Inside PwC’s Shift Toward AI-Enabled Deal Execution
A conversation with PwC UK’s Deals CTO, Attul Karir, on embedding AI into deal workflows to deliver faster, evidence-based insights.
In today’s deals environment, teams are under constant pressure to move faster while maintaining the highest standards of accuracy and judgment. Delivering clear, defensible answers, often across vast and complex datasets, has become central to staying relevant.
At PwC, Attul Karir operates at the center of this evolution. As a Financial Services Valuations Partner and Chief Technology Officer for Deals, he combines hands-on deal experience with a mandate to embed AI, data, and technology into how the firm delivers for clients.
In this conversation, Attul explains how tools like Harvey are reshaping day-to-day work across the Deals team, from interrogating large document sets to translating dense regulatory material and generating executive-ready outputs grounded in evidence. This is changing how quickly teams move from question to answer, while preserving the rigor expected in high-stakes transactions.
He also shares what it takes to scale AI adoption in professional services, including the role of leadership, the importance of structured workflows, and the opportunity to put powerful tools directly in the hands of client-facing experts.
Can you tell us about your role within the Deals team and what your day-to-day work typically involves?
I split my time between two roles. I’m a client-facing Valuations Partner within our Financial Services Deals practice, where I help clients get transactions done by providing valuation insight, often working very closely with our diligence teams. A significant strand of my work also involves advising banks in restructuring and resolution contexts, including resolution valuations.
In parallel, I serve as the Chief Technology Officer for Deals, responsible for embedding technology, data, and AI into the way we solve client problems, evolve our offerings, and collaborate across the firm. That CTO role really boils down to ensuring our client-facing professionals have the best technology at their fingertips to deliver for our clients.
How do you see AI transforming M&A and professional services? What are the primary use cases where you and your team rely on Harvey today? What problems or bottlenecks are you able to solve?
AI is compressing the time from question to defensible answer, without compromising quality. In practice, Harvey is pivotal for three things.
First, ingesting and interrogating large, heterogeneous document sets like deal rooms, internal analyses, and regulatory references. It works across all the information and surfaces the key points with citations. The citation capability is critical, as it reduces hallucination risk and enables rapid verification by users of facts and figures.
Second, translating dense regulatory and technical material into precise, plain-English analysis tied to the exact paragraphs that matter. For valuation work, this is critical when referencing central bank resolution standards or expert witness guidelines.
Third, generative drafting grounded in our own analyses helps us by producing executive-ready, “at a glance” summaries from extensive workpapers and findings, all while preserving nuance around strengths, limitations, and recommendations.
These use cases address bottlenecks that traditionally slow deals down: manual document review, reconciling disparate sources, and producing high-quality output for senior stakeholders on tight timelines. By anchoring outputs to verifiable sources, we maintain the standard of evidence expected in M&A while moving substantially faster.
“By [:Harvey:] anchoring [its] outputs to verifiable sources, we maintain the standard of evidence expected in M&A while moving substantially faster.”
Can you share a specific example where you leveraged Harvey in an M&A, financial due diligence, corporate finance, or restructuring matter?
We recently ran a timed exercise for a globally systemically important bank to test the agility, robustness, and timeliness of its valuation capability within a resolution simulation, that’s effectively a failure scenario. The deliverable was a substantial report destined for board level and supervisory review.
We used Harvey to review the detailed analyses and synthesise observations across multiple sources of data. Ultimately, we used the tool to draft an executive summary that balanced strengths with targeted recommendations. Because Harvey grounded each point with citations back to our analysis and the relevant regulatory guidance, the team could validate conclusions quickly and iterate at pace, giving the client a crisp, defensible narrative for senior audiences.
Personally, I use Harvey to transform materials from my CTO responsibilities into precise content with citations, which has materially improved the speed and reliability of internal reporting and aided my decision making.
Which parts of the Harvey platform have been the most valuable for your team, and why?
For my client work, Harvey’s Vault and citation features are foundational. Vault lets us upload broad, complex document sets, from calculated model outputs to regulatory references, and query them cohesively. The ability to collaborate as a team and see what others have been working on improves understanding and efficiency as we can cut out the repetition of multiple people running the same queries. The latest update that significantly enhances Harvey’s ability to create and edit Excel and PowerPoint documents brings the platform even closer to my work and the deliverables I create for clients.
As the CTO for our Deals business, the development of Agent Builder is strategically important. Our aim is to process map core Deals activities and encode them into modular workflows and agents. This is how we’ll scale consistent quality, accelerate delivery, and ultimately package distinct components of our work into reusable, client-ready assets.
“[Agent Builder] is how we’ll scale consistent quality, accelerate delivery, and ultimately package distinct components of our work into reusable, client-ready assets.”
What impact has Harvey had on productivity, time savings, quality of work, or client outcomes across the Deals team?
The immediate impact is to enable us to deliver at speed with confidence. We answer complex questions faster and with clear source trails, maintaining the quality clients expect. On live deals, that means accelerated cycles without sacrificing analytical rigour; on client development, it means responding to inbound questions in near real time with well evidenced viewpoints.
We are also seeing a “show, not tell” effect in the market. As teams use Harvey in delivery, clients become curious about the tools underpinning our work, which naturally opens doors to joint experimentation and discussions about how their organisations can realise RoI from their tools or by adopting the technology that we use. This has clear implications for pipeline velocity and the depth of our client relationships.
What advice would you give other M&A or professional services leaders who are considering adopting AI tools like Harvey?
Lead from the front. Adoption increases when senior leaders use the tools actively and model best practice for their teams. Start with anchored use cases such as document interrogation with citations, regulatory mapping, and executive summaries so that quality is auditable from day one.
Pair tools with process discipline: map a workflow, define the sources of truth, and encode verification steps. Finally, think about democratisation. The goal is not to concentrate capability with technologists, but to put AI safely into the hands of client-facing experts, changing the tempo of client conversations and enabling “try it now” interactions that were not possible before.
Any other insights you can share?
We see Harvey as a core provider of AI capability across our delivery, so we’re working together to integrate some of the workflows and features into our Connected Digital Experience. This means bringing this capability directly into the tools our clients and teams already use to deliver AI-powered Deals. We also see significant value in collaborative features that enable shared environments with clients, creating a single source of truth for data, analyses, and outputs.
