AI Tool Selection Analysis
Mid-Size Management Consulting Firm — Research, Knowledge Management & Proposal Generation
Date: 2026-05-30 · Prepared by: Resolvix · Status: Sample Deliverable
Deliverable type: AI & Automation — Tool Selection Analysis (~$700)
Industry: Management Consulting
About this sample. This is one example of what a successful Resolvix deliverable looks like at this scope and type — not a template that every engagement follows. Your expert brings their own expertise and judgment to the work: the structure, the emphasis, which angles they dig into, and how they organize their findings will all vary based on your industry, your specific question, and where the research leads. What stays consistent across every engagement is the standard: analysis grounded in evidence, prioritized recommendations, concrete action steps, and a phased implementation plan. All company names, figures, and scenarios in this sample are illustrative.
Executive Summary
Meridian Strategy Group (illustrative; 180 consultants, offices in Chicago, Atlanta, and Denver, ~$68M revenue) is evaluating AI tools to compress research cycles, capture institutional knowledge, and accelerate proposal development. Five tools were evaluated: Perplexity Pro (AI research), Harvey (legal and professional services AI), Glean (enterprise knowledge management), Notion AI (connected workspace AI), and Microsoft Copilot (M365-integrated assistant).
Recommendation: Deploy Glean as the knowledge management foundation (immediate), add Perplexity Pro for research (immediate), and defer Harvey and Copilot pending further evaluation. Do not deploy Notion AI until a Notion migration decision is made.
No single tool addresses all three use cases. The highest-ROI combination for Meridian's current stack is Glean + Perplexity Pro, delivering an estimated $1.9M–$2.6M in annual value through reduced research time and recovered institutional knowledge — at a combined cost of approximately $280K/year.
Firm Context
| Attribute | Detail |
|---|---|
| Consultants | 180 (42 partners/principals, 138 associates/analysts) |
| Offices | Chicago (HQ), Atlanta, Denver |
| Annual revenue | ~$68M |
| Current tech stack | Microsoft 365 (Teams, SharePoint, Word, Excel, PowerPoint), Salesforce (CRM), iManage (document management), custom SharePoint intranet |
| Primary practice areas | Operations, supply chain, digital transformation, private equity due diligence |
| Current AI use | Ad hoc ChatGPT by individuals; no enterprise license or governance |
| Key pain points | Research takes 2–4 days per engagement kickoff; proposal win rate depressed by slow turnaround; institutional knowledge leaves with departing consultants; partners report spending 30–40% of time on content production rather than client work |
Evaluation Methodology
Each tool was scored against nine criteria weighted by Meridian's stated priorities. Weights were assigned based on a priorities survey completed by 12 partners and 8 senior managers during the scoping phase. Raw scores (1–5) were assigned by the expert based on: vendor documentation, independent analyst reviews (Forrester, G2, independent testing), practitioner interviews with users at comparable consulting firms, and hands-on trials where trial access was available.
Weighted Evaluation Matrix
Criteria Weights
| # | Criterion | Weight | Rationale |
|---|---|---|---|
| 1 | Research accuracy & citation quality | 20% | Core use case; hallucinated citations in client deliverables are a reputation risk |
| 2 | Knowledge retrieval from internal documents | 18% | SharePoint and iManage content is where firm IP lives |
| 3 | M365 integration depth | 15% | Firm is deeply committed to M365; friction kills adoption |
| 4 | Proposal / content generation quality | 14% | Proposal acceleration is the #2 stated priority |
| 5 | Data security & enterprise compliance | 13% | Client data confidentiality is non-negotiable; SOC 2 Type II minimum |
| 6 | Ease of use & adoption likelihood | 9% | Prior ad hoc AI use suggests self-serve capability; but governance matters |
| 7 | Pricing & cost predictability | 6% | Budget is firm but not the primary driver |
