Contact

AI

Could AI push aside Salesforce?

by Leave a Comment

Customer Relationship Management (CRM) software is nearly a $119 billion global market today, and Salesforce is the undisputed leader in the category today. But just as Salesforce once disrupted the likes of Oracle and Siebel, AI is now ushering in new players and new rules of engagement that are decisively shifting the battleground. 

Data has become ground zero for AI, and customer data is among the most valuable data out there given its proprietary nature. This makes CRM a natural place for AI to play.  As a result, what used to be a competition over features and market share is now a fight over something much deeper: access to that data and the intelligence, revenue and differentiation that can be derived from it.  

The market for CRM software has become increasingly competitive, but there’s a broader question on people’s mind these days: Will AI interfaces and agents render the CRM software category irrelevant? Will companies even need a CRM at all if they can move to an AI user interface on top of some structured data source?  Will Salesforce get their hat handed to them by their AI-native competitors?

Our take? CRM isn’t going anywhere. And it would be a mistake to underestimate Salesforce. 

However, the competition for incumbent CRM players is going to get more and more intense, and the competition and evolving buyer expectations are going to force these vendors and their service partners to evolve fast.

Rise of the AI-native CRM players

Bessemer’s State of AI 2025 report highlights the shift that’s happening from systems-of-record to systems-of-action, calling it a once-in-a-generation shift. It’s no longer enough to just store customer data and function as a “single source of truth.” Modern platforms are increasingly expected to allow both humans and AI agents to act on that data — proactively, intelligently and securely. 

Many experts believe CRM will look more like an invisible orchestration layer, updating itself automatically, interpreting customer signals across channels, and recommending (or taking) the next-best action before a rep can even open a laptop. With tools like this, productivity could skyrocket, and customer-facing teams can focus on higher value interactions.

It’s already starting to happen as AI-native CRMs introduce faster, cheaper, and more agile workflows: 

  • Sierra, co-founded by former Salesforce co-CEO Bret Taylor, helps companies automate customer support across chat, email and voice while keeping humans in the loop. 
  • Aurasell markets itself as an AI-native GTM platform that aims to collapse different tools such as data enrichment, forecasting and coach-agents into a single intelligent platform.
  • Day AI bills itself as a system of record that reasons and a GTM chief of staff all in one, promising to eliminate manual entry and maintain pipeline in real-time.
  • Attio positions itself as an AI CRM and system of action for modern GTM teams that is highly customizable to specific types of businesses.

And by this time next year, there will likely be far more players on this list.

At the same time, a growing class of AI-forward vendors are building intelligent layers that sit on top of core platforms like Salesforce. They’re not trying to replace the system of record, but rather to extend or automate parts of the CRM experience where the core platform may lag. These solutions tend to do things like automate requirements, setup and configuration (Auctor, Sweep, Swantide), provide process and organizational intelligence (Hubbl), enforce data hygiene and governance (Clarify.ai), or accelerate GTM operations (Vellora). 

Less competitors and more partners, these solutions will help make CRMs stickier, more intelligent, and better aligned with business needs. 

Existing platforms expand 

It isn’t just the AI-natives that are putting pressure on the incumbents. Vendors that traditionally focused down-market or in areas adjacent to sales and marketing (e.g. commerce or business operations) are also capitalizing on the AI disruption and buyer skepticism, and making moves into Salesforce’s territory. 

SMB players, most notably Hubspot, are beginning to take share in the enterprise by promising customers simplicity, affordability and AI-driven productivity. 

Platform players like Shopify and ServiceNow are expanding into CRM territory by building out their platforms and layering in AI-driven workflows. Shopify, for example, has evolved from a pure-play e-commerce solution into a holistic commerce platform that incorporates crucial aspects of CRM such as customer segmentation, personalized marketing, and post-purchase engagement.

ServiceNow, traditionally a leader in other internal workflows like IT and HR, is aggressively moving into the space with acquisitions like Logik.ai and Moveworks, and they are making a concerted effort to build out their partner ecosystem in this area.

This convergence means that the CRM wars are no longer confined to the historical rivalry between established CRM vendors. They now involve broader platform battles where the winners will be determined by the ability to offer the most unified, intelligent, and workflow-centric customer engagement solutions.

The dark horse: The AI platforms

And then there are the broader AI platforms that are quickly making their way into nearly every business process, including CRM.

OpenAI has trained millions of users to treat ChatGPT like a universal interface. We are quickly being trained to see conversational AI as the default starting point for information retrieval, content generation and task completion, and agents (autonomous AI) may soon automate a good portion of our day-to-day workflows.

Anthropic’s introduction of Claude Cowork, an agentic product that plugs into existing tools and platforms to complete multi-step business tasks, triggered a SaaS selloff in recent months as investors worried about the application layer becoming commoditized. And while we believe the ‘SaaSpocolypse’ is overblown – especially when you look at the continued growth for many of these SaaS companies – it’s clear that LLMs are starting to shift user behaviors and how enterprises think about internal processes. 

As these models and workflows become more sophisticated, it stands to reason that many business users will prefer to transact directly inside an AI chat interface without bothering to log into CRM or ERP at all. However, anyone who thinks that this is the death knoll for a category, and that platforms as deeply entrenched as the likes of Salesforce or SAP will go away anytime soon, likely hasn’t worked inside a large, complex enterprise recently.  These vendors are deeply entrenched in business processes and change happens slower than most pundits predict.

