AI shows up in all five tracks at ISPE's 2026 Annual Meeting

Bob Chew read the five tracks of the 2026 ISPE Annual Meeting and Expo aloud and asked Connie Langer how they fit together. Langer chairs the conference's international program committee and works as regulatory intelligence lead for pharma quality, safety and environmental operations at Pfizer Global Supply. Her answer was that a breakthrough means little without a workforce that can execute it and a governance structure that keeps it inspection ready.

The exchange came on the ISPE Podcast, Shaping the Future of Pharma, in a MarketScale Studio segment, with Chew as host and Jose Caraballo, founder and principal of JCO Global Advisory LLC and the committee's executive chair, alongside Langer. The meeting runs October 18 to 21 in Washington, DC, under the theme "from science to systems to patients." For a plant manager or quality head deciding whether to send a team, the design of the agenda is the story.

The timing sits inside a documented mood. The Deloitte Center for Health Solutions, in a report published December 9, 2025, surveyed 280 C-suite executives at biopharma and medtech companies between August and September 2025 and found that more than 75% were confident in their own organizations' financial outlook, while only 41% felt optimistic about the global economy. Deloitte's read is that the coming year rewards organizations that balance bold investment in AI and emerging technology with the ability to adapt to regulatory and economic change.

Confident inside the walls, cautious about everything outside them. That is the audience ISPE's chairs are programming for.

Five tracks built to read as one story

Langer walked through the sequence. Breakthrough to Bedside covers discovery, the science and development, and how a discovery becomes something that can be manufactured, scaled and delivered. Systems that Deliver Reliability picks up from there with the manufacturing, supply chain and technology backbone that has to perform every day.

Digital Data and Decision Integrity is where AI and analytics come in, in Langer's description, turning trusted data into better and faster decisions. People, Capability, and the Workforce of Delivery is about upskilling for an AI-enabled world. Governance, Risk, and Readiness is the oversight that keeps the whole system in control.

Caraballo described the same arc from the other end. Labs discover, development refines, then someone has to design the facilities, systems and processes that produce the medicine, and a supply component still has to get it to the patient. He said the committee wanted any topic under that umbrella to have a place on the agenda.

The shape is not unique to ISPE. The Drug Information Association, in program material for its 2026 Global Annual Meeting, introduced updated tracks that include Track 3, Data, Technology and AI, and Track 9, R&D Quality and Compliance, and said the new set was built to spark cross-functional dialogue. Cambridge Healthtech Institute, in its own program material for the Bio-IT World Expo, describes a Data Science and Analytics Technologies track on May 20 and 21, 2026 focused on moving from mathematical models to production-grade, validated pipelines in regulated environments.

DIA and Bio-IT World Expo both put data and analytics tracks on their 2026 agendas. DIA pairs its data and AI track with a separate quality and compliance track, while the Bio-IT World Expo material describes a single data science track built around trust and validation. For an operations leader choosing which meetings to fund, that overlap means the vocabulary around trusted data will be shared across events, and the ISPE agenda is the one that ties it to the factory floor and the supply chain.

AI has no track of its own on purpose

Chew asked how organizations can use AI, analytics and digital tools to improve decisions while keeping trust, oversight and data integrity. Caraballo's answer reframed how attendees should read the agenda.

Jose Caraballo expects AI presentations across all five tracks even without a dedicated AI track. He described AI as a tool for standardized work, pointing to reports he had read of AI writing code faster, while judgment calls stay with humans who remain accountable. The efficiency, in his view, buys time for the harder questions.

Langer described what that looks like inside a large manufacturer. Her company is investing heavily in training and in giving teams access to new tools, has partnered with universities on general AI training that is mandatory, and lets colleagues opt in to graduate-level classes leading to a certificate in the management of AI. She said the pace of change makes rolling out technology alone insufficient, and that literacy and trust have to be built alongside it.

