That is our quarter. Explore more at our website. These case studies demonstrate how our Agentic SDLC and CI&T Flow are resetting the baseline for enterprise productivity and speed to value. I will now hand it over to Bruno to discuss how we are scaling this hyper-productivity to our global delivery model and our evolved talent strategy.
Thank you, Cesar. Good afternoon, everyone. I am glad to share our operational and talent progress for this quarter. We closed Q2 2026 with roughly 8.1 thousand professionals, with voluntary attrition at 10.1%, continuing to trend toward some of the healthiest levels in our history. At the center of this workforce are our 6.7 thousand AI builders—the result of reskilling 100% of our professionals to work with AI natively. That matters right now. Recent independent research mapped a widening gap in the market for AI deployment talent. Demand for engineers who deploy AI at enterprise scale is growing roughly 50% year over year. Most of our industry is racing to hire into and increase its core pool of talent. We did not have to. We built it from within ahead of the market. Revenue per AI builder continues to grow, reaching over $80 thousand in Q2 2026 on a last 12-month basis, an increase of 7% year over year. This is a direct result of AI monetization and value-based pricing, providing operating leverage and contributing to the expansion of our gross margin. Our momentum is being reinforced by a strategic partnership we announced this quarter—one that speaks directly to the role CI&T plays for large enterprises. We joined Anthropic's Claude Partner Network, certifying more than one thousand AI engineers on Claude and working with Anthropic to help set a new standard for how AI gets deployed inside the world's largest organizations. We already run Claude code extensively inside CI&T Flow, and this partnership expands that work into a joint go-to-market motion with a focus on co-developing industry solutions for financial services, retail, consumer goods, and other verticals. It is designed to open new enterprise accounts and expand our pipeline into verticals we are co-developing, directly feeding our commercial momentum. Large enterprise: meet a partner who can take frontier models into complex, regulated, mission-critical environments and deliver production-grade outcomes. That deployment layer is exactly where CI&T operates. This partnership makes us the connective tissue between the leading model providers and the world's largest organizations. That embedded engineering capability is exactly what enables us to play where the real value is shifting in the AI era. In this chart from Forrester, the clearest way we found to show clients why: most of what is being sold as enterprise AI so far sits on the left side of this chart—automating individual tasks with copilots and agents, augmenting existing workflows end to end. It is a productivity story—efficiency gains on top of an operating model that stays the same. The disruption is what Forrester calls the process chasm. Cross it and the business case changes entirely—from efficiency to new revenue and margin structures, and ultimately to growth, relevance, and the long-term perpetuity of the business itself. CI&T built two offerings specifically for the two quadrants on the right side of that chasm. The one I want to walk you through now—the one gaining the fastest traction with our clients today—is agentic enterprise reinvention. Agentic enterprise reinvention is how we help enterprises redesign the core of their operations, moving from legacy ways of working to agentic-native operations. We do not stop at advisory or isolated use cases. We install real operational capacity inside existing value streams and we stay until it runs on its own. In practice, that means bringing 60% to 80% of a core end-to-end process onto an agentic journey in months. Three things differentiate this from traditional systems integration: who delivers it—small, senior forward deployment engineering teams; how we find the value—three decades of lean-based process transformation along with our industry expertise give us the insight into where reinvention pays off; and how we get paid—increasingly outcome-based with fees tied to business results, not to hours built. And it compounds: we reinvent one core process and it becomes the reference architecture for the rest of the enterprise, giving every account a natural land-and-expand path. That means revenue that scales with the impact we create for clients, not with headcount. Now I will hand it over to Stanley to comment on our financial performance.