Salesforce is under pressure but not out of the game

Salesforce is most certainly under pressure from all sides, but don’t count them out. The CRM giant has reinvented itself many times before and we have confidence they’ll do it again.

Salesforce’s annual R&D investment stands between $5-6 billion and it’s been increasing year-over-year. That spend is primarily going toward rebuilding itself around AI. They are also more focused than ever on profitability, giving them plenty to work with when it comes to acquiring AI-native and foundational capabilities. Their M&A machine has kicked back in aggressively with 12 acquisitions over the last 18 months.

Salesforce’s vision for the Agentic Enterprise, its moves with AgentForce, and its relationships with senior buyers, as well as the biggest AI and data vendors, are keeping Salesforce at the forefront of AI conversations with customers. But two other areas give Salesforce a very compelling moat: it’s extensive service partner ecosystem and the vast amount of customer data they sit on.

As long as they protect those two assets, and use their immense resources to modernize their platform, we have no doubt Salesforce will continue to be a trusted vendor in the AI era.

 

The battle moves to the experience layer

Ultimately though, the winners in CRM will be those that leverage their strengths to stay innovative at the experience layer, where customers and employees actually interact with the system. That’s where AI’s value becomes most visible.  A sales rep doesn’t care if their CRM has the cleanest schema in the world; they care if it helps them close deals faster.  A customer cares less about the underlying data model than they do that their service request is resolved in seconds. 

This is why agentic AI and orchestration tools are so disruptive. They shift the focus from records to results, from storing data to delivering outcomes in the exact moment of interaction. Which is why the winners will be defined not by who owns the record, but by who can orchestrate the action that makes the experience a success. 

For services firms, that requires having a deep understanding of not just tech platforms and how they work, but also how data flows and agents work across systems and within processes. It means having the ability to build experiences around platforms, which requires industry expertise and deep domain knowledge.  All of which are traits that we look for in our future investments.

Read our Tercera 30 report to learn about more trends impacting software vendors and their partner ecosystems.

 

How Oracle and SAP Became Cool Again

by Leave a Comment

Over the past 50 years, Oracle and SAP have evolved from emerging innovators into foundational pillars of enterprise IT, building global empires that power many of the world’s largest organizations. These two legacy tech vendors may lack the flash of newer entrants, but they have steadily expanded their platforms and partner ecosystems over the years, creating what many might consider dull but durable growth in the process. It’s this durable growth that has kept Oracle and SAP firmly anchored in the Tercera 30 since its launch. 

However, these software giants are anything but dull these days. Both are capitalizing on their scale and footprint among large enterprises to capture opportunities in the cloud and AI, delivering a performance jolt not only to the company, but also to their ecosystem of tech services partners. 

In this article, we explain why these ecosystems are seeing a resurgence in the AI era, and why entrepreneurs and investors like Tercera are taking a closer look at these industry titans. 

THE RETURN OF THE TITANS

Oracle and SAP’s rebirth is, in part, a matter of sunsetting the past. Both companies are phasing out support for their legacy on-premises solutions, pushing and pulling customers onto their more modern cloud platforms. 

Oracle and SAP’s rebirth is, in part, a matter of sunsetting the past. Both companies are phasing out support for their legacy on-premises solutions, pushing and pulling customers onto their more modern cloud platforms. 

For example, SAP is pushing all customers of its core on-premise ERP system, ECC, to its newer S/4HANA platform, setting a deadline of Dec 31, 2027 to create urgency. After Dec 31, it will no longer maintain support for the product. Oracle is pursuing a similar strategy with its legacy products, although has been less aggressive about ending support. All this is creating a wave of demand for large and lucrative enterprise migrations who depend on these systems.

Both companies are leaning on their scale, footprint in enterprise IT and product breadth to move forward. Oracle has inked multibillion-dollar contracts with OpenAI, Microsoft, Facebook, and others to provide the cloud infrastructure required to train and run the world’s largest AI models at scale. It is also using its strengths in data, high-performance hybrid cloud infrastructure, and industries to land large new enterprise customers and expand its existing footprint. In its Q2’25 quarterly earnings call, Oracle surprised analysts by announcing four new multibillion-dollar contracts, which increased its remaining performance obligation (RPO) contract backlog by 359% to $455 billion. 

The stickiness of SAP’s ERP platform and its integration into mission-critical processes is giving the company a leg up in the AI game. The company has made a massive investment in modernizing its Business Technology Platform (SAP BTS) and core applications, and is building out AI capabilities on top of BTS with its GenAI tool Joule and a growing number of functional and industry-specific agents. SAP says they now have 400 AI scenarios enabled by Joule and more than 34,000 customers using its AI capabilities. 

Oracle and SAP may have older platforms but that also gives them a competitive moat. Their deep customer relationships, integration into core business processes, substantial R&D budgets, and their deep seated regulatory expertise are not easy to displace, especially for smaller, less-seasoned competitors. 

Their relationships with other AI and data leaders also adds to their moat. Oracle’s alliances with NVIDIA, OpenAI, Google and Meta mean that enterprises can run most large language models and agentic AI workloads directly on Oracle’s cloud infrastructure. SAP’s partnerships with Microsoft Copilot and OpenAI allow customers to embed its LLMs into applications via SAP Business Technology Platform. SAP has also deepened its partnership with Databricks, making Databricks a native component of SAP’s Business Data Cloud. 

