Choosing an intelligence tool: the 7 criteria that matter

The intelligence tool market has exploded since 2023. Between established enterprise platforms (Meltwater, Talkwalker, Signal AI), information aggregators (Feedly, Inoreader), AI-native new entrants (Perplexity and similar) and specialised sector solutions, a consulting firm can face more than thirty credible options, with few frameworks to compare them.

The trap is choosing the most powerful tool rather than the most appropriate one. A 12-person firm has different needs from a large bank's economic intelligence department. The criteria that matter for one are often irrelevant to what matters for the other.

Here are the seven criteria we have identified as decisive for mid-sized consulting firms, and the weighting to give them depending on your context.

Criterion 1: source coverage

This is the most obvious criterion and yet one of the most poorly evaluated during trial phases. The question is not "how many sources does the tool index?", enterprise platforms claim millions of sources, a figure that means almost nothing for a sector-focused consulting firm. The question is: does the tool cover the sources that matter for my intelligence profiles?

A concrete evaluation grid: take your five most important sources for each of your two or three main profiles. Check whether the tool indexes them correctly, with what latency, and whether full content or only headlines are accessible. Also test primary regulatory sources (official journals, regulator websites): many consumer-grade tools ignore them or handle them with several days' delay.

A tool that covers your key sources at 80% with a few hours' latency is often better than one claiming to cover the entire web but processing your priority sources with a two-day lag.

Criterion 2: AI filtering quality

This is the hardest criterion to evaluate without real-world testing, and the most decisive for the daily experience. Poor filtering means the consultant still spends 45 minutes a day sorting irrelevant articles before reaching useful information. The promised productivity gain is cancelled out.

To assess filtering, the most reliable method is a known-corpus relevance test. Take 50 articles you actually used in notes over the past three months. Configure the tool's intelligence profile with the same parameters. Replay collection over the same period and measure recall rate (how many of the 50 relevant articles does the tool find?) and precision rate (of 100 articles retained by the tool, how many would genuinely have been useful?).

A good tool targets a recall rate above 85% and a precision rate above 60%. Below that, filtering creates more work than it saves.

Criterion 3: deliverable formats

For a consulting firm, the intelligence tool is not an end in itself: it is a deliverable production infrastructure. The question is not "can I access information within the tool?" but "does the tool allow me to quickly produce something I can send to the client?"

Formats to check:

• Structured PDF export with white-label capability (logo, colours, firm or client branding)
• Structured summary adapted to different reading levels (short flash / developed note)
• Automated weekly digest by email or web link
• Integration with presentation tools (PowerPoint, Google Slides) for meeting deliverables

A tool that produces beautiful internal visualisations but only exports as CSV or raw tables requires an additional editorial layer. For an active firm, that layer represents 2 to 4 hours per week, a significant fraction of the promised productivity gain.

Criterion 4: value for money

The price spectrum in this market is remarkably wide: from €29 per month for tools targeting independent consulting firms to €30,000 per year for enterprise platforms. The correlation between price and quality is weak, several leading tools are out of reach for 10 to 30-person firms not because their features are useless, but because their revenue model is designed for teams of 10 dedicated analysts, not for a firm where intelligence is one activity among many.

The right approach: calculate the cost per active intelligence profile and per deliverable produced. If the tool costs €800 per month and you produce 12 deliverables per month for 4 clients, the unit cost is €67 per deliverable. If that deliverable is billed (or produces equivalent value), the ROI is immediately calculable.

The detailed comparison between the main solutions on the market is available in our comparison table.

Criterion 5: integrations

For small firms, integrations are often a secondary criterion, the tool is used as a standalone component. For mid-sized and large firms, they become critical because they determine whether the tool integrates into the existing workflow or imposes a parallel flow that teams abandon after three months.

Integrations that genuinely matter: Slack and Teams (to distribute alerts without forcing a change of communication tool), email (for automated digests and critical alerts), CRM (to link intelligence profiles to client accounts), and API for firms with technical customisation capacity.

Integrations that are often sold but rarely used: most "social media analytics" connectors (useless for strategic intelligence), integrations with BI tools (too complex for real-world consulting use), and proprietary connectors to databases you do not subscribe to.

Criterion 6: support and onboarding

An intelligence tool is never truly operational on day one. Intelligence profiles need configuring, relevance criteria need adjusting, alert thresholds need calibrating and teams need training. The quality of support during these first weeks often determines whether the tool is genuinely adopted or used below its potential.

For small firms, support is all the more critical because there is no internal IT manager capable of configuring the tool alone. Evaluate: is there a guided onboarding? Is documentation available in your language? Does support respond within 24 hours to technical questions?

For larger firms, the question is different: is there a dedicated Customer Success Manager capable of understanding the business stakes of a consulting firm, not just the technical aspects of the tool?

Criterion 7: scalability

This criterion is often overlooked at purchase and regretted 18 months later. The question is not "does the tool meet my needs today?" but "will the vendor still be here in three years, can the tool evolve as the firm grows, and can the accumulated data be exported if I switch tools?"

Three practical checks:

• Vendor solidity: is the company established and profitable, or is it a startup at risk of acquisition or closure?
• Data portability: can you export the history of your profiles and briefings if you leave? In what format?
• Roadmap: is there a clear vision of the tool's evolution, and is it publicly communicated?

A tool with limited scalability in a market moving as fast as AI intelligence is a risk to anticipate.

How to weight these criteria for your decision

The grid above gives an indicative weighting by firm size. In practice, two parameters modify it significantly.

If you want to sell intelligence as a service (client subscriptions, add-on offerings), criteria 3 (deliverables) and 4 (pricing) move to the front: your margin on the offering depends directly on the quality of exports and the cost of the tool.

If your intelligence covers non-anglophone markets, criterion 1 (sources) becomes determinant: verify coverage on your specific geographies before any other criterion.

Sentinel Briefing was designed starting from the analysis of these seven criteria applied to independent consulting firms: AI filtering calibrated to intelligence questions, white-label PDF export, entry price at €29/month, support, and architecture designed to evolve with firm size. Our detailed comparison positions the main available solutions across these seven dimensions.

The right tool is not the most complete one, it is the one your consultants actually use.