Business and Market Research
Competitive intelligence research workflow: From Scattered Sources to a Cited Brief
Build a repeatable competitive intelligence research workflow for collecting competitor sources, tracking contradictions, and producing a defensible cited brief.

A competitive intelligence research workflow should produce more than a competitor profile. It should move from a defined decision question to a traceable source corpus, an evidence matrix, analysis, a cited brief, and a scheduled refresh cycle.
The point is not to collect every public mention of a rival. The point is to build a defensible answer to a strategic question: what changed, how confident are we, what evidence supports it, and what still needs checking?
Updated November 19, 2026.
What a competitive intelligence research workflow should produce
Competitive intelligence is the systematic gathering, monitoring, analysis, and dissemination of external information with strategic value, as San José State University’s School of Information defines it in its overview of competitive intelligence. VCU’s research development guide similarly frames CI as ethically collecting, analyzing, and disseminating information to support strategic decisions through actionable insights (VCU Scholars Compass).
That definition matters because it separates competitive intelligence from three things it is often confused with:
A competitor profile: useful background, but usually static.
A market summary: helpful context, but often too broad for a decision.
A due diligence review: deeper investigation of a specific entity, transaction, or risk event.
Competitive intelligence monitors markets, competitors, positioning, products, hiring, partnerships, customer signals, and strategic moves over time. Due diligence asks a more specific risk question about a company, asset, vendor, investment, or transaction. If that is the job, use a separate due diligence research workflow rather than stretching CI into entity-risk investigation.
A strong CI workflow should produce six outputs:
A decision-linked intelligence question
The brief should support a real choice: pricing, market entry, product strategy, partnership selection, investment analysis, sales positioning, or executive planning.
A traceable source corpus
Selected webpages, PDFs, filings, reports, transcripts, videos, podcasts, spreadsheets, and notes should sit in one working library with source dates and access dates.
An evidence matrix
Each material claim should be tied to the exact evidence behind it, its source type, confidence level, and contradiction status.
A cross-competitor analysis
Competitors should be compared across the same dimensions, not summarized in isolated profiles.
A cited brief
The reader should be able to move from claim to source without guessing where an assertion came from.
A refresh plan
Competitive intelligence decays. Product pages change, hiring signals disappear, filings age, and messaging shifts. The brief needs a maintenance schedule.
This workflow works beside discovery tools, market-intelligence databases, analyst research, monitoring platforms, and human source evaluation. Search and monitoring help find signals. The research workflow decides what belongs in the evidence base, how much weight it carries, and what conclusion it can support.

Define the intelligence question before collecting competitor research sources
The fastest way to ruin a CI project is to start collecting sources before defining the decision.
A broad instruction like “research our competitors” creates a folder full of tabs, screenshots, PDFs, and half-paraphrased notes. It usually ends with a polished but fragile summary: confident language, weak provenance, and unclear relevance.
Start with the decision.
Useful CI questions look like this:
Product positioning: Which competitors are repositioning toward enterprise buyers, and what public evidence supports that?
Market entry: Which adjacent players appear to be entering the UK mid-market segment?
Pricing strategy: Which competitors have changed packaging, discounting, or usage limits in ways that affect our pricing?
Partnership selection: Which ecosystem partners are expanding with our competitors, and where are exclusivity risks visible?
Investment analysis: Which public signals suggest a competitor is reallocating capital toward a specific geography, product line, or customer segment?
Executive planning: Which competitor moves matter enough to affect next quarter’s roadmap or sales narrative?
Then define the comparison set. Include obvious direct competitors, but do not stop there.
A serious comparison set may include:
Direct competitors selling similar products to similar buyers.
Adjacent alternatives solving part of the same problem.
Emerging entrants with credible funding, hiring, or product momentum.
Substitutes that change customer behavior without looking like category peers.
Partners or platforms that could become competitors through bundling or integration.