Partnerships with AI-native companies, and AI-native services partners, are essential as both companies reposition and reinvent themselves for the AI era.

These partnerships with AI-native companies, and AI-native services partners, are essential as the companies reposition and reinvent themselves for the AI era. They carry brand baggage as legacy players, which newer entrants will continue to exploit. We’re already seeing cloud-native startups attempting to chip away at their markets. Yet these behemoths are not going to go quietly and their moat is real. 

THE OPPORTUNITY FOR TECH SERVICES FIRMS

The biggest near-term opportunity for services partners is with migrations. The transition from SAP and Oracle’s legacy on-premises systems to cloud platforms is far more than a mere software upgrade. These are multi-year, multi-million dollar transformations that can ripple across every division and process a company runs. Doing this successfully requires untangling and reconnecting integrations that are deeply embedded into mission-critical systems. It also requires understanding and optimizing data across customized business processes, and orchestrating change across global teams.

The biggest near-term opportunity for services partners is with migrations. The transition from SAP and Oracle’s legacy on-premises systems to cloud platforms is far more than a mere software upgrade.

ERP systems are deeply embedded into business operations, and complex to migrate. Which is why a large majority of ECC clients have not yet migrated to S/4HANA. Yet the pressure is mounting. Organizations that don’t move onto these more modern platforms are at risk if vendors discontinue maintenance for their systems. They also won’t be able to take advantage of new tools and AI advancements, potentially putting them behind the curve.  

SAP and Oracle do not have sufficient in-house resources to support all these migrations. We have heard from partners that SAP, for one, has raised its bar for the clients it directly supports from $1B to $3B in customer revenue, pushing smaller and midsize enterprises into the open arms of IT services firms. 

For IT services firms, the opportunity presents a windfall of high-value work that could last a decade or more. Some work will extend well beyond the formal sunset dates. Helping organizations move to cloud and AI-enabled systems is a sizable journey, not a single event.

For IT services firms, the opportunity presents a windfall of high-value work that could last a decade or more.

Implementation revenue can spike in the initial 18 months, but the need for continuous improvement, support and training could provide a sizable recurring revenue stream.  How significant is the opportunity? I’ve met several companies that say they can generate as much revenue from <20 clients on SAP’s S/4HANA as from 250 still on legacy systems.

The AI opportunity extends far beyond migrations and implementations. Today, Oracle and SAP applications collectively house much of the world’s most valuable private enterprise data. This is decades of structured, governed data spanning finance, supply chains, manufacturing, HR, customer behavior, and regulatory operations. 

The data in these systems is not only high-quality and deeply contextual, but also uniquely tied to real business processes, decisions, and outcomes. As AI shifts from generic models to domain-specific intelligence, these systems become a true treasure trove.

The data in these systems is not only high-quality and deeply contextual, but also uniquely tied to real business processes, decisions, and outcomes. As AI shifts from generic models to domain-specific intelligence, these systems become a true treasure trove. The foundation for training and deploying AI agents that can reason, predict and act within specific industries and workflows. 

FROM UNCOOL TO UNMISSABLE

Coolness may be in the eye of the beholder, but power is measurable. Oracle and SAP have made deliberate bets that place them at the gravitational center of the AI era, arguably the most consequential enterprise technology shift of the decade. This should be a tailwind for the SAP and Oracle partners that can make the evolution alongside these vendors, and an opportunity for native AI firms looking for a place to play.

This should be a tailwind for the SAP and Oracle partners that can make the evolution alongside these vendors, and an opportunity for native AI firms looking for a place to play.

Both SAP and Oracle already have large, highly mature service partner ecosystems, yet the number of partners in both ecosystems is still growing – a testament to the opportunities ahead. 

If you are a partner in one of these ecosystems, looking for growth capital to tackle this moment of opportunity, we’d love to hear from you here. 

Read more about SAP and Oracle in the Tercera 30

Tercera 30

Forecast vs. Reality: 2025 Predictions in Review

by Leave a Comment

As we do every year, we started out 2025 with a set of predictions for the tech services market for the 12 months ahead. Now, with the end of the year just a few weeks away, it’s time to review just how right (or wrong) we were. 

We’re proud to say this year’s forecasts were our most accurate to date. With more predictions than usual (10 in total) we scored almost all As and Bs with two Cs in the mix. Not bad, but it’s a sign to push ourselves to be bolder next year. And maybe do fewer predictions.

Here’s a closer look at the 10 IT services trends we said would shape the year, what we thought would happen, and what actually did. 

Prediction 1: Pockets of growth return

Grade: A-

The services sector did see growth in 2025, but not uniformly. According to Gartner, year-over-year (YoY) growth in worldwide IT services spending remained well below the highs of 2021 and 2022, and is expected to land around 6.5% YoY for 2025. Certain segments such as data and AI grew faster, while other segments like commerce and digital engineering services saw greater headwinds.  

The software sector in general performed well. The median growth rate of all the publicly traded software companies we track in the Tercera 30 was around 17%. Of course some companies performed better than others. The hyperscalers (Microsoft, AWS, Google, Oracle) reaccelerated, while data players like Snowflake, Databricks, MongoDB, Rubrik outperformed as well. The foundational AI platforms (Anthropic and OpenAI) saw jaw dropping growth, while continuing to raise capital to advance their models, build out offerings and invest more deeply in infrastructure and their partner ecosystems. All of which provided opportunities for services firms who have figured out how to play here.