For a CIO or quality VP signing off on an AI pilot, the operating question becomes who on the team can tell a fast answer from a right one. Deloitte's December 2025 outlook makes a related point in its own terms, noting that most surveyed leaders expect real transformation in 2026 to involve innovative thinking, agile operating models and strong external partnerships.

Training the people already on the payroll

Langer's point about pace has a headcount behind it. Cora Systems, a project management software vendor, wrote in an April 13, 2026 guidebook on its own site, citing a UK BioIndustry Association forecast, that the UK life sciences sector alone may need as many as 145,000 new and replacement workers by 2035. The same guidebook describes thousands of postings across Europe and the US unfilled in specialized roles such as data science and biomanufacturing, and calls the talent constraint structural rather than temporary.

The complication is that hiring is not expanding to meet it. NES Fircroft, a staffing firm, wrote on its own site on December 29, 2025 that life sciences job openings are decreasing, competition for roles is intensifying and candidates should expect longer searches, while employers remain cautious because the post-pandemic rebound did not materialize as hoped. In the same post, the firm says digital fluency in AI, data analytics and automation is becoming an essential skill.

A staffing firm reports life sciences job openings falling while employers ask for more digital fluency. Put next to the mandatory training program Langer describes, the picture is an industry that expects to upskill the people it already has rather than hire its way to AI readiness.

Caraballo's prescription matches that reading. He told listeners to start with curiosity, then find the courage to enroll in a class or take advantage of company training, and to spend time with the tools to learn what they can and cannot be trusted to do. Without sufficient upskilling, he said, the technology will hit a wall and will not be adopted in day-to-day work.

Langer added that the next generation needs deep domain expertise paired with digital tools, and that technical fluency plus strong critical thinking is what will set people apart. For a plant manager, that reads as a staffing spec. The operator who can question a model's output is the one who makes the model worth deploying.

Governance as the answer to complexity

Chew asked how leading organizations are moving from reactive compliance to a proactive, risk-based approach. Langer said industry, academia and regulators are all moving in that direction, and pointed to the FDA's Quality Management Maturity pilot, now under way, which in her description is meant to foster quality culture, recognize organizations already operating at a mature level and identify room to grow.

She also described her company's use of the ICH quality guidelines: pharmaceutical development and quality by design, quality risk management, an effective pharmaceutical quality system, and the lifecycle management strategies now available. None of the concepts are new, she acknowledged, but used together they give a framework for building quality in from the start.

Caraballo pushed back on the idea that governance is a burden. A company with one product and one customer can be run by its operations manager, he said, but as products, jurisdictions and customers multiply, the complexity demands rules to protect what is critical. Readiness, in his view, includes practicing what you have learned and challenging what you think you know, so that you can demonstrate execution under inspection.

Chew pressed on whether anyone is using AI to make quality risk management quantitative rather than a room of experts scoring a risk a four or a five. Caraballo called it an active area of investigation and set a test for the model's training data.

He added that operations change fast enough that the model has to be kept current, so it does not assume only three failure modes exist when a fourth simply has not happened yet, and that a useful model recognizes when a question falls outside its data. What AI adds, he said, is the ability to surface signals and relationships across volumes of data no human could review, which an expert then investigates. Caraballo's test hands a quality head two questions for any vendor: what failures the model was trained on, and how it flags the limits of what it knows.

The patient in the room

Asked what she most wants to hear in Washington, Langer named the patient testimonials. Every time she hears one, she said, it renews her energy for the work, and her hope is that attendees leave with that same sense of renewal alongside new knowledge.

Caraballo's advice for what to bring home was practical. Find the presenters who applied a technology and reported the lessons learned, ask them whether it worked, and bring the ideas back to use. The presentations he finds most interesting are the ones where a team tried something and says what happened.

The next milestone is October 18, when the meeting opens in Washington, DC, and Chew closed with a reminder to register and book hotel rooms before they sell out. The measure Caraballo offered for whether the industry's AI talk has become practice is how many presenters can report what went wrong as well as what went right.