Set the dimensions before searching. Otherwise the loudest sources will shape the brief.
Common dimensions include:
Product capabilities
Customer segments
Pricing and packaging
Distribution channels
Partnerships
Hiring and organisational changes
Messaging and positioning
Geographic expansion
Financial performance
Regulatory exposure
Strategic priorities
Customer proof and implementation patterns
Also set the boundaries: time period, geography, source cutoff, intended audience, and required confidence level. A board memo about market entry needs a different evidence threshold than a quick sales enablement update.
A compact evidence plan is enough. For example:
Decision: Should we adjust enterprise positioning in the Canadian market?
Question: Which competitor moves indicate a shift toward enterprise customers in Canada, and what public evidence supports that interpretation?
Competitors: three direct rivals, two adjacent platform vendors, one emerging entrant.
Dimensions: packaging, enterprise features, partner activity, hiring, customer stories, and messaging.
Period: last 18 months, with older evidence allowed only for baseline context.
Output: cited brief for product and sales leadership, with confidence levels and open questions.
That plan keeps the research from silently mixing a current product page, a two-year-old conference quote, and a speculative news article into one undifferentiated conclusion.
Build a balanced source-collection process across public evidence
Competitive intelligence depends on source diversity. A competitor’s homepage tells one story. Documentation, filings, hiring pages, customer case studies, earnings calls, webinars, podcasts, and partner announcements often tell a more complicated one.
Start with first-party competitor material:
Websites and landing pages
Product pages
Documentation and help centers
Changelogs and release notes
Customer stories
Pricing or packaging pages
Support articles
Hiring pages
Press releases and public announcements
Preserve the page URL and capture date. Web content changes, and the access date often becomes important when a claim later disappears or shifts.
Then collect formal disclosures where relevant:
Annual reports
Quarterly filings
Investor presentations
Regulatory disclosures
Earnings-call transcripts
Prospectuses or offering documents
Public grant, procurement, or contract records
For analyst and financial work, filings and transcripts deserve special treatment because they are dense, source-rich, and easy to over-summarize. A source-cited workflow for analyzing SEC filings with AI is useful when financial evidence is central to the competitive question.
Add market context:
Analyst and market reports
Trade publications
Reputable news
Conference presentations
Webinars
Podcasts
Executive interviews
Industry association material
Then add customer, partner, and ecosystem evidence:
Implementation stories
Partner announcements
Procurement documents
Case studies
Public reviews
Job postings
Marketplace listings
Integration directories
Public statements from customers or resellers
Label relationships and incentives. A customer case study approved by a vendor is not the same kind of evidence as an independent user complaint, a regulator’s document, or a procurement award notice.
Use multiple formats deliberately. CI sources rarely arrive in one neat format. A competitor’s strategic move may be visible across a product page, a PDF investor deck, a podcast interview, a hiring page, and a customer webinar. Those formats should enter the same working corpus rather than staying scattered across browser tabs, folders, transcripts, and personal notes.
Discovery and evidence collection are separate stages. Search engines, market-intelligence databases, paper discovery tools, analyst portals, and monitoring software can help locate candidates. The corpus is where selected evidence is retained, evaluated, and compared.
Do not collect forever. Stop when each comparison dimension has adequate evidence, the remaining gaps are explicit, and additional sources are unlikely to change the decision. “We found 200 sources” is not a better outcome than “we found 37 decision-relevant sources and know exactly where the gaps are.”

Record provenance and assess source quality before synthesis
Every source in the corpus needs a record. At minimum, capture:
Title
URL or file reference
Publication date
Date accessed
Author or organisation
Source type
Competitor involved
Geography
Time period described
Relevant comparison dimension
Then capture the exact passage, page, timestamp, table, quotation, or figure that supports each finding. A paraphrased note is not enough when the conclusion will guide strategy.