We predicted tailwinds in industries such as insurance and healthcare. According to Gartner, IT service spending in insurance in 2025 is projected to come in at 8.1% YoY growth in 2025, exceeding broader IT service spending estimates. We’re seeing many AI native software and services firms leaning into this vertical, putting pressure on incumbent vendors like Guidewire and Sapiens, although Guidewire is showing some acceleration as of late.

IT service spending in Healthcare and Life Sciences is projected to grow closer to 6% YoY.  The healthcare sector in the U.S. saw more headwinds than expected, with tariffs impacting the supply of medical equipment and pharmaceuticals, along with policy changes and funding reductions to programs like Medicaid. 

Prediction 2:  Game time for GenAI: production projects double

Grade: B

We were directionally right that 2025 would be a breakout year for GenAI moving beyond pilot mode. However, the picture was more nuanced than a clean 2x conversion rate. Production deployments expanded meaningfully, yet the actual rate of pilot-to-production success remains hard to pin down and, in some cases, hotly debated.

A widely referenced MIT study suggests that only around 5% of GenAI projects were considered successful, while a McKinsey 2025 state of AI survey found 88% of companies are now using AI in at least one function. A recent study from Wharton and GBK Collective added more optimism to the mix, finding that three-fourths of enterprises report positive ROI on their GenAI investments. Most of the respondents in the Wharton study (88%) say they’ll increase spending in the next 12 months, which is perhaps the biggest sign that companies are beginning to see results.

 

In 2025, the AI narrative pivoted rapidly to agents. Every software vendor in our 2025 Tercera 30 has announced investment or offerings here, and it’s on the agenda for their customers and partners as well. A Tercera survey of 320+ services leaders this summer indicates that more than a quarter of IT services firms now have clients with AI agents in production. 

 

Prediction 3: AI-native consultancies challenge public firms

Grade: A

This one proved spot on. AI-native consultancies are starting to materially challenge larger public firms, not just on cost and speed, but increasingly on specialization and IP-led delivery. 

Press coverage throughout the year spotlighted emerging AI boutiques winning deals from the likes of McKinsey, Bain and the Big Four. AI-native players aren’t just putting pressure on the GSIs, but targeting the mid-market as well. Along with the boutique firms that have specialized there.

What we didn’t anticipate was how quickly the big consultancies would counter, leaning hard into a build–buy–partner playbook to keep pace.  All of the GSIs are pushing hard to retool and build for the AI era.  

Some did this through M&A. For example, IBM this year acquired Hakkoda and Datastax, buying AI capability to complement their existing organic efforts. 

Many are leaning into build mode. Genpact’s Chief Agentic AI Officer, Jinsook Han, spoke at Tercera’s annual Leadership Retreat about the company’s GenpactNext strategy and how its Advanced Tech Solutions group are putting verticalized AI agents to work. 

Others are leaning more into partnering. For example, we’re seeing both large and boutique consulting firms choosing to partner with AI-native Services-as-Software platforms like Auctor AI and Bridgetown Research to accelerate workflow and expand capabilities.

Whether the strategy is to build, buy or partner (or increasingly a combination of all three), it’s clear that AI is forcing firms of every size to adapt.

Prediction 4: Focus shifts from hours to outcomes 

Grade: C+

We were a bit too optimistic here. While firms did make progress experimenting with outcome-based pricing models, most customers and firms remain firmly attached to the by hour, time-and-materials model. Procurement teams for the most part still default to traditional structures, and sales and delivery teams are still learning how to scope, deliver and measure true outcome-based work. 

Many providers are experimenting with new pricing models, finding different ways to capture value from their AI investments and adapt to competitive pressure in the sales cycle. Experiments range from billing for agents or ‘digital FTEs’ based on the value they deliver, to carving out pieces of a project that can be delivered as a fixed fee with a bonus if certain metrics are hit, to true outcome-based engagements.  

There are signs the larger firms are making progress here. McKinsey claims that a quarter of its global fees are now outcome based, and Globant’s subscription-based, tokenized AI pods that move from hourly billing to AI-metered usage.

As early experiments play out, and AI moves deeper into production and operations, pricing models will shift. However, it’s going to take longer than many of the headlines suggest. 

Prediction 5: The HAIDaMo mafia tightens its grip

Grade: A+

This year the hyperscalers, AI providers, and data modernization (HAIDaMo) players reinforced their position at the center of the modern enterprise IT stack. 

Of the publicly traded companies we track in the Tercera 30, the players in this category generated 78% of the total revenue growth. This excludes high-flying private companies like OpenAI, Anthropic and Databricks, but it does include Nvidia which continues to grow at an unprecedented rate. 

The big 3 hyperscalers (AWS, Microsoft Azure, Google Cloud) continued to consolidate wallet share, now holding 62% of the cloud market.  In Q3, AWS revenue grew 20%+, and is now at a $132 billion annual run rate (yes, billion). Microsoft’s Intelligent Cloud group, which houses Azure, grew 28%+ at an annual run rate of $123 billion. Google’s cloud group grew 34% in Q3 with an annual run rate of nearly $61 billion, while its newest Gemini release is setting new benchmarks for the competition.

It’s not just the hyperscalers. IBM doubled down on AI, hybrid cloud, and quantum computing, and its recently announced Confluent acquisition is poised to further strengthen its capabilities here. Oracle reasserted itself, signaling to investors that it expects its cloud business to grow more than 40% in its next fiscal year while deepening co-sell arrangements that make its database accessible across all major clouds.