Thank you, Bruno, and good afternoon, everyone. Let me walk you through our financial results for the second quarter of 2026. As Cesar mentioned, we delivered record net revenue of $142.8 million, up 21.9% year over year, entirely organic, and 14.1% at constant currency—above our guidance of at least $140 million. This performance reflects the strength of our go-to-market execution. Through the quarter, we saw our sales pipeline expand and our conversion rates improve—the direct payoff of deliberate commercial initiatives and the tangible results AI deployment delivers for our clients. Beyond the headline number, what matters is how broadly this growth is spread across our footprint. This slide shows the composition of our growth and the message is clear: our momentum is not carried by any single vertical or client. Every region contributed. Latin America was the largest engine, expanding 32.1% year over year. New markets grew 26.3% and North America added a consistent 10.2% on a large mature base. Financial services, our largest vertical, continued to grow strongly, up 36% year over year. Technology and telecommunications accelerated to 68% growth—a robust turnaround from the contraction we saw just a year ago. Others grew 25%, life sciences 16%, and retail and industrial goods 11%. The one exception is consumer goods where demand has been softer, a headwind we view as temporary. The composition by client cohort tells an equally healthy story: clients outside our top 10 grew 24.1%, outpacing the 19% growth of our top 10, reinforcing that our momentum is not dependent on any single account. Taken together, this confirms that our AI deployment is a global catalyst driving deeper penetration across every region and every client tier we serve. As you may recall from our last quarter's call, we said that as our new engagement models gained traction, they would begin to expand our gross margin—and that is exactly what we are seeing. This quarter, our adjusted gross margin expanded sequentially from 30.6% in Q1 to 32.4%, an increase of 1.8 percentage points as those models lift the value we capture per engagement. That said, on a year-over-year basis, adjusted gross margin declined, driven by a foreign exchange headwind as our productivity gains offset the impact of the payroll tax resumption. Looking ahead, we expect gross margin to continue improving over the coming quarters as adoption of these models broadens across our book of business—a core driver of the profitability expansion we are working toward. Adjusted EBITDA was $19 million, with a 13.3% EBITDA margin. The year-over-year compression reflects two main factors. The first is the appreciation of the Brazilian real against the US dollar, which we have flagged before. On an FX-neutral basis, adjusted EBITDA would have been $20.8 million, a 15.6% margin, giving a clearer view of our underlying performance. The second factor is deliberate and reflects two distinct components related to our sales efforts. Part of it is a targeted investment specific to 2026, including scaling our Agentic SDLC initiative to capture the current acceleration in demand for AI deployment. This is not a permanent addition to our cost base. The other part is structural: an expansion of our commercial organization to support new offerings, practices, and vertical initiatives as well as commission expenses, which will remain part of how we go to market going forward. Together, these investments are funding the 40% pipeline expansion year over year, as Cesar mentioned, and the higher conversion rate already showing up in our top line. This is a conscious trade-off between near-term margin and durable, higher-quality growth. Importantly, the underlying trend is encouraging. Our adjusted gross margin expanded sequentially, showing that the pressure at the EBITDA level comes from our deliberate investments and external headwinds, not from our core delivery economics, which are in fact improving. Looking ahead, we expect our adjusted EBITDA margin to improve sequentially, while these investments position us to fuel growth into 2027 and beyond. Moving to our bottom line, adjusted profit was $8.7 million in the second quarter, with a 6.1% margin. This reflects the same two main factors I just described: the appreciation of the Brazilian real and our deliberate investment in growth. Adjusted diluted earnings per share was $0.07 versus $0.09 in the prior year. We see this as an investment cycle, not a new baseline. As these investments continue to fuel our growth into 2027 and beyond, we expect profitability to recover and our capital discipline to keep amplifying returns for shareholders. This quarter, debt discipline included repurchasing $2.8 million in shares, continuing our ongoing buyback program even as we invest in growth. Combined with these repurchases, our weighted average diluted share count is down 3.6% year over year, meaning each remaining share now carries a larger claim on our future earnings. I will now turn the call back to Cesar to discuss our business outlook and the strategic path forward for the remainder of 2026.
Thanks, Stanley. We continue to see an improving demand environment as enterprises increase their spending on AI deployment, and we are pleased with the evolution of our AI monetization efforts through new value-based commercial models. For the third quarter of 2026, we expect revenue of at least $145.7 million, a 14.4% increase over the third quarter of 2025 or 12.3% at constant currency. For the full year, we are raising our revenue guidance to the range of $566 million to $578 million, implying organic growth of 15.5% to 18%. Our revised outlook includes a positive FX impact of approximately 400 basis points. Alongside that, we now expect adjusted EBITDA for the full year in the range of 15% to 17%, reflecting deliberate investment in the commercial engine that drives demand and accelerates monetization, with sequential margin improvements through the second half as planned. This is a forward-leaning choice to move first on AI deployment and expand our wallet share. To be clear, this does not trade away financial discipline. Profitable, cash-generative growth is still the bar we hold ourselves to. What we are building is a company that scales revenue with less headcount and grows more profitable as the new commercial models mature. With that, we are ready to begin the Q&A session. Thank you.
