Most companies that say they have a CI program don't. They have people gathering information. They have subscriptions to databases. They might have someone who forwards conference highlights. But that's not a program. That's activity without infrastructure, and activity without infrastructure produces insights that sit on a shelf.
The difference between CI activity and a CI program comes down to three things: a defined process, stakeholder alignment, and measurable output. I've built CI functions from scratch — inside organizations that had never thought systematically about competitive intelligence, and inside organizations that thought they had it covered when they didn't. The pattern is consistent. The ones that succeed share the same structural foundations. The ones that fail skip the same steps.
One of the challenges a lot of teams have today is how you define competitive intelligence: where does it start and where does it end? If companies define competitive intelligence as a news clipping exercise — desktop research, scanning headlines, forwarding articles — that program will be fragile. It might survive for a year, maybe two, but it won't earn its seat at the table. If you define CI as the analytical process of integrating different types of information from market and competitive sources to develop strategic knowledge, then you have a much broader remit. And a much stronger foundation.
This guide walks through the structural decisions, from assessing where you are today to choosing an organizational model, building the team, establishing the process, and creating the governance that keeps it running. Whether you're standing up a CI function in a 200-person biotech or restructuring one inside a global pharma organization, the fundamentals are the same.
For the broader CI framework these decisions sit within, see What Is Competitive Intelligence?.
Before building anything, figure out where you're starting from. Most organizations underestimate how much CI already happens informally — and overestimate how organized it is.
The CI maturity model tracks how functions develop from informal and ad-hoc to systematic and embedded. At the lowest level, CI happens incidentally: a sales rep mentions a competitor move in a meeting, someone forwards a news article, a medical science liaison picks up a data point at a conference. There's no system. No process for evaluating what it means. No feedback loop.
At the next stage, CI becomes reactive. Someone is assigned to "do CI," usually in addition to other responsibilities. They respond to requests, pull together information when asked, and maybe run a monitoring tool. The work is useful but project-based — it starts and stops based on immediate needs.
A mature CI function operates continuously. Intelligence requirements are defined in advance, collection is systematic, analysis follows structured methodologies, and outputs are integrated into decision-making forums. CI is embedded in the planning cycle, not bolted on.
The honest answer is that most organizations sit somewhere in the reactive middle. They have some CI capability, often stronger in certain therapeutic areas or business units than others, but lack the infrastructure to make it consistent.
Assess your starting point honestly, because it determines what to build first. If you're at the ad-hoc stage, start with intelligence requirements and a single high-priority deliverable. If you're reactive, focus on process standardization and stakeholder alignment. If you're already systematic, the opportunity is optimization — better technology, deeper analysis, stronger integration with strategy.
Not every asset in your portfolio needs the same level of CI investment. A CI intensity classification — based on commercial value crossed with competitor threat level — helps allocate resources where they matter most.
High commercial value plus high competitive threat demands dedicated, continuous CI. A Phase III asset approaching launch in a crowded indication with multiple competitor readouts expected — that's the highest intensity. An early-stage asset in a less competitive space with lower commercial potential can be monitored with lighter coverage.
This classification prevents the common mistake of spreading CI resources evenly across the portfolio instead of concentrating them where the competitive stakes are highest. When I've helped organizations implement this, the conversation shifts from "we need more CI" to "we need CI in the right places" — which is a much more productive starting point.
Three dominant organizational models exist for CI functions. The choice says a lot about where CI sits in the decision-making hierarchy, and it has real consequences for what the function can deliver.
A dedicated CI team organized by business unit or therapeutic area, reporting to a CI Head. Companies like a major biotech company, a top-10 pharma company, and a top-10 pharma company run versions of this model.
The advantages are real: scale, specialization, standardization of methods, clear career paths for CI professionals, and a degree of analytical objectivity that comes from sitting outside any single brand team. The CI Head can champion competitive readiness across the organization and drive best-practice sharing.
The head of a CI Center of Excellence at a top-10 pharma company, describes the scope this way: "We work with teams from the bench to the board." That's the promise of centralized CI — intelligence covering the full spectrum, from early-stage molecule assessment through commercial strategy.
The risk is creating barriers between the CI function and the business units it serves. A centralized team that's excellent at analysis but distant from decision makers doesn't earn its seat at the table. A percentage of time inevitably goes to functional work — internal meetings, methodology development, reporting — that doesn't directly serve the brand teams.
