95% of
AI projects fail.
13 leaders share how to succeed.
Learn how to turn operational chaos into calm. B2B service providers share how to fix your foundations, scale efficiently, and keep clients coming back.
Chapters
Featuring
Jon McClay
Jessica Samadi
Kathiravan Udayakumar
James Hall
Kit Cox
Kumar Lalit
Modhura Roy
Mukta Agarwal
Randima Krishnaratne
Satheesh Neelam
Stephanie Hamon
Suri Babu Komma
Vipul N Mehta
Still figuring out your operations?
Join our exclusive webcast
Service leaders from Vistra, Anthesis, and Enate share the unfiltered truth about implementing AI and fixing operational foundations. Limited spaces.
Introduction
The uncomfortable truth about running B2B services today
Most organizations score below 30 out of 100 on the Operational Excellence Index (OEI) - Mckinsey [1]
Let’s be honest, in day-to-day service delivery, how often do clients say they’re truly satisfied? It's a rarity. And for those on the inside running the show, it's a matter of chaos behind the calm. Before the morning coffee has even cooled, there are three urgent escalations demanding attention. One client in Singapore is questioning a payroll deadline, another in London asking why their fund administration report is delayed. And because teams are spread out all over the world using their own systems, even knowing the status of client work becomes a headache.
Randima Krishnaratne, Global Programme Director at A.P. Moller Maersk says,
"Some of our clients operate in more than 100 countries, meaning we're navigating 100 different cultures and communication styles. Each person expects a customized response, which is challenging to deliver at scale."
The brutal reality about running B2B services in 2025 is that margins are shrinking, clients are more demanding than ever, and the unique competitors who get operations right are stealing market share. Add to that the promise and pressure of automation, and it's no wonder service providers are stressed to breaking point. Customer satisfaction has become the last frontier of competitive advantage. When products and pricing are on par, operational excellence becomes the true differentiator between keeping clients and losing them. Yet most providers are struggling with the basics. The service works, but only just, and there’s no capacity for anything extra. If this sounds familiar, it’s because it’s become the status quo. Modern operations are broken, and it’s hurting the bottom line.
Vipul Mehta, Senior Vice President of Private Equity & Portfolio Transformation at SaxeCap describes the shift,
"A couple of years ago, clients were comfortable with certain processes being performed status quo. Now, during renewal discussions, they come with substantially higher expectations."
The answer won’t be found in yet another spreadsheet or the minutes of a bi-weekly automation meeting. It lies in taking a step back and acting holistically. That’s what this playbook is about. We interviewed *13 leading B2B service providers across various industries to share the good, the bad, and the ugly of running services today. Together, we’ve built a vision for the future. One where operations are organised and structured, work gets done on time, and clients stick around for the long haul.
Read on to see how, together, we can fix the foundations and build better services, for customers, clients, and the industry at large.
Chapter 1.
Poor visibility is driving clients away
In the conversations we have with operational heads, transformation leads, and C-suite executives, there's always a common theme. Industry leaders want to make their operations more efficient and structured, yet struggle to answer fundamental questions about their own operations. For instance, a few of the queries we hear time and time again are…
- What are our people working on?
- How busy are they?
- How efficient are they?
- Do we need more resources?
- Are processes performing as well as they should?
This is crucial information, yet most leaders don't have the answers because information gets collected at the client level in manual Excel trackers, shared Outlook mailboxes, and disconnected systems. Almost no time tracking exists of how long it takes to complete work, with estimates based on what teams think they should tell the client rather than what's really happening. The result is that it's nearly impossible to aggregate data across teams to get a business-wide perspective on performance, bottlenecks, or capacity, leaving leaders scratching their heads while clients demand faster, more reliable service delivery.
Kumar Lalit, Vice President and Partner in Applied Advisory at Genpact, puts it simply:
"If you can't measure [operations], you can't improve it."
The consequences of this visibility gap extend beyond day-to-day operations. A recent MIT study looking at AI ROI [2] found that while companies pour over half their AI budgets into sales and marketing tools, the highest returns actually come from back-office automation and operational improvements. Without clear visibility into where inefficiencies are, organizations are essentially investing blind, throwing technology at the wrong problems while the real opportunities for transformation remain hidden in their operational data.