| 8 | Customization & fine-tuning potential | 3% | Longer-term consideration; not blocking for Year 1 |
| 9 | Vendor stability & roadmap | 2% | All five vendors are well-capitalized; low differentiation on this criterion |
Raw Scores by Tool (1–5 scale)
| Criterion | Weight | Perplexity Pro | Harvey | Glean | Notion AI | Microsoft Copilot |
|---|---|---|---|---|---|---|
| Research accuracy & citation | 20% | 5 | 3 | 2 | 2 | 3 |
| Internal knowledge retrieval | 18% | 1 | 1 | 5 | 3 | 3 |
| M365 integration depth | 15% | 2 | 2 | 4 | 1 | 5 |
| Proposal / content generation | 14% | 3 | 4 | 3 | 4 | 4 |
| Data security & compliance | 13% | 3 | 5 | 5 | 3 | 5 |
| Ease of use & adoption | 9% | 5 | 3 | 4 | 4 | 3 |
| Pricing & cost predictability | 6% | 4 | 2 | 3 | 4 | 3 |
| Customization potential | 3% | 2 | 4 | 4 | 3 | 3 |
| Vendor stability & roadmap | 2% | 3 | 3 | 4 | 3 | 5 |
Weighted Scores
| Criterion | Weight | Perplexity Pro | Harvey | Glean | Notion AI | Copilot |
|---|---|---|---|---|---|---|
| Research accuracy & citation | 20% | 1.00 | 0.60 | 0.40 | 0.40 | 0.60 |
| Internal knowledge retrieval | 18% | 0.18 | 0.18 | 0.90 | 0.54 | 0.54 |
| M365 integration depth | 15% | 0.30 | 0.30 | 0.60 | 0.15 | 0.75 |
| Proposal / content generation | 14% | 0.42 | 0.56 | 0.42 | 0.56 | 0.56 |
| Data security & compliance | 13% | 0.39 | 0.65 | 0.65 | 0.39 | 0.65 |
| Ease of use & adoption | 9% | 0.45 | 0.27 | 0.36 | 0.36 | 0.27 |
| Pricing & cost predictability | 6% | 0.24 | 0.12 | 0.18 | 0.24 | 0.18 |
| Customization potential | 3% | 0.06 | 0.12 | 0.12 | 0.09 | 0.09 |
| Vendor stability & roadmap | 2% | 0.06 | 0.06 | 0.08 | 0.06 | 0.10 |
| Weighted Total | 100% | 3.10 | 2.86 | 3.71 | 2.79 | 3.74 |
Score Summary
| Tool | Weighted Score | Rank | Deployment Recommendation |
|---|---|---|---|
| Glean | 3.71 | 1 (tied) | Deploy immediately |
| Microsoft Copilot | 3.74 | 1 (tied) | Defer — see analysis |
| Perplexity Pro | 3.10 | 3 | Deploy immediately |
| Harvey | 2.86 | 4 | Defer — evaluate in 6 months |
| Notion AI | 2.79 | 5 | Do not deploy at this time |
Why Copilot scores highest but is deferred: Copilot's score reflects its deep M365 integration and enterprise-grade security — genuine strengths. However, Meridian's SharePoint environment is disorganized and inconsistently tagged, which is the primary source of institutional knowledge retrieval failures. Deploying Copilot into a messy SharePoint before remediating that content will produce low-quality retrieval output and kill adoption. The recommended sequence is: use Glean to surface and organize the knowledge that exists, clean SharePoint structure as a byproduct, then evaluate Copilot in 12 months against a remediated content environment. Glean's connectors and search intelligence also make it a better diagnostic tool for understanding where the knowledge gaps actually are.
Tool-by-Tool Analysis
Perplexity Pro — Deploy (Score: 3.10)
What it does well for Meridian: Perplexity's real-time web search with inline citations is best-in-class for external research tasks — industry sizing, competitor landscapes, regulatory updates, and public company analysis. Unlike ChatGPT (which has a knowledge cutoff and frequently fabricates citations), Perplexity returns verifiable source links with every response, which is essential for consulting deliverables. The "Deep Research" mode (multi-step reasoning across 20–30 sources) can compress a 4-hour analyst research task to 25–35 minutes for the right query types.
Where it falls short: Perplexity has no access to internal documents — it is entirely externally focused. It also struggles with highly structured quantitative analysis (financial modeling, regression interpretation) compared to a human analyst. The API for internal integration exists but requires development effort not currently planned.
Specific fit for Meridian: The PE due diligence practice is the highest-value immediate fit. Pre-engagement market scans (TAM sizing, competitive landscape, regulatory environment) are currently a 2–3 day analyst effort. Perplexity Deep Research can produce a first-pass draft in 2–3 hours. Analyst then validates, supplements, and formats — net time savings of 60–70% on the research phase.
