Megha Sinha, founder and CEO of Kamet Consulting Group explains why the systems most manufacturers use to manage post-approval changes are no longer fit for purpose, and what a viable alternative looks like in the AI age.
Kamet Consulting Group
There is a category of operational risk in pharmaceutical manufacturing that absorbs enormous resources, drives significant cost overruns and creates genuine supply continuity exposure - yet rarely features in boardroom conversations or investor presentations. It is post-approval change management, which sits at the intersection of regulatory affairs, supply chain, quality and manufacturing operations. It is unglamorous, technically demanding and almost universally under-managed. Yet the scope can be all-consuming.
A large pharma company will typically assess around 6,000 post-approval changes annually. Around 3,300 of this will necessitate regulatory submissions, generating an estimated 90,000 country-level filings across more than 140 health authorities worldwide. A single manufacturing site transfer might involve 10 or more functions simultaneously - regulatory, supply chain, quality, manufacturing, labelling, legal, pharmacovigilance and IT - each operating to its own timeline, across what could be 50-100 countries each with their own submission categories, review windows, and rules governing implementation grace periods. Achieving full global approval for one change can take up to five years, which seems astounding in today’s fast-paced market. For most companies, that entire process is orchestrated through spreadsheets and email, even now.
The failure mode no one owns
What makes this particularly difficult to address is that it rarely presents as a crisis. Missed deadlines are simply absorbed, and budget overruns explained away. Supply gaps are patched through expedited shipments and the considerable efforts of in-country affiliate teams already operating at capacity. No single incident is serious enough to trigger a strategic response — but the cumulative impact on operational performance, cost and risk is substantial.
The propensity for failure is structural. No single individual or function owns post-approval change management, end to end. Regulatory affairs tracks submissions in one system; supply chain manages inventory transitions in another; quality owns change control in a third; artwork and labelling sit in a separate workflow entirely. No one has a unified view of how a delay in one country’s approval cascades into a supply gap in three others — or how an inflight submission in one market blocks a new variation from being filed in a second.
The consequences are tangible. In one large-scale API site transfer spanning more than 50 markets, a missed grace period deadline in a strict jurisdiction created a supply gap lasting several weeks, with revenue at risk in the millions. The regulatory submission had been
approved on time. The problem was a coordination breakdown between the supply chain team planning the manufacturing cutover and the regulatory team tracking grace period trigger dates - each working to a different timeline, with no one connecting the two until buffer stock could no longer be built. In a separate case, an artwork pipeline misaligned with a regulatory submission sequence resulted in a four-month delay to market entry and around $200,000 in packaging write-offs. A divestiture requiring transfer of marketing authorisation holder status across more than 80 countries, meanwhile, generated inconsistent filings and rejected submissions, resulting in a remediation programme that took over two years and cost several million pounds to resolve.
Converging pressures are making a bad situation worse
Several forces are compounding an already strained baseline. The M&A and divestiture activity of the past five years has been considerable in scale, and much of its operational aftermath is still affecting regulatory operations teams today. Supply chain restructuring is accelerating in parallel, as companies nearshore and reshore manufacturing in response to geopolitical risk, the US BIOSECURE Act and the supply disruptions of the COVID era. Every facility move or CDMO switch generates a cascade of post-approval changes across every registered market.
Regulatory requirements across more than 156 countries continue to diverge rather than harmonise. Manufacturing grace periods vary from zero to 365 days, with different trigger events per market. And the experienced regulatory operations professionals who historically held all of this complexity in institutional memory are steadily retiring, without that knowledge being captured in systems. The same-sized teams are absorbing two to three times the change volume with the same manual tools they have always used.
A business decision treated as a compliance exercise
There is also a more fundamental problem that rarely surfaces in leadership conversations: many lifecycle changes are initiated without any clear understanding of whether they make financial sense. Manufacturing proposes a cost-saving initiative - a supplier switch, a site consolidation - and the projected savings look compelling. But once regulatory fees across 50-plus markets, artwork updates, dual production runs, packaging write-offs, resource costs and two to three years of execution time are factored in, the actual cost can equal or exceed the intended savings. Because planning is fragmented across functions, no single point exists at which someone calculates total execution cost against projected benefit. Changes are treated as compliance questions rather than business decisions to justify or challenge.
The smarter approach that AI enables
Solving the problem of lifecycle change management requires three intertwined considerations. Organisationally, companies need to assign a designated programme lead for each lifecycle change - someone with cross-functional visibility and genuine coordinating authority. Most do not have this today. From a process perspective, there needs to be standardised planning frameworks that map regulatory pathways by country, identify cross-functional dependencies and sequence execution with defined escalation paths. From a technology perspective, companies should establish a structured regulatory intelligence layer: a validated database
covering not just what to submit by country and change type, but what happens after submission, and the manufacturing and supply chain constraints that apply while approval is pending.
AI is now genuinely capable of encoding practitioner expertise at scale - computing cross-functional dependencies, flagging risks before they materialise and providing the aggregate portfolio view that no human team can sustain manually. The probability of AI delivering near-term operational return in lifecycle change management is meaningfully higher than in many areas currently attracting far greater industry investment. The bottleneck is not technology; it is the absence of recognition, at a sufficiently senior level, that this problem category warrants attention and budget.
Companies that build structured regulatory intelligence, establish cross-functional governance, and connect planning to execution in a single unified view can expect to deliver change faster, spend less doing so, and more effectively protect supply continuity than those still relying on spreadsheets and institutional memory. Happily, it has never been easier than now to make the change.