分析師問答
Alright. We will now begin the Q&A session. I will announce each participant's name. Once you hear your name, please unmute your line and ask your question. Then when you are done, please mute your line. The first question comes from Puneet Jain from JPMorgan.
Hi, thanks for taking my question. I want to follow up on margin guidance. The cut of around 200 basis points: I understand a lot of it is discretionary investments and then currency. Can you break down that impact for us? How much of the incremental impact is FX versus investments, and why should we expect the level of investments to go down? We are still in very early stages of AI build-out and ramp-up—why will this level of investment not stay where it is into next year and beyond?
I may start here. Puneet, thanks for the question. Let's take the full picture. If you see seven quarters of double-digit growth, growing four times faster than our peers, we are gaining market share and wallet share, which means we are deepening relationships with our clients. We have broad-based growth across regions and segments, and everything is pure organic. Our pipeline is growing 40%. Everything is funded by the investment we have been making ahead of the pack and specifically in this second quarter we heavily invested. It is a response to the surge in demand for AI deployment as we reshape and redesign our go-to-market. Going forward, part of this investment will be reduced through the quarters—specifically the Agentic SDLC deployment element—but another part is structural and will remain. If you look at last year's quarter, sales were 8% of net revenue and this quarter sales were 12%, so going forward we expect to be more in between. This will be more than compensated by what you see in gross margin improvement as adoption of the new models broadens, and we will also get operating leverage on top of SG&A as a whole. As a consequence, EBITDA should improve over time. And all of this already accounts for the FX pressure we have highlighted.
Got it. That is helpful. On your top line, it seems like the financial services vertical is doing really well, including the top client that picked up nicely sequentially this quarter. How broad-based is that growth in financial services and what are your expectations for the rest of the year for that vertical?
You are right: financial services was one of our stronger verticals, growing 36% year over year. We grew across the board—retail 11%, tech and telco 68%, life sciences 16%. Only consumer goods was down, by 9%. Even if you exclude our top client from the top 10, the top 10 still grew 16% year over year. We continue to see expansion across all verticals and client cohorts. What is behind this is a solid increase in AI deployment demand and the investments we've made to capture that momentum. We've expanded our sales organization and reach across verticals to speed up growth, increase wallet share in our portfolio, and land new clients. Financial services will continue to be our number one vertical; the use cases for efficiency and customer experience in banking are very clear now in terms of impact. But we also see other verticals evolving—retail with agent commerce will be a big trend—and every industry will have powerful use cases to explore. We are preparing our offerings and teams to capture that.
Just to add: financial services grew 36% year over year while our top client grew 27%. So if you exclude the top client within that vertical, other clients grew faster than the headline number.
Great. That answers my question—thanks.
Thank you, Puneet. Our next question comes from Steven from Wedbush. Steven, please go ahead.
Alright, thanks guys. I want to start on the Agentic SDLC you pointed out in the quarter. Specifically about the pipeline you are seeing there: you mentioned a 40% year-over-year growth in the pipeline, but what percentage of that was specifically tied to Agentic SDLC? Does this carry any higher average deal sizes, and can you talk a little bit about the metrics there?
Sure. I can start and Bruno can jump in. Roughly 35% to 40% of our demand we articulate as Agentic SDLCs, especially transforming the current engagements we already have—old traditional digital engagements now being reshaped as Agentic SDLC engagements with different commercial models and a different margin profile. This is one offering where we are very competitive; I think we are five to ten times ahead of typical competitors. We have a lot of space for replacing underperforming competitors because we've invested heavily in CI&T Flow and in reskilling our teams over the last three years. I see the performance gap versus competitors increasing, which gives us a lot of room to replace and land on clients and also open new avenues of growth.
And the second question: specifically on deal sizes for Agentic SDLC versus your traditional deal sizes?
We are seeing an increase in deal size. Over the last two quarters the size of deals has been larger. We are not yet certain whether this is a long-term trend or a transitional momentum as engagements move from previous recurring models to Agentic SDLC. We disclosed that our pipeline is 40% larger year over year with a solid conversion rate. Deals are larger now, but whether it becomes a sustained long-term trend depends on how the market evolves.