Snowflake and Databricks posted strong performance for the year providing tailwinds to partners, while the big names in AI (namely Anthropic, OpenAI and Palantir) remained gravitational forces. 

All this level of momentum, and the unprecedented capital still pouring into this space, means these ecosystems are only getting more and more competitive. Partner programs here are tightening, expectations for technical depth are rising, and the barrier to meaningful differentiation is getting higher. 

Prediction 6:  Partners trade monogamy for polyamory

Grade: A

As expected, in 2025 services firms continued to expand beyond single-ecosystem strategies, with diversification becoming the norm rather than the exception. Relying on one ISV to scale to $50 million is increasingly rare for an IT services firm, and most now recognize the need to work across multiple ecosystems to fully address client needs. This is especially true with AI and agentic processes which often span different systems.

Tercera survey data backs this up. In a 2024 survey of IT service leaders, firms reported maintaining 3-6 partner relationships, with larger firms naturally carrying wider portfolios. In our 2025 survey, 70% said they had grown their partnerships over the past 12 months — 55% slightly and 15% significantly. High-growth firms pushed even harder: 82% expanded their partner ecosystems, and they were over two times more likely to grow partnerships “significantly.”

We also saw firms deliberately branching into adjacent partner ecosystems, and beginning to experiment with an expanding universe of AI and agentic partners. This is a space that’s both full of opportunity and notoriously difficult to navigate. The shift toward an “ecosystem of ecosystems” is no longer a conceptual ideal, it’s becoming table stakes.

Prediction 7: The return of on-prem priorities

Grade: B

With enterprise buyers prioritizing AI strategies and holding off on large transformational projects, we expected a slight downtick in cloud migrations and an uptick in on-premise and data center investments.

Investment in these areas did surge. According to the Gartner report mentioned above, spending on data center systems is expected to grow a whopping 46.8% this year.  Much of that is driven by the AI juggernauts. However, enterprises are also investing here as they lean into their own AI capabilities, and manage more data-intensive applications that have certain bandwidth and latency requirements and bring compliance complexities.

Yet, cloud migrations marched on. More than 90% of enterprises now have some workloads in the public cloud, and as the growth of the hyperscalers suggest, that will only increase. 

What really defined 2025 is hybrid computing – not a pure on-prem return. There is a growing desire among enterprises to mix infrastructure types based on specific use cases and to increase flexibility and resiliency. 

What really defined 2025 is hybrid computing – not a pure on-prem return. There is a growing desire among enterprises to mix infrastructure types based on specific use cases and to increase flexibility and resiliency.  All of this benefits vendors and partners with the capabilities to work across diverse, complex enterprise architectures.

Prediction 8: Employers retain control

Grade: A

Employers (mostly) stayed lean through 2025, prioritizing AI investments, employee productivity and increasing margins over pure growth. This has kept a firm grip on wages, and meant power remains in the employer’s court.

According to the National Center for the Middle Market, business services firms grew employment 8.3% YoY (as of July 2025), several points below recent years. Layoffs across tech continued in smaller, serial waves: Challenger reported 71,321 planned cuts in November 2025 and 1.17 million year-to-date, the highest since 2020. Hiring plans sank to their lowest level since 2010. All of which reinforce employers’ leverage.

However, within the AI sphere, the market played by a different set of talent compensation rules. AI engineers, ML specialists, and other top talent commanded outsized pay. Well-funded players battled for scarce expertise. While for most roles, employers held the upper hand, for elite AI talent, employees could name their price. 

Return-to-office mandates strengthened as we predicted, especially in fast-moving AI environments where in-person collaboration and a “9-9-6” pace (9 am to 9 pm, six days a week) became the norm and even a badge of honor. 

Prediction 9: Africa on the rise

Grade: C

We were directionally right on this one, but other regions saw greater investment this year and large scale tech services investment in Africa is unfolding slowly.

Africa continues to gain traction as a source for talent, especially in Kenya and South Africa. Many African countries are seeing positive signs of investment. For example, Kenya’s investment authority projects the country’s Global Business Services sector will reach $700 million by 2025 and climb toward $1 billion by 2030. South Africa is also strengthening its position, with its GBS workforce more than doubling since 2019, and revenues nearly tripling through 2024.

Talent supply continued to expand as well. Andela, the Linux Foundation Education, and the Cloud Native Computing Foundation (CNCF) collaborated to launch the Kubernetes African Developer Training Program in 2025 to train tens of thousands of African technologists in cloud and AI-adjacent skills. Vendors like ServiceNow are also investing in training and supporting partners in the region. All of this is part of a broader wave of AI-native talent entering the workforce thanks to the continent’s relatively young population. 

Still, constraints remain. Complex engineering is still maturing in the region. While the pool of junior talent is large, there are fewer mid-level and senior specialists available than in some other regions. And while cost advantages exist, they’re not as dramatic as some buyers might anticipate. 

In contrast, other regions like India, are seeing greater investments by big tech players. Amazon recently said it would invest $35 billion into India between now and 2030, and Microsoft said recently they are committing more than $17 billion to the region. 

India is seeing greater investments by big tech players. Amazon recently said it would invest $35 billion into India between now and 2030, and Microsoft said recently they are committing more than $17 billion to the region. 

Africa is still well positioned to be a talent destination for software and services firms in the future. However, 2025 was just not the breakout year we predicted. 