Source evaluation should happen before synthesis, not after the executive summary is written. Harvard’s Guide to Using Sources puts the basic rule plainly: before relying on a source, evaluate whether it is appropriate for the work (Harvard Guide to Using Sources).
For CI, source quality is not just “credible” or “not credible.” The better labels are more specific:
Direct observation: a product page, documentation page, changelog, filing, or pricing page showing the thing itself.
Company claim: a press release, investor deck, executive interview, or marketing page.
Independent reporting: reputable journalism or third-party reporting with named sources or documents.
Customer or partner evidence: case studies, procurement records, implementation material, reviews, or partner announcements.
Analyst interpretation: market research, analyst commentary, benchmark reports, or expert notes.
Unverified signal: hiring pattern, social post, forum report, review trend, or indirect clue that may deserve follow-up.
Record incentives and likely bias. A competitor’s product page is useful for understanding claims and positioning, but it is designed to persuade. An investor presentation may highlight growth and downplay risk. A customer case study may be approved by both sides. A paid analyst report may be rigorous, but the scope and assumptions still matter.
Separate publication date from event date. A newly published article may describe a product launched two years ago. An updated webpage may preserve old claims. A job posting may reflect future intent, current hiring, or an evergreen requisition.
Preserve context around every citation. AI-generated summaries and cited answers can help locate relevant passages, but the underlying wording, scope, caveats, and surrounding paragraphs still need human checking. A sentence that looks decisive in isolation may be qualified two paragraphs later.
Organise the corpus and build an evidence matrix
A durable CI corpus needs structure that survives follow-up questions.
Use folders, spaces, or tags for:
Competitors
Source types
Comparison dimensions
Time periods
Evidence status
Geography
Confidence level
Open questions
Contradiction status
Then build an evidence matrix. One row should represent one meaningful claim or observation, not one source.
Useful fields include:
Competitor
Dimension
Claim or observation
Exact evidence
Source
Date published
Date accessed
Source type
Evidence strength
Incentive or bias
Interpretation
Confidence
Contradiction status
Next verification step
For a reusable structure, see this guide to building an evidence matrix from multiple sources. The same logic applies to CI: the row unit is the evidence-backed claim, not the document.
Keep facts, interpretations, and hypotheses separate.
A job posting for enterprise account executives in Germany is an observation.
A possible expansion into the German enterprise segment is an interpretation.
A predicted enterprise product launch next quarter is a hypothesis.
Those three statements require different language and different confidence levels.
The matrix also prevents a common CI failure: isolated competitor summaries. If competitor A is summarized by product features, competitor B by customer logos, and competitor C by funding history, the reader cannot compare like with like. Use the same dimensions across the comparison set, even when evidence is uneven.
Track absence explicitly. A missing public disclosure is not proof that an activity does not exist. A private company with fewer public sources should not automatically appear less active than a public company with filings, calls, and analyst coverage.
Most importantly, record contradictions instead of flattening them.
Examples:
A product page claims broad enterprise capability, but documentation shows limited availability.
An executive says the company is not focused on a geography, but hiring and partner announcements suggest activity there.
A customer case study implies deep integration, but support docs show the feature is in beta.
A market report describes category leadership, while recent customer reviews show implementation friction.
A good matrix keeps those tensions visible. For broader comparison logic, the method in how to compare sources in a literature review transfers well: compare evidence, themes, convergence, contradiction, and significance rather than stacking summaries.

Use a multi-source research workspace without outsourcing judgment
This is where Otio fits: after discovery, as a permanent multi-source research workspace for the selected evidence.
A CI project usually contains webpages, PDFs, filings, reports, spreadsheets, videos, podcasts, transcripts, and researcher notes. Keeping those in separate tools creates predictable failure modes: lost capture dates, missing quotations, duplicate sources, untraceable summaries, and a painful rebuild when an executive asks a follow-up question three weeks later.
In Otio, the useful workflow is straightforward:
Add competitor webpages, reports, PDFs, transcripts, videos, podcasts, and notes into one library.