A CI professional embedded in each product team or business unit, reporting to the insights lead or BU head. Companies like a major pharma company, AbbVie, and a top-10 pharma company (in certain divisions) have run this model.
The advantage is proximity. You're closer to decision makers, you understand the brand context, and you can be more agile. There's less CI bureaucracy. The embedded analyst can influence product strategy directly because they're in the room when decisions are made.
The disadvantages are structural. Without a common CI culture, there's no standardization, no best-practice sharing, and limited career progression for CI professionals. Each embedded person develops their own approach, which makes quality inconsistent across the portfolio. And because they report to the brand team, objectivity can be harder to maintain — the CI professional may feel pressure to produce intelligence that supports the team's existing assumptions rather than challenging them.
A single generalist insights manager combines CI with market research, business analytics, and forecasting. Companies like Shire, Celgene, and a major pharma company have used this approach.
You get the holistic brand view — the same person who understands the market research also understands the competitive dynamics. End users receive integrated insights rather than siloed deliverables.
But the risk is losing CI specialization. The insights manager usually gravitates to the area where they have the most expertise — typically market research — and competitive intelligence becomes an afterthought. There's no champion for competitive readiness, no one pushing the organization to think systematically about competitive threats. Jack of all trades, master of none.
The operating model matters less than two things: executive sponsorship and a competitive mindset.
If the CEO and senior leadership aren't actively driving a competitive mindset, CI will always be treated as a support function rather than a strategic input. You can have the best tools, the best analysts, and the best processes, but if competition isn't part of how the organization thinks and makes decisions, insights will sit on a shelf.
The right model depends on your organization's size, culture, and maturity. A small biotech with 200 people doesn't need a centralized CI team — it needs one sharp CI professional embedded in the leadership team with access to external specialists. A global pharma with 20 therapeutic areas probably needs centralized coordination with some degree of local embedding. The model should serve the intelligence needs, not the organizational chart.
A CI function needs at minimum three role types, though in smaller organizations one person may cover more than one:
CI Director or Manager. Owns the CI strategy, manages stakeholder relationships, sets intelligence requirements with leadership, and ensures CI is integrated into decision-making. This role requires deep industry knowledge, analytical capability, and — critically — the interpersonal skills to influence executives and earn the function's seat at the table.
CI Analyst. Executes the collection plan, conducts secondary and primary research, develops competitive landscapes, produces deliverables, and supports analytical frameworks. Strong research skills, ability to synthesize information from diverse sources, and domain expertise in relevant therapeutic areas.
CI Coordinator. Manages operational logistics — vendor relationships, source databases, knowledge management systems, event planning, distribution lists. This role is often undervalued but essential. Without someone keeping the machinery running, even talented analysts spend their time on administration instead of analysis.
In larger organizations, you'll add specialized roles: collection managers for conferences, primary research leads, technology specialists for AI-powered monitoring. But the three-role foundation covers most CI operating needs.
CI requires a specific combination of capabilities that's hard to find in a single person. The core skills: deep industry knowledge, strong secondary research capability, analytical thinking, interpersonal ability for elicitation and networking, and communication skills. That last one matters more than most CI teams admit. The ability to translate complex competitive data into a clear, concise briefing that changes an executive's decision — that's what separates a CI professional from a researcher.
CI investment should scale with the lifecycle stage and competitive intensity of your portfolio. From the programs I've built and managed, these ranges hold across most pharma organizations:
Phase I and Phase II: $100K–$300K annually. 0.1–0.3 FTE dedicated to CI. At this stage, you need news monitoring, competitive landscapes, timelines, clinical trial trackers, target product profile comparisons, and conference coverage. Primary research is limited. The focus is establishing the competitive baseline.
Phase III and Launch: $300K–$1M+ annually. 0.3–1 FTE. This is where CI investment should peak. You add strategic workshops, competitive threat assessments, competitive simulations, scenario planning, and rapid response capability. Early warning becomes critical. Internal key meetings, guideline monitoring, and building the internal network intensify.
Growth and Loss of Exclusivity: $200K–$500K annually. 0.2–4 FTE (depending on portfolio complexity). Focus shifts to competitor lifecycle management, indication prioritization, and generics or biosimilar CI — manufacturing intelligence, regulatory routes, vertical structures, legal basis analysis.