Security note: Perplexity Pro's enterprise tier (Perplexity Enterprise Pro) includes SOC 2 Type II compliance and a data processing agreement. Client-confidential data should not be entered into the standard Pro interface — only the Enterprise tier with DPA in place should be used for client work. This governance rule must be communicated at deployment.
Pricing: Enterprise Pro at approximately $40/user/month. For 180 consultants, full deployment = $86K/year. Recommendation: initial rollout to 60 heaviest research users (all associates and analysts), expand based on adoption data. Year 1 cost: $29K.
Glean — Deploy (Score: 3.71)
What it does well for Meridian: Glean is purpose-built for enterprise knowledge retrieval. It connects to SharePoint, iManage, Salesforce, Teams, and email simultaneously, indexes all content, and makes it searchable via natural language. The killer feature for a consulting firm is "who knows what" — Glean surfaces subject matter experts within the firm based on the documents they've authored and the projects they've worked on, not just explicit org-chart relationships. When a Denver associate is staffed on a manufacturing turnaround engagement, Glean can identify that a Chicago partner has authored three past deliverables on exactly this topic and two Atlanta managers have recent client-facing experience — connections that currently only happen by accident.
Knowledge retrieval quality: Glean uses a combination of semantic search (dense retrieval via embeddings) and keyword search, reranked by recency and the user's personal work context. This means it does not simply return documents containing the search terms — it understands intent and surfaces the most relevant content. In independent tests at professional services firms, Glean typically reduces "time to find the right document" by 55–70% vs. SharePoint native search.
Where it falls short: Glean does not generate new content — it retrieves and summarizes existing content. It is not a writing tool. Glean's summarization of retrieved documents is useful for quick orientation but not sufficient for client-ready output without significant human editing.
Security: Glean is SOC 2 Type II certified. It respects SharePoint and iManage permissions — a consultant only sees documents they already have access to. It does not train its general model on customer data.
Critical prerequisite: Glean's retrieval quality is a direct function of SharePoint content quality. If documents are poorly named, inconsistently structured, or stored in disorganized folder hierarchies, Glean will surface that disorganization efficiently. Before Glean deployment, a 30-day SharePoint remediation sprint is required (see Recommendations).
Pricing: Glean typically prices at $18–$25/user/month for enterprise contracts. At 180 users: $39K–$54K/year. Negotiate a pilot for the Chicago office (approximately 80 users) at $17K–$26K for 6 months before full deployment.
Microsoft Copilot — Defer 12 Months (Score: 3.74)
What it does well: Copilot's deep integration with Word, PowerPoint, Excel, Teams, and Outlook is genuinely differentiated. The ability to draft a proposal in Word while pulling context from a Teams meeting transcript and a referenced SharePoint document is a workflow that no other tool in this evaluation can match. Copilot for Sales (Salesforce integration) is also relevant to Meridian's BD team.
Why defer: Three reasons. First, Meridian's M365 tenant has not been configured for Copilot prerequisites — Microsoft requires SharePoint Copilot-readiness configuration (sensitivity labels, oversharing remediation, content classification) that is estimated to take 8–12 weeks of IT effort. Second, as noted above, Glean deployed first will clean up the knowledge retrieval environment that Copilot will depend on. Third, Copilot for M365 is priced at $30/user/month — at 180 users, that is $64.8K/year on top of existing M365 licensing. Proving ROI from Glean and Perplexity first gives the firm evidence to justify this additional spend to the partnership.
Recommended evaluation trigger: Reevaluate Copilot at Month 12 after Glean deployment, once SharePoint content quality has improved and Copilot-readiness configuration is complete.
Harvey — Defer 6 Months (Score: 2.86)
What it does well: Harvey is purpose-built for professional services document work — contract analysis, due diligence document review, and structured legal/regulatory analysis. Its document reasoning capability (understanding a 200-page contract and answering specific questions about it) is meaningfully better than general-purpose LLMs for document-heavy tasks. For Meridian's PE due diligence practice, Harvey's ability to process data room documents at scale is potentially high-value.
Why defer: Harvey's strongest use case (data room document review) is not Meridian's highest-volume pain point today — it is a specific sub-workflow in one practice area. Harvey is also the most expensive tool in this evaluation at $75–$125/user/month for the professional services tier, and it requires dedicated implementation effort for custom workflow configuration. The ROI case is strongest for 15–20 senior associates and managers in the PE practice; it is not a firm-wide tool. Recommend a focused 6-person pilot in the PE practice at Month 6, with a go/no-go expansion decision at Month 9.