Prediction 10:  All eyes on the Trump Administration

Grade: A

Retrospectively, this one was too much of a layup. Of course all eyes are going to be on the Trump Administration. No matter what side of the aisle you are one, this Administration is hard to pull your eyes away from.

Here’s what we know.  As expected, the Administration did freeze spending, cut programs and implement tariffs. However, it happened far faster (and far more unpredictably) than we expected. Our view was that the government is vast and complex, and change would take time. But less than a year into the Administration, the federal government has already undergone significant change, with certainly more to come.  

Another part of our prediction was that federal spending would remain lucrative for software and services companies as the administration invested in automation and commercial software to create efficiencies. That one held true, with defense, cybersecurity and AI acting as counterweights to softness in certain civilian agencies. 

The Federal IT Dashboard shows $102.3 billion in requested FY2025 IT spending — a 10.8% increase over the $92 billion request in FY2024. Historically, actual spending exceeds requests, so the final tally is likely to be even higher.  

Those servicing the defense sector saw spending increase, with Palantir emerging as a clear winner, reporting 52% YOY growth in U.S. government revenue in Q3. Vendors like Salesforce and ServiceNow leaned in hard with efficiency messaging, and price cuts to stay entrenched.  Yet some vendors saw a target on their back. 

The administration’s moves around tariffs, immigration, and its unfortunate crusade against DEI had significant impacts well beyond just spending. Retailers and construction have had to rethink supply chains and absorb costs related to tariffs; education and healthcare institutions that are heavily dependent on government grants have had to reconsider investments; and services firms are keeping a very careful eye on the HIRE Act which could impose new taxes on offshore payments. 

Not quite honor-roll status, but comfortably above average

Our forecasts were more on the mark than previous years and that’s saying something considering the pace of change. With just a solid 3.4 GPA, we’re definitely on the merit roll.

Stay tuned for our thoughts on what services leaders and investors should pay attention to in 2026. If you’d like to be among the first to review our prognostications, subscribe to our monthly newsletter or weigh-in with your own predictions by hitting us on LinkedIn.

The 2025-2026 Tercera 30

by Leave a Comment

The Top Software Ecosystems for Tech Services Partners

We know from experience that not all software partners are created equal, and that choosing the right one can make all the difference. 

The Tercera 30 profiles some of the most entrenched and promising software partner ecosystems set to thrive in an AI-enabled future. Each of the names on this year’s list is an innovator at the core, and depends on service and solution providers as part of their growth strategy.

Enter your information to access the full report, including vendor profiles and key trends shaping the enterprise software space.

Building the Tercera 30: Where AI Helped and Hindered

by Leave a Comment

At Tercera, we spend considerable time advising business leaders on how AI is impacting the way work is delivered. So we are always looking for ways to take our own medicine and responsibly weave AI into our own processes.

When it came time to develop this year’s annual Tercera 30 report, we knew this was the perfect opportunity to leverage AI in a purposeful way. Over the last four years, we’ve built up a significant amount of data on software vendor ecosystems. This, along with our deep connections with software channel partners, bankers, analysts, and service providers in these ecosystems, gives us a lot to work with. However, we are still a small team without a dedicated research arm. 

We started by looking at our existing processes and where we could leverage the foundational enterprise AI tools we already use (e.g., ChatGPT, Claude) and the AI capabilities embedded within our existing productivity tools (e.g., Google, Box, HubSpot, and Microsoft). These processes include: data collection, data validation, report development, and marketing and promotions.

From there, we looked at how we could use fit-for-purpose AI platforms and agents to do things that would’ve been difficult or impossible to do without giant budgets, or an army of analysts. This included interviewing or doing deep research on smaller, less established software vendors where we may not yet have direct relationships, or getting direct insights from senior enterprise IT leaders to understand how they’re buying, building or adopting AI platforms. 

The two tools we ended up leaning on the most were ChatGPT Deep Research and Bridgetown Research’s agentic research platform. 

ChatGPT: The Swiss Army Knife for Research 

ChatGPT proved most helpful at filling in basic data in our repository of company information. In batches, we were able to create tables including information such as website, company description, public/private, etc. formatted in a way we could easily import into our existing data set. Of course, there were more than one instances of hallucination or incorrect information, so it still required a fair amount of double-checking data against past data we had collected and other sources. This should require less training in the future based on this year’s usage. We found consistent value in ChatGPT’s Deep Research function to develop a base-level understanding of corners of the market that we were unsure about. 

Bridgetown Research: Agents to Go Wider, Deeper, Faster

We were first introduced to Bridgetown Research through one of our Advisors, and were immediately impressed by their vision to transform how primary and secondary data are gathered and analyzed, achieving scale, speed, and cost efficiencies that were previously difficult or nearly impossible. At its core, the Bridgetown platform uses autonomous AI agents to conduct expert interviews, collect data, and run custom analyses. It’s still early days, and we only scratched the surface of what Bridgetown’s platform can do, but we were impressed enough that we are considering incorporating it deeper into our AI-stack for value creation advisory work. 

Lessons Learned for Using AI in Research

When it comes to incorporating AI deeper into our processes, we learned first-hand just how important it is to experiment. Some of the tools and ways we leveraged AI in this process exceeded expectations. At times, it was more trouble than it was worth. 

It reinforced to us that AI can be extremely powerful as part of the research process. However, its impact is best felt when applied not generally, but purposefully. 

Here are 5 other lessons we learned along the way.