Got it. One quick follow-up: on the geographic split moving forward, Latin America was another solid quarter of growth at 32%. North America decelerated sequentially from 16% in Q1 to 10% in Q2. Is there anything to point out there from a demand perspective? Any competitive displacement or something specific within North America?
No, that was seasonality, Steven. We can expect acceleration again through the year. Q2 had some seasonality—gaps in contract renewals and other ad hoc situations—but nothing systematic.
Thanks, Steven.
Thank you, Steven. Our next question comes from Bryan Bergin from TD Cowen. Bryan, please go ahead.
Hi, thanks. I wanted to ask on tech and telecom—really strong growth there in the quarter, the second consecutive strong number. First, is this driven by a handful of large transformation wins or broader demand across the client base? Maybe give more detail on what is driving it and how to think about it going through the second half.
Thanks, Bryan. For telco, we have some big telcos as clients and we're getting a lot of traction—especially around Agentic SDLC—and it resonates with the level of differentiation we can showcase in terms of productivity. We also see a new trend where tech companies are increasing their spending with us. This is correlated with AI deployment demand and with the partnerships we're announcing—not only the typical hyperscalers, but also new players like Anthropic. So it is a new revenue source and it's adding to our growth.
Understood. Quick follow-up on margin recovery path: to achieve the full-year EBITDA margin target it looks like you need to meaningfully improve from just over 14% in the first half. Can you categorize the biggest drivers of that improvement through the second half? How much comes from commercial model benefits versus moderating investment spend versus moderating FX headwinds?
Let me start. We have roughly a 2.2 percentage point headwind from FX compared to Q2 2025. That includes efficiency gains that compensate for the payroll tax resumption in Brazil, so you already see efficiency gains rolling in. Going forward, we expect more conversion of new models that have higher gross margins, which will drive improvement at the gross margin level. Then, on the P&L, we will see operating leverage in SG&A. In particular, you won't see another Agentic SDLC investment component as heavy as the one in Q2, which provides some alleviation. Of course, there is a structural sales component that will remain, but the combination of better gross margins and operating leverage should bring the sequential EBITDA improvement implied in our guidance.
To add color: the sales investment was seasonal in Q2 and we expect it to recede somewhat. Where sales went from roughly 8% to 12% of revenue this quarter, we anticipate a long-term normalization around 10%—so that peak should recede a bit and contribute to margin recovery.
Understood. Thank you.
Thank you, Bryan. Our next question comes from Maria Clara from Itaú. Maria Clara, please go ahead.
Hi everyone, thanks for the opportunity. I have two questions. First, more color on the increase in pipeline: which industries have been outperforming, and do you already see a trend of new clients gaining traction? Also comment on the evolution of the new monetization formats within those new potential deals. Second, on gross margin expansion: Stanley mentioned these new monetization formats are already helping margin expansion—can you elaborate on the profitability boost coming from those models and the long-run profitability expansion potential? Thank you.
Let me start with the second question. We are introducing four new value-based models: fixed price with higher margin, output-based throughput models, price-per-consumption, and outcome-based models. Depending on the mix, these models give us 3 to 15 percentage points higher contribution margin than traditional time-and-materials. That is a significant improvement. In the first half of 2026, 40% of new engagements were already based on these new models, and as we execute and renew contracts we expect gross margin to expand sequentially. On the pipeline and demand: we group demand under the AI deployment umbrella into three sets of offerings. First is AI adoption—particularly Agentic SDLC and software engineering productivity gains. Second is IP-based solutions, like our Modernization Studio, which streamlines conversion of legacy systems into modern AI-based architectures; the same applies for data modernization—preparing clients for AI by reducing fragmentation and building the right foundations. Third are industry-specific use cases for efficiency or customer experience—financial services, retail, etc. A new trend is enterprise reinvention services: transformational, outcome-based engagements where we help clients reinvent core processes around AI. That is increasingly relevant in our pipeline.
If I may add: we're excited about customer experience work. The first wave of AI demand focused on efficiencies and internal operational excellence. Now we're seeing the first large-scale programs geared toward consumers and customers of our clients, which is a completely new type of demand and we expect it to be exponential as clients grow confident in the models and results. That creates significant potential for growth and CI&T is well positioned to capture it.
Quick follow-up: do you foresee meaningful revenue potential from this next step of AI sponsorship tied to customer experience in 2027?