Jon McClay of Baker McKenzie comments on the core challenge,
"I can't imagine incorporating AI effectively if your data remains fragmented across on-premises silos."
Meanwhile, clients remain unaware of the operational challenges behind the scenes. They're focused on results.
Mukta Agarwal, Client Partner for Government & Regulatory, Banking & Financial Services at Cognizant, reflects that "clients are eager to get results quicker and faster,"
but without fundamental data at hand, providers can't reliably meet service level agreements. Without understanding processes inside out, performance becomes guesswork rather than strategic management.
Service providers can't afford to run sophisticated services without visibility anymore. It's not only an operational challenge, it's also incredibly time-consuming. Staff sifting through thousands of emails daily or performing data entry manually makes service slow, unreliable, and inefficient. It's an issue that Jessica Samadi, Director of Operations at a leading FS firm, knows well:
“Within Corporate Services, handling requests for certificates of incumbency and good standing involves considerable manual work despite being high-volume, routine tasks."
In 2025, with all the technology at our disposal, the question remains: Why does manual work still dominate service delivery?
Many service leads have attempted technology solutions. They've tried implementing CRM systems, project management tools, or workflow solutions, only to find these approaches too rigid, complex, or misaligned with how work actually gets done. The result is abandoned investments that never quite delivered on their promise.
Service providers don't just want to deliver better service, they have to. It's their core business, and without operational foundations that support systematic improvement, they risk being replaced by competitors who've already solved the visibility problem.
Companies succeeding today understand that operational excellence isn't about working harder. It’s about having X-ray vision into what's actually happening on the ground so you can make improvements.
Chapter 2.
Why 95% of businesses get zero ROI from AI
Why GenAl pilots fail: top barriers to scaling Al in the enterprise
Users were asked to rate each issue on a scale of 1-10
"People are falling into the trap again of going for the shiny new tools, jumping to the solution, forgetting that they first need to identify what problem they're trying to solve, how big that problem is, and whether it warrants a Gen AI solution."
Stephanie Hamon, Head of Legal Operations Consulting at Norton Rose Fulbright
MIT's recent study of over 300 AI initiatives revealed something surprising. 95% of businesses are seeing zero ROI from their AI investments. Despite $30 to $40 billion being invested collectively, most organizations have little to show beyond sophisticated chatbots and pilot projects that don’t scale [4]. This is the reality three years since ChatGPT's debut, despite AI dominating boardroom agendas, conference keynotes, and transformation budgets across the industry. The promise that intelligent systems will slash costs, accelerate delivery, and finally solve the productivity puzzle hasn’t come to fruition.
The disconnect between AI's promise and its performance isn't a technology problem, it's an operations problem. Service leaders are attempting to patch AI onto broken processes, expecting transformation from systems that have no structured data to learn from, no clear workflows to improve, and no visibility into what actually needs to change.
Stephanie explains the importance of getting operations in order before deploying technology solutions: "
Before implementing any AI solution, you need to address three critical areas: people, processes, and data. Particularly important is understanding the processes you're trying to improve. There's no point fixing something that isn't broken or improving a process that's fundamentally unsuitable."
The reality is that most service operations are disjointed, messy and primarily run through email. Work flows through shared mailboxes with no tracking. Critical processes exist only in the heads of senior team members. Data lives in scattered spreadsheets that don't communicate. Performance metrics are estimates at best. In this environment, AI becomes a shiny, pointless widget rather than a measurable solution.
Interestingly, MIT researchers discovered something crucial. Companies that buy specialized AI tools succeed 67% of the time,[5] while those building internal solutions succeed only 33% of the time. The biggest ROI comes from ditching outsourcing contracts, reducing agency fees, and streamlining operations. This gap between investment and opportunity goes some of the way to explaining the failure.
Modhura Roy, AI Strategy & Operating Model Transformation at Cognizant, cements this point.
"Before implementing AI solutions, you must standardize the core operating processes in which AI will be injected. If processes are fragmented, AI will only exacerbate the problems."