Organise them by competitor, dimension, source type, and evidence status.
Ask cited questions across the corpus.
Pull exact passages into notes.
Compare claims across sources.
Draft a cited brief.
Review every material answer against the original source.
The point is not to let AI decide the strategy. The point is to reduce the mechanical cost of finding, comparing, and citing evidence.
For example, a researcher can ask:
Which sources mention enterprise packaging changes for competitor A?
Compare competitor A and competitor B on partner strategy using only sources from the last 12 months.
Which sources contradict the claim that competitor C is expanding into healthcare?
What evidence is missing for pricing changes across the comparison set?
Draft a brief section on geographic expansion with citations to source passages.
Otio’s webpage summarizer is useful for turning competitor pages and public announcements into reviewable source notes. Its AI PDF reader helps with filings, investor decks, analyst reports, and long PDFs where page-level evidence matters. Its multiple AI models support different analytical passes over the same corpus, such as one model extracting claims and another challenging the interpretation.
Use that multi-model capability carefully. Agreement between models is not verification. It is a prompt to inspect the underlying sources more efficiently.
For financial and market work, Otio’s AI for financial analysis workflows are most useful when filings, earnings transcripts, and investor materials are part of the CI corpus. They do not remove the need to check definitions, reporting periods, non-GAAP adjustments, and management incentives.
Set the boundary clearly. Otio is not a complete real-time competitor monitoring system. It does not discover every relevant source, replace market-intelligence databases, replace analyst research, or remove the need for human source evaluation. It is the workspace where selected evidence becomes a traceable analysis.
[[OTIO_INLINE_PROMO:%7B%22title%22%3A%22Ready%20to%20compare%20your%20CI%20sources%20in%20one%20corpus%3F%22%2C%22description%22%3A%22Bring%20selected%20webpages%2C%20PDFs%2C%20transcripts%2C%20videos%2C%20and%20notes%20into%20one%20library%2C%20then%20ask%20cited%20questions%20while%20checking%20passages%20and%20contradictions%20yourself.%22%7D]]
Turn the evidence matrix into a cited competitive intelligence brief
A CI brief should not read like a dump of research notes. It should answer the decision question, show the evidence path, and expose uncertainty where it matters.
Start with an executive takeaway. This should include:
The answer to the decision question
The most material competitive signals
What changed or appears to be changing
Which claims are high confidence
Which claims are interpretations
Which open questions affect the decision
Then add a short methodology section. Keep it concise, but include enough detail for the reader to judge scope:
Competitors included
Comparison dimensions
Time period covered
Geography
Source types reviewed
Search or collection boundaries
Evidence cutoff date
Important gaps
Next, present competitor-by-competitor findings using the same dimensions for each competitor. This is where many briefs weaken. If one competitor gets a rich narrative and another gets three bullets because there were fewer sources, the comparison may reflect public-source availability more than competitive reality.
After competitor sections, add a cross-source comparison. This is the analytical core.
Look for:
Convergence: multiple source types support the same conclusion.
Contradiction: sources disagree on scope, timing, availability, or strategic intent.
Asymmetry: one competitor is stronger on distribution, another on product depth, another on pricing flexibility.
Change over time: messaging, hiring, release notes, and customer proof point in a new direction.
Missing evidence: the absence of support for a claim that the organisation has been assuming.
Signals needing follow-up: weak but potentially material clues.
Assign confidence levels to material findings. Citations improve traceability, but they do not make a claim objective. The University of Southern California’s research guide defines a citation as a formal reference to a source consulted and used in writing (USC Libraries). That formal reference tells readers where a claim came from. It does not settle whether the source selection was balanced, whether incentives shaped the evidence, or whether contradictory evidence was missed.
Use confidence language like this:
High confidence: supported by direct first-party evidence and at least one independent or corroborating source; current; few contradictions.
Moderate confidence: supported by credible evidence, but with limited source diversity, date issues, or partial contradictions.
