9 Minutes - Read Now 
ABB has been in business for more than 140 years. It employs more than 110,000 people worldwide, and its work touches how the world manufactures, moves, and powers itself. And 18 months ago, it launched the ambitious project to bring its fully outsourced due diligence program fully in-house. And along the way, it rebuilt its own process from the ground up and added an AI tool to the mix, resulting in a success story that shows what can be done when strategic planning crafts a solution where AI amplifies human expertise rather than replaces it.
Fixing the House While Living in It
ABB’s original due diligence function came together quickly under a deferred prosecution agreement that gave the company little choice but to build something fast. It leaned heavily on outside vendors, ran at the group level, and grew by addition rather than design. Every new requirement got bolted onto what already existed instead of triggering a rethink of the whole process. The result? A program that technically existed but wasn’t built for the volume or speed the business actually needed.
Two costs came with that setup. The obvious one was money: the more outside vendors and outside people involved, the higher the bill. The less obvious one was time: external teams never fully learn a company’s customers, vendors, or risk footprint, no matter how long they’re on contract. Every unfamiliar case meant more rounds of emails, and high-risk cases could take as long as 50 days. Multiply that delay across every deal waiting on a screening result, and due diligence stops being a compliance function and starts being a business problem.
The Case for Building It In-House
The instinct when a process is broken is to look for a better vendor or a better tool. ABB’s team started somewhere else: with the process itself. Before deciding who should run due diligence, they designed what the ideal version of it would look like, what the business needed from it, and what had to change to get there.
The design process revealed that bringing things in-house would save ABB considerable money. After all, every external analyst who rolled off a contract took months of institutional knowledge with them, and every replacement had to be trained back up to speed on ABB’s customers and risk profile before becoming useful. An internal team not only keeps that knowledge, it creates a career path. Due diligence expertise built in-house becomes a pipeline into the wider legal and compliance organization, not a revolving door of contractors.
Not every ethics and compliance program receives a deferred prosecution agreement as a mandate to rebuild from the ground up, and that matters. But the business case for insourcing wins on the same cost math, the same speed argument, and the same reduction in dependency on people who don’t know the company.
Hiring for Judgment, Not Process
ABB built a team of 19 people across three hubs: Mexico, Krakow in Poland, and an Asia-Pacific hub split between Malaysia and India. The regional split wasn’t about convenience. It matched where ABB’s sales channels and vendor risk actually concentrate, and it put language skills and jurisdictional knowledge where the case volume was heaviest.
The bigger decision was who got hired into those seats. Most people can learn a due diligence process given enough time. What’s harder to teach is the instinct to look past what a procedure defines as a risk and notice what’s sitting just outside it. The cases that turn into real findings are rarely the ones that trip an obvious rule. They’re the ones where an analyst has a gut feeling that something doesn’t add up, and the standing to say so.
That instinct only holds up when backed by expertise the analyst doesn’t have to carry alone. ABB paired its analysts with a formal escalation network of internal subject-matter experts across anti-bribery and corruption, sanctions, sustainability and human rights, and antitrust, so any case that outgrew what a generalist could responsibly assess had somewhere specific to go. Early in the build, the team also brought in a secondee from an outside due diligence firm to help design the process itself. That combination answers a skepticism that comes up constantly when compliance functions consider insourcing: the assumption that a newly built internal team can’t match what an established external firm already knows about sanctions, anti-bribery, or export control law. The answer isn’t pretending a small internal team can be expert in everything. It’s building the bench of expertise around the team and giving analysts a fast, clear way to reach it.
Where AI Earns Its Place
Compliance software often lives disconnected from the case management system an analyst already works in. When a tool like that becomes one more login instead of one less step, it can die a slow death of low adoption no matter how capable it is. ABB’s rule for any new technology, AI included, was that it had to work inside the process already running, not next to it.
That’s the job AI does inside ABB’s due diligence workflow. Xapien handles entity verification up front, confirming an analyst screens the right person or company before applying any risk logic to them. It turns scattered public information into a single readable narrative report, so analysts spend their time interpreting the gray areas instead of hunting down source material. And when a case escalates to a subject-matter expert who has a full-time job outside of due diligence, an AI-generated summary hands them the two or three things that actually need a decision instead of a full case file to read cold.
None of that changes who makes the call. AI speeds up the collecting and the summarizing so people have more time for judgment, not less of a role in using it. When an AI-generated risk read doesn’t match what an analyst already knows about a relationship, that gets a follow-up question, the same way an unexpected answer from a colleague would. And there’s a boundary that doesn’t move regardless of what a tool can technically do: anything that requires attorney-client privilege still goes to an attorney. AI was never going to solve for that, and ABB never asked it to.