These aren't aspirational numbers. They're benchmarks from working CI programs. If your organization is spending less than $100K on CI for a Phase III asset in a competitive indication, you're underinvesting. The cost of CI is measured in thousands. The cost of not doing it — missed competitive signals, flawed launch strategy, wrong positioning — is measured in tens of millions.
For organizations outside pharma, the principle is the same even if the dollar amounts differ: CI investment should correlate with competitive intensity and the magnitude of the decisions it informs.
The CI process follows a disciplined, cyclical structure. It's not linear — it loops back continuously as new intelligence reshapes requirements and collection priorities.
Intelligence Requirements → Collection Plan → Secondary Collection → Human Collection → Processing & Analyzing → Final Analysis → CI Review → Strategy — and the loop starts again.
Everything starts here. A successful CI process begins and ends by understanding the company or product strategy. Intelligence needs are expressed as Key Intelligence Topics (KITs) and Key Intelligence Questions (KIQs). KITs are the big-picture strategic issues. KIQs are the specific questions that must be answered to address each KIT.
Properly defining requirements demands effective stakeholder interaction and an experienced CI team. The CI professional's job is to anticipate decision-making needs — not just transcribe what stakeholders ask for. A brand director who says "tell me what the competitor announced at the conference" may actually need to know "how does the competitor's data change our positioning?"
KIQs get prioritized: high priority means dedicated human source collection, medium means secondary sources with human follow-up, low means report as received. Review KITs every six months with senior leadership, quarterly with the operational team. This isn't optional — intelligence requirements that don't evolve with the business become irrelevant.
For the full KIT/KIQ methodology, see Key Intelligence Questions: The Foundation of Any CI Program.
With intelligence needs defined, review what you already know to identify gaps. The collection plan determines who will gather what, when, where, and how. It considers the right mix of sources, redundancy for critical questions, cueing strategies, and cost-effectiveness.
Build the collection plan on your understanding of the target environment and past experience. It should specify source types — from investigators and KOLs to corporate and government sources — and identify which events, databases, and human networks will be used.
A great deal of useful CI can be obtained from public sources — the internet, journals, commercial databases, regulatory filings. The goal is not to gather copious amounts of information but to produce highly relevant analysis. Public domain data establishes an accurate baseline, guides further collection, and often yields findings that scanning-and-clipping efforts miss.
But secondary research is available to everyone. It's necessary — essential, even — but it's not sufficient. CI uses secondary research primarily to develop targeting data: leads on sources, organizations, individuals, and events that will feed the primary collection effort.
This is where CI becomes an art. After mining public information, systematic collection of non-public information provides competitive advantage through superior knowledge. Non-public information comes primarily through direct or indirect questioning of people — what the intelligence community calls HUMINT.
What really is unique to competitive intelligence is human judgment and the ability to gather non-public information by talking to people. AI won't replace that. Technology won't replace that. Relationships will always be the differentiator. The source network — contacts linked by trust and reciprocity — is the CI function's most valuable asset.
Processing converts raw information into a form useful for analysis. Analysis then converts processed information into intelligence that answers the company's questions — combining discrete data points to produce knowledge that no single source could provide.
Outputs should be clear, concise, and organized to facilitate integration into decision-making. Decision makers value brevity over volume. And dissemination must be targeted — intelligence should go only to individuals who need to act on it. Over-distribution erodes perceived value and increases the risk of strategic information leaking.
A structured CI program produces six core deliverable types:
Not every organization needs all six from day one. Start with KITs/KIQs and one or two deliverables that address your most urgent intelligence gaps, then build out systematically.
CI infrastructure requires more than people and processes. You need systems to store, retrieve, and protect intelligence.
At minimum, a CI function needs three things: a source database (tracking who you've spoken with, what they know, their reliability, and contact history), an electronic filing system organized by product, therapeutic area, and intelligence category, and a secure distribution mechanism for sensitive intelligence.
The filing structure matters more than the platform. Organize by company, therapeutic area, indication, and product — not by date or project. The analyst who needs information six months from now has to be able to find it without depending on the person who filed it.
Three years ago, we had people doing news monitoring who would spend probably about three hours per day logging into email, Google Alerts, different access feeds and different databases. All of that can now be centralized and processed through AI-powered platforms.