The businesses seeing the biggest returns from AI aren't necessarily the ones with the biggest budgets or most sophisticated technology. They're the ones that fixed their operational foundations first. They mapped their processes. They consolidated their data. They created visibility into their workflows. Only then did they apply AI strategically to specific, well-understood problems. Not only this, but they started small with experimentation and pilots rather than jumping in two feet first.
Suri Babu Komma, Vice President of Digital Transformation and AI Solutions at EXL, advocates for this approach.
"Instead of proposing multi-year, multi-million pound projects, we identify targeted solutions through one or two-month sprints, allowing clients to reinvest savings into additional improvements."
The path forward isn't about abandoning AI altogether, it's recognizing that scaling digital transformation requires operational maturity. Service providers need structured workflows, consolidated data, and clear visibility before AI can deliver on its promise. Those who get their house in order first will be the 5% seeing real returns. And for real returns to happen, orchestration becomes essential.
Chapter 3.
Orchestration is the bedrock of decent service
B2B service providers are all-too-familiar with the challenge of managing work across multiple systems, teams, and different locations. Despite having thousands of technology tools at our disposal and supposedly being more connected than ever, the big picture remains messy to say the least. Some have found the answer lies in process orchestration.
If you’re wondering what the term ‘process orchestration’ means, you’re not alone. Kit Cox, Founder and CTO of Enate, explains orchestration using a simple but powerful analogy:
"You can think of orchestration like an orchestra and processes as all the people playing their instruments. For an orchestra to make a great sound, every player has to play their part at exactly the right time to deliver an amazing outcome. And orchestration is about delivering the right work, to the right resource, at the right time, to deliver a great service for customers.”
It’s essentially ensuring people, systems and processes are working together from one place.
Orchestration has been backed by leading analyst firms, Forrester[6] and Gartner [7] and the global process orchestration market size is valued at USD 7.32 billion and projected to grow at a CAGR of 21.1% from 2025 to 2030 [8].
The need for orchestration becomes particularly important when operating at scale with high-volume, highly-repetitive work across thousands of employees. As Kit adds,
"Orchestration really shines when you're delivering B2B services at scale. If you've got hundreds or thousands of people involved in delivering consistent but sometimes variable services across lots of clients and lots of jurisdictions, you need orchestration to help you get control and manage that business more effectively."
Rather than managing individual tasks or processes in isolation, orchestration provides one single view of how work flows through the entire operation. James Hall, CEO of Enate, coined the term "operational soup", when work is carried out, but there's no clarity on how much work has been done, by whom, or how exactly it's processed.
The ‘operational soup’ situation was one that Felipe Araya, Global Head of Operations at TMF identified with prior to using orchestration.
TMF Group is a global business offering accounting, corporate secretarial, HR administrative and capital services to businesses across 50+ countries. Before implementation, TMF struggled to collect and consolidate data on service delivery across 90+ global offices. They knew what services were being delivered but had no visibility into who was doing what, when, or how long anything actually took.
The business was receiving lots of end-customer requests through various channels such as email, self-service and call centre tickets. This was making it difficult to see the full picture and work efficiently. When work requests are scattered across various pockets, there is the potential for crossed wires, wasted time and it creates a general margin for error. TMF came to us after realizing they could house all of these work processes under one roof and track and manage all tasks using the Enate platform.
In addition, TMF also wanted to manage their hybrid workforce more efficiently by introducing UIPath RPA bots to automate dull, repetitive tasks and allow employees to take on more challenging work.
TMF decided to try a process orchestration solution, enabling 4,500 users to communicate and work through one platform. The difference was huge. Felipe admitted that...
"Having orchestration implemented across our departments can be likened to having X-ray vision into your operations."
Using orchestration, TMF achieved a £32 million uplift in profit margin, showing the true power of getting your operations properly in order.
Satheesh Neelam, Global Client Partner at Cognizant, identifies with the opportunity to orchestrate. He said,
"Orchestration represents the single greatest opportunity for organizations to move beyond quick wins and establish a proper foundation for AI adoption."