Yes, we believe the customer-experience area has a lot of exponential potential. As clients gain confidence in the results and expose these experiences to customers, that will create a new wave of demand, and we are very well positioned to participate.
Thank you. Very clear.
Thanks, Clara. Our next question comes from Gustavo Farias from UBS. Gustavo, please go ahead.
Two questions. First, on AI deployment demand: can you share color on how much of it is AI deployment per se versus legacy modernization required for AI deployment? And how much of the guidance raise was supported by the new partnerships with Anthropic and Mistral? Second, on the margin outlook: can you confirm if the Agentic SDLC investments are mostly concentrated in Q2, and how do you think of the structural expansion in the commercial department—interim or should we expect normalization next year?
Thanks, Gustavo. Roughly 30% of demand within our AI deployment umbrella is legacy modernization and data modernization—foundational investments companies need to fully explore AI's reinvention potential. On sales investments, we expanded from 8% to 12% of revenue this year, and we plan to stabilize around 10% next year. This will be more than offset by the new gross margin from the new offerings and commercial models. So we went from 8% to 12%; about half of that increase is a transition of current SDLC engagements to agentic models and new commercial models, and part of it is stronger global reach, new vertical geographies, and capabilities in sales. We will use gross margin improvement and dilution of G&A to scale CI&T profitably. In short, this is a sustainable long-term investment to accelerate growth, increase wallet and market share, and scale more profitably as the new commercial models mature. Regarding Anthropic and Mistral, these partnerships expand our technical options and go-to-market motion, helping us win accounts that require frontier models and strong deployment capabilities.
Very clear, thank you Cesar.
Thanks, Gustavo. Our next question comes from Luke Morrison from Canaccord. Luke, please go ahead.
Hey team, congrats on the quarter. Quick question on consumption pricing and the new pricing models, like the agent computing unit model: as underlying costs for running models fall and token costs fall, you're billing on consumption—how do you prevent that becoming a deflationary force and shrinking revenue over time?
That is not a simple question. Cost per token is indeed decreasing, but model capabilities and usage are increasing, so consumption can rise even as token cost per unit falls. We also have options like open-weight models through partnerships such as Mistral, which give us different cost-performance trade-offs. Importantly, price-per-consumption will be one element of a portfolio of commercial models; it won't be the majority. We combine models per client and engagement. We're not betting everything on a single consumption model. It's a long and complex game of adjusting pricing and cost structure, but consumption pricing is only a part of our overall approach.
Fair enough. Also, you have partnered with multiple frontier model providers—how should we think about being model-agnostic versus going deep with one provider, and to what extent do your clients care what sits underneath CI&T Flow?
They care. Clients are sensitive about privacy and data controls, so Flow is designed to be agnostic: Flow is connected to more than 37 models and controls that complexity for clients. We must be model-agnostic and multi-model because we serve clients across geographies and sectors with different footprints and requirements. We help clients choose what is best for certain tasks, manage FinOps, and automate different types of workstreams. That knowledge is something we've built over nearly four years with Flow, and it drives a lot of the conversations and deal flow we are seeing.
Thank you. Our next question comes from Cesar Medina from Morgan Stanley. Cesar Medina, please go ahead.
Hi, thanks for taking my questions and congrats on the results. Can you confirm that 40% of your revenues in the first half are linked to these new pricing mechanisms? If correct, how much of this is in the pipeline?
Yes: 40% of the new bookings in the first half were under the new commercial models. We have many long-term contracts that will be converted incrementally to the new models as they renew, so I estimate it will take roughly 18 months to have the majority of our book re-priced under the new models as renewals and new sales combine. We don't expect 100% of revenue to be one single model—time and materials still fits certain types of work, like forward deployment engineering by design—so the mix will remain diversified. But we expect to move to a higher contribution margin profile across our book over the next 18 months under current market conditions and if we continue converting legacy engagements into these new models.
Understood, thank you and congrats again.
That concludes our Q&A session. Thank you all for attending our event today. I will now invite Cesar to proceed with his closing remarks.
Sure. Thanks, Bruno, Stanley, Eduardo. Thank you all for joining us today. I want to thank our teams around the world for their hard work and dedication—I am glad you are seeing our transformation happening. And, of course, a special thank you to our clients for trusting CI&T as their AI deployment and innovation partner. That is it. We will see you soon. Stay well. Bye.