1. ChatGPT is a utility player, but not the best tool for everything

General-purpose GPTs are table stakes for the modern worker. Like Word or Excel, they are a base tool you should reach for by default to draft, summarize, run first-pass research, and structure messy information. In our process, ChatGPT sped up recurring tasks like populating company metadata into import-ready tables and spinning up quick briefs so we could engage experts faster. But just as a growing company eventually moves from tracking pipeline in a spreadsheet to implementing a CRM, you’ll eventually outgrow what a generalist GPT can do on its own. 

The real transformational change in our research process came from pairing a generalist GPT with a purpose-built tool.

Bridgetown Research’s platform expanded our reach by using agents to coordinate expert and buyer interviews and generating targeted insights we could not reliably get from a generalist model alone. Its “Bridget” assistant answered research questions more effectively because it was tuned to market-research workflows and grounded in the data we supplied. 

As scope and complexity increase, you need tools that are designed for the specific job.The model you choose matters: different GPTs handle retrieval, structure, and citation differently, and task–tool fit drives quality. 

2. There is still a role for traditional research tools

AI helped us move faster and cover more ground, but it did not replace the basics. We still leaned on the traditional research and analysis tools. For example, we commissioned a third-party research firm to survey 321 current services leaders to gauge how teams were actually implementing AI and thinking of partnership strategies. 

We also built a straightforward scoring model to evaluate ISVs on service potential and growth. The weightings and KPIs came from proven ways of consolidating and analyzing data, which kept the results anchored in reality rather than a black box. Then came the hands-on work that turns data into insight. We used Excel to clean, weight, and combine variables, and to turn the results into a clear 2×2 grid that made the tradeoffs easy to see. 

Just as important, we talked to people. Interviews and live conversations with services leaders, ISV channel leaders, and analysts gave us context, helped validate surprising findings, and turned the numbers into practical guidance. AI gave us speed and reach. Traditional research tools and real conversations made the output trustworthy and useful. 

3. Challenge the output at each step 

Reviewing any AI output is essential, but it’s just as important to understand the inputs and sources from which the AI tools are pulling. In more than a few instances, we found data was wrong because it was pulled from a source that was wrong and not reliable. A prime example of this was a higher-than-expected rate of misidentifying whether or not companies were publicly traded. Double-checking outputs like this is a given for anyone using AI to produce deliverables. And while this may seem like a no-brainer, Deloitte’s recent blunder shows how easy it is to skip this step in the name of efficiency and speed. 

Another problem we encountered was a feedback loop involving our own data from the report. As we began to query in more specific areas, multiple different AI research platforms began to feed us back data from our past Tercera 30 reports. While upon first glance this was incredibly self-validating, the danger is exactly that. We ran the risk of blindly validating ourselves with stale data and commentary without challenging our narrative based on other resources and data from the past year. Luckily, we were able to identify this and use our previous thought leadership as contextual input, while ensuring we incorporated third-party resources. GPTs and agents are designed to interact with data that already exists and consolidate. Even then, the outputs are not consistently reliable (although accuracy continues to increase). 

4. Agents are powerful, but they can’t replicate a human experience

While agents can do many things, the power of a human relationship remains one thing that they can’t replace. We found great value in using the Bridgetown platform agents to conduct interviews with key market stakeholders, but we also found tremendous value in calling upon our professional relationships to validate (and debate) the selections we made for this list. 

Not only do we know where these individuals stand and what might be influencing their opinions, but through these conversations, we garnered insights that likely would not be shared with an agent that people don’t have a personal relationship with.

Exposure to expert and on-the-ground opinions is not only crucial to our process but the secret sauce to the Tercera 30.

While anyone with enough time and resources can replicate the base process that we have built, it’s tough to replicate relationships and discernment.

In a similar vein, it is dubious to expect agents and GPTs to automatically reflect the decades of institutional knowledge and context that we bring to the analysis and scrutiny of the data we gather for the Tercera 30. Unlike relationships, this can be trained into models over time, but it requires the foresight and structure to do so. Even then, these models will need to be updated by the rapid changes and learnings that we gain over time. 

5. AI cannot run the whole project (yet)

AI gave us speed and reach, but it couldn’t (and we didn’t expect it to) fill all of the gaps across an end-to-end project like this. At this point, AI wasn’t able to design the right models for our goals, drive the sequencing of work across teams, or keep all the moving pieces on track. We still needed people to frame the problem, pick the methods, set quality thresholds, and decide when to pivot. Beyond basic, repetitive tasks, AI still isn’t reliable at carrying that context through the full lifecycle.

The creative side is another limit. While AI can draft copy, suggest layouts, and generate visuals, the final deliverables still required human taste and judgment to tune the voice, ideate and contextualize data visuals, and make choices that fit our brand and audience. The tools are getting better every day, and we will keep experimenting, but they are not ready to own creative direction or produce truly differentiated deliverables by themselves. AI is a strong contributor. It is not the project owner.

The Future

AI tools are moving fast. What was unreliable this cycle could be ready by our next round of research. We will keep testing and adopting the best options for our work, and we will fold standout platforms into the ever-growing list of companies we consider for the Tercera 30.

Blurred lines: software and services converge

by Leave a Comment

How AI is redrawing the map for tech services firms

One of Tercera’s long-standing mantras for tech services leaders has been to pick a lane – services or software. You can’t be both. 

But in the AI era, that mantra no longer holds true. 