Notion AI — Do Not Deploy (Score: 2.79)
The fundamental problem: Notion AI only operates within Notion. Meridian does not use Notion — the firm's knowledge management, document storage, and collaboration all happen in M365/SharePoint/iManage. Deploying Notion AI would require migrating a meaningful volume of content to Notion, which creates a parallel system that competes with SharePoint and iManage, fragments knowledge, and introduces redundant licensing cost. The only scenario in which Notion AI makes sense for Meridian is if the firm decides to migrate its knowledge management infrastructure to Notion — a significant project that is out of scope for this evaluation and would need to be evaluated on its own merits. Do not deploy Notion AI without that migration decision.
Licensing & Total Cost Summary
| Tool | Users | Unit Price | Year 1 Cost | Recommendation |
|---|---|---|---|---|
| Glean | 80 (pilot, Months 1–6) → 180 (Month 7+) | $22/user/month | ~$37K pilot + $24K expansion = ~$61K | Deploy |
| Perplexity Enterprise Pro | 60 (associates/analysts) | $40/user/month | ~$29K | Deploy |
| Microsoft Copilot | — | — | — | Defer |
| Harvey | 6 (PE practice pilot) | $100/user/month (est.) | ~$7K (pilot only) | Defer to Month 6 |
| Notion AI | — | — | — | Do not deploy |
| Total Year 1 | ~$97K |
ROI Estimate
Research time savings (Perplexity, 60 users):
60 users × 4 hours/week research × 50% efficiency gain × 46 work-weeks = 5,520 hours recaptured annually. At a blended loaded cost of $95/hour for associates and analysts: $524K in recovered capacity.
Knowledge retrieval savings (Glean, 180 users):
Industry benchmark: knowledge workers at professional services firms spend 15–20% of their time searching for information. At Meridian, assuming 18% (conservative), 180 consultants × 2,000 hours/year × 18% × 55% retrieval time reduction × $110 blended loaded cost = $1.24M in recovered capacity.
Proposal acceleration:
Meridian produces approximately 85 proposals per year. Partners estimate current proposal development takes 18–22 hours of billable-equivalent effort per proposal. Estimated 30% reduction via AI-assisted first drafts: 85 proposals × 6 hours saved × $195 partner-equivalent rate = $99K in recovered senior capacity.
Total estimated annual value: $1.86M
Year 1 tool cost: $97K
Net Year 1 value: $1.76M
ROI: 18:1
Recommendations
-
Deploy Glean in a Chicago-only pilot (80 users) for the first 6 months before expanding firm-wide. Chicago has the largest staff concentration and the most complete SharePoint content — it is the highest-signal pilot environment. Define success metrics before launch (search query volume, document retrieval satisfaction score, internal referral rate to subject matter experts). Set a go/no-go expansion gate at Month 6 before committing to the full 180-seat license.
-
Run a 30-day SharePoint remediation sprint before Glean goes live. Assign one IT staff member and two consultant volunteers to audit the top 10 SharePoint document libraries: standardize folder naming conventions, delete duplicate documents, apply consistent metadata tags (practice area, client sector, engagement type, year). This is not glamorous work, but it is the single biggest determinant of Glean retrieval quality. Glean's implementation team can provide a content audit report as part of onboarding — use it.
-
Deploy Perplexity Enterprise Pro immediately to all associates and analysts (60 users), with a governance policy issued simultaneously. The governance policy must specify: (a) Perplexity may be used for external research only on the Enterprise Pro tier; (b) no client-identifying information, confidential data, or privileged information is to be entered into any AI tool without explicit IT approval; (c) all AI-generated research content must be source-verified by the analyst before inclusion in a client deliverable. This policy is non-negotiable before any AI research tool goes live.
-
Do not allow informal ChatGPT use to continue without governance. The current state — individuals using personal or ad hoc ChatGPT accounts for client work — is the highest data security risk in Meridian's current AI posture. Client confidential information is almost certainly being entered into tools with no enterprise DPA, no access controls, and no data retention policy. Issue a firm-wide AI acceptable use policy within 30 days. The policy should not prohibit AI use (which would be unenforceable) — it should channel use to the approved enterprise tools with appropriate controls.