The shift is real. AI handles the monitoring layer — source screening, news summarization, initial categorization. But the technology is not the differentiator. It's what the CI team does with the time it gets back. According to a senior strategy executive at a top-10 pharma company from the pharma company's Enterprise Strategy put it: "I want AI to do the data gathering so I can spend time on strategic thinking."
of a mid-size pharmaceutical company describes AI as an equalizer for smaller organizations. Before AI, his team relied heavily on external vendors for competitive analysis. Now they can establish strategy and test different assumptions at a fraction of the cost. That's the democratization effect — CI capabilities that used to require a large team and significant budget are becoming accessible to organizations that could never have afforded them.
The practical question for any CI function building its technology stack: what can be automated (standard monitoring, source screening, initial categorization) versus what requires human judgment (strategic interpretation, KOL engagement, executive influence)? Get that boundary right, and technology amplifies the function. Get it wrong, and you're spending money on tools that produce noise instead of intelligence.
For guidance on selecting and evaluating CI technology, see CI Software and Tools.
A CI program without governance is a liability. This isn't optional — it's what separates professional intelligence from ad-hoc information gathering.
Every CI function needs documented policies covering information security, distribution controls, ethical guidelines, and legal compliance. The compliance manual should address: how information is classified (public, internal, restricted), who has access to what, how sensitive intelligence is stored and transmitted, and what happens when someone encounters a grey area.
CI is legal when conducted ethically using publicly available information and voluntary conversations. It is explicitly distinct from industrial espionage. But laws differ significantly by jurisdiction — bribery statutes, confidentiality protections, data protection regulations, and interception laws all vary across the UK, US, EU, and beyond.
Two practical self-tests cut through the ambiguity. The "grandmother test": would you be comfortable explaining to your grandmother what you did to get this information? The "Wall Street Journal test": would your company want this method reported on the front page of the Wall Street Journal?
Some basic minimum guidelines: never lie when collecting intelligence. Don't use information clearly labeled confidential. Protect source privacy. Outsourcing collection to a vendor doesn't eliminate your responsibility. And recognize that not having a CI program could itself be considered a failure of fiduciary responsibility — management has an obligation to be aware of the competitive environment in which the firm operates.
CI ethics and compliance aren't one-time activities. Build them into the onboarding process for new CI staff, create annual refresher training, and extend awareness to the broader organization. Employees and strategic partners need to understand not only the value of CI but also the importance of protecting the company's own sensitive information — because your competitors are running their own collection programs.
This is where most CI programs succeed or fail. Not on the quality of the analysis, but on whether anyone acts on it.
CI is fundamentally a human process that relies on a network of people within the company to be successful. Optimally, CI should have strong support from an internal champion at the top of the company who can influence acceptance and support.
This awareness is needed at all levels, but is particularly vital among upper management since executive input is required for intelligence efforts to be proactive and optimized to support key decisions. In an effective CI system, senior management demands intelligence input to support decision-making and planning, and a high level of interest and appreciation for CI is clearly visible at the executive level.
Without that, you're pushing rope. I've seen technically excellent CI functions produce brilliant analysis that changed nothing — because no one with authority had asked for it or was waiting to act on it.
One of the most practical lessons I've learned: don't mention the KIT. When engaging decision makers, use their vocabulary, not CI jargon. A brand director doesn't care about your Key Intelligence Topics. They care about whether their launch strategy will survive contact with the competitor's Phase III data.
Calibrate to the client's leadership style. Know your customer. Some executives want the detailed analysis. Others want one slide and a recommendation. Some want to debate the implications. Others want you to tell them what to do. The CI function that can flex across these styles builds credibility faster than the one that delivers the same format to everyone.
The importance a firm attaches to CI will be signaled to employees by the extent to which it is recognized through formal and informal rewards — recognition, basic compensation, incentive compensation, promotions. If CI contribution appears in job descriptions and performance evaluations, people contribute to it. If it doesn't, they won't.
Build CI into the planning calendar: brand planning meetings, portfolio reviews, launch readiness assessments. Make CI a standing agenda item, not a special topic. The more CI becomes part of how the organization operates — rather than something the CI team does separately — the more it influences decisions.
It's worth investing time in a CI mission statement, values statement, and vision statement. Not as a branding exercise, but as a tool for alignment. When the CI team can articulate what it exists to do — for example, acting as a lighthouse that foresees competitive changes and enables the organization to act before they arrive — that clarity helps stakeholders understand what to expect and what to demand.
If you can't measure it, you can't improve it. And more importantly, you can't defend the budget.