AI is not just disrupting how services are delivered, it’s collapsing the traditional boundaries between what a services firm is and what a software company does. 

AI has begun to take over tasks that once sat squarely in the domain of professional services: pulling and analyzing data, creating reports and nurture flows, developing and testing applications, migrating legacy code to modern tech stacks…you name it. More and more of those capabilities are being embedded directly into software workflows. Salesforce’s AgentForce, SAP’s Joule, and ServiceNow’s AI Now platform are great examples of this at play.

At the same time, forward-thinking services firms are starting to leverage AI to evolve their own offerings. They’re using co-pilots to move faster, and building their own agents to automate full processes that could once only be done by humans. And a growing number of firms are building their own software – not accelerators and reference architectures, but full platforms.

The rise of Services-as-Software

In our latest research report, A New Services Playbook for the AI Era, we explore the rise of a new category called services-as-software. Analyst group HFS, which coined this term, defines this category as AI-first offerings that automate traditionally people-based services, delivered through repeatable, software-like experiences, sold in a subscription or consumption model. 

And we’re seeing this model everywhere. 

Publicis Sapient’s AI-powered Slingshot platform promises to automate everything from prototyping to coding to maintenance. Accenture’s AI Refinery promises to help companies deploy agentic AI safely. Infosys markets a collection of solutions, including hundreds of industry-specific agents, under the Infosys Topaz brand.

It’s not just the large global firms though. Dozens of services-as-software startups are coming onto the scene to automate segments of the services space. Companies like Bridgetown Research, Auctor AI, Echelon AI and Rocket (just to name a few) are automating tasks that used to be purely human, and customers and software partners are starting to take notice.

And in between, are thousands of mid-size services firms. These firms don’t have the luxury of building from scratch, or the big R&D budgets to invest in developing sophisticated platforms from scratch. They are feeling the pressure, but also aren’t sitting still. The smart firms are getting creative with their investments. They are co-developing alongside customers and partners, extending platforms and diving into vertical use cases. They’re embedding tools and third-party services-as-software within delivery processes, and redirecting funds from stagnant areas or raising capital) to build out IP. They’re consultants – they adapt! 

This blurring of lines is forcing service leaders to take a hard look at their future direction. Are they growing a services firm or building a software company? Or increasingly common, how do they successfully do both when the DNA of services and software companies is very different?

This blurring of lines is forcing service leaders to take a hard look at their future direction. Are they growing a services firm or building a software company? Or increasingly common, how do they successfully do both when the DNA of services and software companies is very different? This decision has significant repercussions on many other downstream decisions – from how budgets are allocated, to go-to-market strategies, to the types of employees that are hired, to how they engage with their partners.

When Software Players Move Deeper Into Services

There’s no question the partner playbook is evolving. We’re seeing more software and AI companies leaning more heavily on their captive professional services teams. Not to replace partners, but to drive deeper consumption of their AI capabilities and maintain proximity to the customer where models are being trained and tuned. Their ‘land and expand’ mindset now includes not just license growth, but usage, outcomes and training. All of which require more human intervention.

This trend isn’t new, but it is accelerating. Some vendors like Palantir and IBM have had a blended services and software model for years. Palantir, which started as a services company and gradually became more productized, has been winning deals for years with its white glove Forward Deployed Engineers who work alongside customers to ensure the software performs and evolves. IBM has been focused on software and services for decades. Its global consulting business generates nearly a third of IBM’s total revenue. 

Some vendors, like OpenAI, are newer to the consulting game. OpenAI’s move into consulting felt like the shot heard round LinkedIn with many people suggesting this was the end of consulting. Yet OpenAI has been clear that (for now) their consulting arm will only tackle engagements that begin at $10 million. That leaves a lot of real-estate for other partners to play.

In our Tercera 30 research, one of the metrics we consider (for public companies that report it) is the percentage of revenue delivered through internal services, and how that is trending over time. We look at this metric because internal services arms can make it more difficult for services partners in the ecosystem to compete. 

Typically, software firms try to keep services to a relatively small percentage of revenue. Partly because margins are lower on services and partly because it keeps the company focused squarely on building the core product. However, as services margins increase and software companies focus on consumption, we wouldn’t be surprised to see these numbers will tick up. 

That in no way spells doom for external services partners. Services partners will still be invaluable to growth, and may become even more important as software firms look to drive value and consumption of their offerings. However, the partner playbook will change. Something that we tackle in our new research report.

The future belongs to the bold

This new world brings uncertainty, yes. But it also brings tailwinds for those that choose to ride them. According to our latest survey of 320+ tech services leaders:

  • 90% believe that their partners’ investments in AI will create new business opportunities for their firms
  • 95% say they’re already getting AI-related support from partners
  • 47% tell us that a significant portion or nearly all of their revenue is coming from AI work

In other words, most firm leaders see the trend not as a threat, but as an opportunity. An opportunity to build differentiated solutions, not commoditized services. To create more value for customers and the partners that they work with. And to capture a larger share of that value as well.

We’ve said before that AI changes the game in services. It resets the scoreboard. We believe the most successful companies in the AI era will see themselves as hybrid organizations, firms that blend automation with human expertise, and product thinking with an advisory mindset. Those that focus on long-term customer value, not just successful implementations and projects. Those whose growth isn’t completely tied to how fast they can hire. 

This next era of technology demands bold moves and shaking up beliefs that have held for decades. We’ve had to shift our own thinking because of this. So what long held mantras are you re-evaluating?