Low confidence: based on indirect signals, one-sided claims, old evidence, or unresolved contradictions.
Hypothesis: plausible interpretation that needs more verification before being used as a decision premise.
Cite the exact material behind each important claim. Depending on the source, that may be:
Webpage URL and access date
Report page
Filing section
Transcript passage
Video timestamp
Podcast segment
Table or figure
Quoted passage
Archived page reference
End with implications, open questions, and recommended decisions. Do not turn a weak signal into a forecast. If the evidence supports “competitor A appears to be testing enterprise packaging,” write that. Do not write “competitor A will dominate enterprise accounts next year” unless the evidence can carry the claim.
For the adjacent writing workflow, use this guide to creating a source-linked research brief.

Maintain the brief with refreshes, change logs, and explicit uncertainty
A CI brief is not finished when it is delivered. It becomes stale at different speeds depending on the source type and the decision it supports.
Set refresh triggers based on the decision and market. Do not assume every source needs the same schedule.
A practical refresh plan might separate:
Product pages and documentation: review when product strategy, sales enablement, or feature comparison matters.
Pricing and packaging pages: review before pricing decisions, renewal campaigns, and board updates.
Hiring pages: review when expansion, capability-building, or market-entry signals matter.
Filings and earnings calls: review on reporting cycles.
Webinars and conferences: review around major industry events.
News and analyst commentary: review when material market events occur.
Partner pages and marketplaces: review when ecosystem strategy is central.
Keep the original evidence and add new observations as dated entries. That distinction matters. A changed fact is not the same as a changed interpretation.
Use a change log with entries such as:
New sources added
Pages removed, changed, or archived
Claims revised
Contradictions resolved or newly discovered
Confidence levels changed
Evidence gaps closed
Evidence cutoff date updated
Decisions affected
Before reusing an older brief, recheck high-impact claims against current first-party material and independent evidence. A six-month-old source may still be useful background, but it should not silently function as current evidence.
Label each statement in the maintained brief:
Observed fact
Interpretation
Forward-looking hypothesis
Also include the evidence cutoff date prominently. Readers should know whether the brief reflects sources reviewed through last week, last quarter, or last year.
Scheduled refreshes are not complete real-time monitoring. If the organisation needs alerts, coverage guarantees, CRM integrations, battlecard automation, or specialist monitoring workflows, keep that function in a dedicated monitoring system. Feed selected evidence into the research corpus when it becomes decision-relevant.
The operating principle is simple: discovery tools find signals, monitoring tools detect change, and the CI research workspace turns selected evidence into a cited, reviewable brief.
Build your next competitive intelligence corpus in Otio, bring your web pages, reports, videos, and notes into one workspace, and generate a cited first draft you can review against the original sources.
FAQ
Q: How is competitive intelligence different from competitor monitoring software?
A: Monitoring software helps detect changes or mentions. Competitive intelligence turns selected evidence into a decision-focused interpretation with provenance, confidence levels, caveats, and human review.
Q: Can a competitive intelligence brief be objective if every claim has a citation?
A: No. Citations improve traceability, but objectivity also depends on source selection, publication date, incentives, missing evidence, contradictory findings, and the distinction between observation and interpretation.
Q: What should a team do when competitor sources disagree?
A: Keep the contradiction visible in the evidence matrix, compare source dates, incentives, scope, and directness, then assign a confidence level. If the disagreement affects the decision, record the next verification step.
[[OTIO_FOOTER_PROMO:%7B%22title%22%3A%22Apply%20this%20workflow%20to%20your%20own%20source%20set%22%2C%22description%22%3A%22Upload%20or%20link%20your%20competitor%20evidence%2C%20organize%20it%20by%20dimension%2C%20and%20use%20cited%20answers%20to%20draft%20a%20brief%20you%20can%20verify%20against%20the%20originals.%22%7D]]