Vendors receive screening across the board. Customers and sales channels get a tiered approach: one baseline level of screening for everyone, and a deeper, more resource-intensive process reserved for the higher-risk sales channels where it matters most. There’s no universal answer for how much scrutiny a third party deserves. The level of screening should match the level of risk, not apply itself out of habit.
The Guardrails That Come Before the Rollout
Adopting an AI tool for due diligence raises the question of what happens to the data. Whether a tool operates in a closed environment, one where information doesn’t flow back out to train someone else’s model, matters as much as what the tool can find. Doing due diligence on the AI tool itself, before it touches a single case, is the first step. Skip it, and every efficiency gain downstream sits on a foundation nobody checked.
ABB tracks cost, speed, escalation volume, and satisfaction. The business case started with cost, which fell once the team came in-house. Then came speed: cases that used to bounce back and forth between analysts and the business for extra questions saw that back-and-forth cut by more than half. Escalations to subject-matter experts who already had full-time jobs elsewhere dropped as analysts got better tools for finding answers themselves. And through all of it, satisfaction with the process on the business side stayed above 90%.
The team also learned to listen to two different audiences. End users, the people actually going through onboarding and screening, will tell you directly what’s working and what isn’t. The people several levels above them mostly hear about due diligence when something goes wrong: a deal that took too long or a frustrated customer. Both signals matter, but chasing only the loudest one risks overcorrecting for a single bad case.
Nobody Gets to Call This Finished
ABB is now extending the same in-house approach to work that used to go entirely to outside due diligence firms, including the higher-stakes screening that comes with an acquisition or an investment, where speed matters even more and paying for a five-day turnaround from an outside firm stops making sense. On the technology side, the next step is tighter integration: pulling tools like Xapien into ABB’s case management system through APIs instead of running them as a separate application analysts have to remember to open.
There’s a trap worth naming here. Once a team hits a new baseline, cheaper, faster, and better than before, it’s tempting to treat that as the finish line. It isn’t. The number that got a budget approved two years ago won’t get one approved again. The only way to keep justifying the same size team is to keep finding higher-value work for the time AI and a better process just freed up.
The Real Lesson
AI as a force multiplier is an easy phrase to say and a specific thing to build. ABB’s version of it didn’t start with a shopping list of AI vendors. It started with a process the team wanted to run, built the internal expertise to run it well, and only then figured out where a tool could handle the collecting and summarizing so people could spend more time on the judgment calls that actually require a person. That order matters more than any individual piece of technology. Get it backward, buy the tool first and design the process around it, and a compliance program ends up with expensive software nobody quite uses. Get it right, and the people running the program aren’t managing risk from the sidelines. They’re making the business faster and safer at the same time.
To hear this conversation in full, including how ABB and Xapien think about tooling integration, privilege, and what the next 18 months may bring, watch the full webinar, “AI as a Force Multiplier: ABB’s Due Diligence Transformation,” on demand: https://ethisphere.com/webinars/how-abb-transformed-its-global-due-diligence-function-with-ai/
Learn how to transform compliance from business blocker to strategic advantage at XapienXchange on 7 October 2026 at 2 Savoy Place, London. Register here: https://xapien.com/conference/#register-now
Author Bio
Bill Coffin | Editor-in-Chief, Ethisphere Magazine, Ethisphere
Bill Coffin is an internationally recognized writer and editorial leader in the fields of ethics, compliance, insurance and risk management. His work has appeared in The Wall Street Journal, BusinessWeek, Forbes and Fortune, as well as Compliance Week, National Underwriter, Risk Management, Best’s Review and other B2B media. Previously, Bill served as Director, Senior Editor at AIG, leading content strategy for one of the world’s largest insurance companies.
Kyle Lewis | Head of Integrity Operations, ABB, Guest Contributor
Kyle Lewis is Head of Integrity Operations at ABB, where he leads two complementary teams: a global Due Diligence organization supporting third-party risk management activities across ABB, and the Motion Integrity Operations team responsible for integrity program execution, change management, communications, learning, digital transformation, and integrity support for China.
Jessica Denny | Head of Comms and Marketing, Xapien, Guest Contributor
Jessica Denny is Head of Comms & Marketing at Xapien, a purpose-built AI-native due diligence platform that delivers complete visibility of third-party risk with unmatched speed, accuracy, and depth, enabling organizations to make rapid, confident decisions in an evolving risk landscape. Jess is an expert in international strategic communications, and she has obtained CAMS Certification from the Association of Certified Anti-Money Laundering Specialists. She was also Xapien’s first client.