Six dimensions of intelligence quality provide the foundation for measurement:
Track turnaround time by deliverable type, cost per deliverable, and coverage against the collection plan. The coverage assessment matrix — mapping intelligence requirements against actual coverage — reveals blind spots before they become problems.
Survey key stakeholders annually on CI relevance, quality, and impact. But go beyond satisfaction scores. The real question is: can the CI function point to specific decisions that were influenced by intelligence it provided? Track those decision impacts over time. That's your ROI story.
For a comprehensive measurement framework, see Measuring CI Impact and ROI.
Having built CI programs across different organizational contexts, I see the same failure patterns:
Starting with technology instead of requirements. Buying a CI platform before defining what intelligence you need is like buying a telescope before deciding what you're looking for. Start with KITs/KIQs, then choose technology that supports those requirements.
Underinvesting at launch. The Phase III and launch window is where CI has the highest impact — and where most organizations underinvest. If you're spending $100K on CI for a billion-dollar launch, that's not prudent cost management. That's a strategic gap.
Treating CI as a research function. CI produces intelligence, not research. The difference is actionability. If your CI team is delivering 50-page reports that no one reads, the problem isn't the quality of the research. It's the absence of the "so what."
No feedback loop. A CI function that delivers intelligence but never learns whether it was accurate, useful, or acted upon will never improve. Build feedback into the process — after every major deliverable, after every competitive event, at every planning cycle.
Ignoring internal intelligence. 50% of all your CI needs can be found within your own company — you just don't know where it is. The field force, the MSLs, the medical affairs team, the regulatory group — they all interact with the competitive environment daily. A CI function that doesn't systematically tap internal knowledge is working with half the picture.
Expect 6–12 months to establish the foundations: intelligence requirements, organizational model, collection plan, initial deliverables, and stakeholder alignment. A fully mature CI function — embedded in decision-making, running all six core deliverables, with a trained internal network — takes 18–24 months. The CI maturity model isn't something you skip through. Each stage builds on the previous one.
It depends on the lifecycle stage and competitive intensity of your portfolio. Benchmarks from working CI programs: $100K–$300K during Phase I and II (0.1–0.3 FTE), $300K–$1M+ during Phase III and launch (0.3–1 FTE), $200K–$500K during growth and loss of exclusivity (0.2–4 FTE). These include staff, vendor costs, technology, and primary research. Organizations outside pharma can start smaller — a single dedicated CI professional plus focused vendor support for the highest-priority competitive questions.
It depends on your organization's size, culture, and maturity. Centralized CI gives you scale, specialization, and standardization. Decentralized puts you closer to decisions. Integrated gives the full brand picture but risks losing CI depth. The model matters less than executive sponsorship and a competitive mindset. If competition isn't part of how the organization thinks, no organizational model will fix that.
Define your Key Intelligence Topics and Key Intelligence Questions with key stakeholders. Review what you already know to identify gaps. Then develop a collection plan for the highest-priority questions. Don't try to build the entire function on day one — start with one high-value deliverable that demonstrates CI's impact on a specific decision.
Start by solving a problem they already have. Find the decision maker who's facing a competitive question without good intelligence, and deliver a CI product that directly informs that decision. One high-impact deliverable creates more buy-in than a hundred presentations about CI methodology. Then connect that success to a broader CI mandate. Executive sponsorship grows from demonstrated value, not from theoretical arguments.
For a founding CI hire, prioritize industry knowledge and interpersonal skills over research skills. You can outsource secondary research. You can buy monitoring tools. What you can't outsource is the ability to understand the business context, anticipate stakeholder needs, and build the relationships that make CI actionable. The first CI hire should be someone who can sit in a brand planning meeting, identify the unasked competitive questions, and deliver intelligence that changes the conversation.
AI doesn't change the fundamentals — you still need defined requirements, a collection plan, human sources, and stakeholder integration. What AI changes is the resource allocation. Monitoring and data processing that used to require significant FTE can now be automated, which means the same budget buys more analytical and strategic capacity. For smaller organizations especially, AI is an equalizer — it makes CI capabilities accessible that previously required a large team and a large budget. But AI doesn't replace the need for human judgment, source relationships, or strategic influence. Those remain the CI function's core assets.
Salvador Carlucci is the founder of the Atacana Group and has spent over two decades building competitive intelligence functions for pharmaceutical and healthcare companies worldwide. He is co-author of the seminal articles on CI in biotech published in Nature Biotechnology.
© CompetitiveIntelligence.com