Unlike some transformation projects that drag on for years, orchestration solutions can be deployed in weeks, not months. They integrate with existing technology infrastructure rather than requiring organizations to start from scratch, and many are no-code enabling employees to build processes from the ground up using citizen development.
The business impact of orchestration is proven in a matter of months. Enate’s CEO, James Hall cites average improvements of 20% productivity gains and 33% increases in customer satisfaction. And Felipe Araya hits on something more fundamental,
"Don't underestimate the power of analyzing data in your orchestration platform. Implementing the software is 50% of the work; analyzing and acting upon what the data is revealing is where the real improvement comes from."
TMF's results validate this approach, their £32 million margin improvement wasn't about working harder or using more resources. It was about finally being able to see and control how work actually flows through their organization, and make improvements to processes accordingly.
Felipe's advice to other service providers is as follows...
"If you're thinking about implementing orchestration software, stop thinking and do it. The process may be difficult at first but eventually you'll ask yourself 'how did we manage without it?'"
The companies delivering brilliant service today are using orchestration to turn operational chaos into systematic excellence. The question for most service providers isn't whether they need better operational control, it's how much longer can they really afford to operate without it?
Chapter 4.
The rise of Services-as-Software
For decades, the B2B service delivery model has followed the same old formula. Organizations sell people's time, success is measured by meeting service level agreements, and scaling means adding headcount. Software is a part of it, but the service is very much delivered by people. However, according to Phil Fersht, CEO at HFS Research [9], this entire model of delivering service is dissolving. The industry is entering the era of Services-as-Software, where tech solutions don't just support services, they become the service itself. He says,
“The reality is that most people-based services, once they become predictable and routine, eventually become automated.”
Modhura Roy, AI Strategy & Operating Model Transformation at Cognizant, echoes this sentiment.
"Traditional 'software-as-a-service' business models may not be sustainable five years from now, giving way to more disruptive 'services-as-software' models. The industry is fundamentally transitioning from human-delivered services to software-led services."
Services-as-Software represents a complete glow-up of service delivery. Rather than providing tools for humans to use, these autonomous platforms complete work, learn from outcomes, and optimize performance with very little need for human intervention. Fersht describes it as the point where:
"The line between services and software is blurring and eventually vanishing."
Organizations no longer buy software to help deliver services, they buy software that forms the bedrock of the service.
Three external forces are driving this shift. Firstly, client expectations have evolved beyond traditional metrics. Suri Babu Komma, Vice President of Digital Transformation and AI Solutions at EXL, says:
"Today, it's about business outcomes rather than just SLAs. In accounts payable [processes], clients previously focused only on processing metrics like invoice volume and accuracy. Now they expect insights about optimizing cash outflow, vendor relationships, and fraud analysis."
Secondly, economic pressures call for new commercial models. The traditional approach of billing for time and resources no longer meets client demands of knowing how efficient a service is and where they’re getting value. Thirdly, AI has matured to the point where when operations are in order, it can genuinely replace, not just augment, vast swathes of human service delivery. HFS Research found that a whopping 60% of enterprises plan to replace professional services with AI within three to five years [10].
The transformation has already started. McKinsey reports that organizations who have a high level of AI maturity can handle 70-80% of customer interactions through self-service channels [11]. One standout example is Aviva, which saved £60 million [12] in 2024 through AI transformation of their motor claims operations, reducing liability assessment time by 23 days for complex cases.
Kumar Lalit, Vice President and Partner in Applied Advisory at Genpact, describes the change.
"There’s a shift from FTE-based commercial models to outcome-based and value-based models. Rather than charging for the number of employees deployed, service providers can offer commercial arrangements based on specific outcomes, such as reducing invoice processing costs from $7 to $2 per invoice and sharing the resulting value."
By 2035, Fersht predicts Services-as-Software will grow into a $1.5 trillion market [13]. Service delivery will be primarily technology-driven, with human intervention reserved for exception handling and strategic decisions. Kathiravan Udayakumar, Oracle Practice Leader at Cognizant echoes this sentiment.
"I believe we're witnessing a fundamental shift towards outcome-driven services. Currently, the industry operates primarily on FTE, time-bound or scope-bound service models, but this is changing as clients increasingly measure value by outcomes rather than resources deployed."
The path forward requires a total rethink of the way that services are run, managed and delivered. Organizations are going to have to shift from just managing teams to orchestrating AI agents, and from time-and-materials pricing to outcome-based commercial models which is no easy feat.
The choice facing service leaders is clear. Pivot into technology companies that happen to employ people, or risk losing it all as clients turn to providers who've already made this leap.
Chapter 5
10 tips for running better services
- Ask uncomfortable questions
Do you know what your team's currently working on, how busy they are, or where the bottlenecks sit? If you're drawing a blank, that's your starting point. - Stop the spreadsheet spaghetti
Before you even think about AI or automation, pull all that information trapped in shared mailboxes and Excel trackers into one place where you can actually see what's happening. - Fix broken processes first
Nobody wants to hear this, but throwing technology at a broken workflow just gives you a faster mess. Document what actually happens (not what the handbook says should happen) before you automate anything. - Run pilots, not marathons
Those three-year digital transformation programmes? 70% fail. Pick one specific problem instead, run a 1-2 month pilot, measure what changes, then decide if it's worth scaling. - Move beyond "did we hit our SLAs"
Your clients don't just want to know you processed 500 invoices on time. They want to understand how you're helping to improve their cash flow or reduce their fraud risk. - Buy an off-the-shelf solution
Companies using purpose-built solutions succeed twice as often as those building their own from scratch. Unless you're a software company, don't try to become one. - Measure how long things actually take
Not how long you think they take, or how long you tell the client they take. Real numbers on real work. Without this, you're guessing at capacity and pricing. - Test outcome-based pricing on something small
Pick one high-volume process and run a commercial model based on actual results rather than hours worked. Start slow and scale-up. - Know your use case
Be clear on what you want to achieve with AI. For instance sentiment analysis for customer satisfaction or IDP for invoice processing. - Learn from what the data tells you
Implementation is only half the job. The real margin improvement comes from looking at what the data reveals and doing something about it.
Join our
exclusive webcast
If you've made it this far, you're probably thinking one of two things. Either "this all makes sense, but where do I actually start?" or "I need to hear from someone who's actually done this before committing to anything."
Fair enough. That's exactly why we're hosting a webcast with service leaders who've been in your shoes and come out the other side.
Join us for an exclusive webcast: AI Dreams vs. Reality: How to implement AI for service delivery success
We're bringing together industry leaders from Vistra, Anthesis, and Enate. This group has tackled the messy reality of fixing operations and they’re going to be sharing the unfiltered truth about running better services.
What you'll learn
- How to fix your operational foundations before scaling AI
- What "brilliant service" actually requires from an operations standpoint
- Practical steps to prepare for the future of service delivery
The panel
- James Hall, CEO, Enate (Moderator)
- Kiran Sinharoy, Vistra
- Will Harris, Industry Advisor
- Kit Cox, Founder & CTO, Enate
Save your spot
Spaces are limited because we want to keep this intimate and valuable. If you're serious about fixing your operations and getting AI to actually deliver, this is worth an hour of your time.
About this playbook
This playbook was created by Enate, a process orchestration and AI solution purpose-built to help businesses run smooth services at scale.
References
[1] Next-generation Operational Excellence, McKinsey & Company, January 2024
[2-5] The GenAI Divide: State of AI in Business 2025, MIT NANDA Report, August 2025
[6] Announcing the Evaluation of the Process Orchestration Market, Forrester, September 2025
[7] Embrace the future of automation with BOAT, Gartner, July 2025
[8] Process Orchestration Market Size, Share & Trends Analysis Report, Grand View Research, July 2024
[9] Services-as-Software presents a $1.5 Trillion Opportunity, HFS Research, February, 2025
[10] Software Services are Blurring, HFS Research, December 2024
[11] The Next Frontier of Customer Engagement: AI-enabled Customer Service, McKinsey & Company, March 2023
[12] The Future of AI in the Insurance Industry, Mckinsey & Company, July 2025
[13] Services-as-Software Market, HFS Research, July 2025