-
Evaluate Harvey as a 6-person PE practice pilot at Month 6. Select the 6 most frequent data-room users in the PE practice. Define a specific test workflow: processing a data room for a lower-middle-market manufacturing target. Measure: hours spent on document review pre- vs. post-Harvey, analyst confidence in findings, partner review time required. Use that pilot data to make a rigorous go/no-go decision before expanding to the full PE team.
-
Schedule a Copilot readiness assessment for Month 9. Engage Microsoft's FastTrack team (included with enterprise M365 licensing) to assess Meridian's current SharePoint configuration against Copilot readiness requirements. This assessment is free and provides a remediation roadmap. Begin SharePoint sensitivity labeling and oversharing remediation in parallel with Glean deployment so that Copilot prerequisites are met by Month 12.
-
Appoint a Partner-level AI Champion to own adoption and governance. AI tool adoption at professional services firms fails most often due to absence of a visible, credible internal advocate. The AI Champion role does not require technical expertise — it requires partnership authority and willingness to model behavior. Responsibilities: co-design the governance policy, champion the Glean pilot, communicate firm-level metrics to the partnership quarterly, make the go/no-go call on tool expansion.
Action Steps
| # | Action | Owner | Time | Tied To |
|---|---|---|---|---|
| 1 | Issue firm-wide AI acceptable use policy (draft for managing partner approval) | IT Director + General Counsel | 0–14 days | Rec 4 |
| 2 | Appoint AI Champion from partner group | Managing Partner | 0–14 days | Rec 7 |
| 3 | Execute Glean enterprise agreement (Chicago pilot, 80 seats, 6 months) | IT Director / CFO | 0–21 days | Rec 1 |
| 4 | Execute Perplexity Enterprise Pro agreement (60 seats) | IT Director / CFO | 0–21 days | Rec 3 |
| 5 | Begin SharePoint remediation sprint (top 10 libraries) | IT + 2 consultant volunteers | Days 14–45 | Rec 2 |
| 6 | Glean onboarding and connector configuration (SharePoint, iManage, Salesforce, Teams) | IT + Glean implementation team | Days 21–45 | Rec 1 |
| 7 | Perplexity orientation session for all 60 analysts/associates (1-hour live training) | AI Champion + IT | Days 21–35 | Rec 3 |
| 8 | Glean Chicago pilot launch; baseline metrics established | IT / AI Champion | Day 45 | Rec 1 |
| 9 | Harvey pilot scoping: select 6 PE practice users, define test workflow | PE Practice Lead | Month 5 | Rec 5 |
| 10 | Glean Month 6 go/no-go review: adoption data, satisfaction survey, retrieval quality audit | AI Champion + IT | Month 6 | Rec 1 |
| 11 | Harvey pilot launch (if scoping approved) | PE Practice Lead + IT | Month 6 | Rec 5 |
| 12 | Glean firm-wide expansion (if go/no-go approved): 100 additional seats | IT Director | Month 7 | Rec 1 |
| 13 | Microsoft FastTrack Copilot readiness assessment | IT Director | Month 9 | Rec 6 |
| 14 | Harvey go/no-go decision for PE practice expansion | PE Practice Lead + CFO | Month 9 | Rec 5 |
| 15 | Copilot deployment decision based on readiness assessment | AI Champion + IT + CFO | Month 12 | Rec 6 |
Implementation Plan
Phase 1: Foundation Tools & Governance (Months 1–6)
Objective: Establish enterprise AI governance; deploy Glean and Perplexity Pro; measure adoption and ROI.
Milestones:
- Month 1: Acceptable use policy issued; Glean and Perplexity contracts executed; AI Champion appointed
- Month 1.5: SharePoint remediation complete; Glean connectors live
- Month 1.5: Perplexity orientation training complete for all 60 associates/analysts
- Month 2: Glean Chicago pilot live; baseline metrics captured (searches per day, user satisfaction score, time-to-document benchmark)
- Month 4: Mid-pilot Glean satisfaction survey; identify adoption laggards; address friction points
- Month 6: Glean pilot go/no-go review; Harvey PE pilot scoped
Go/No-Go Gate (Month 6 — Glean expansion):
- ≥ 65% of Chicago pilot users active (≥1 Glean search/week) over the last 30 days
- Average "search relevance" satisfaction score ≥ 3.5/5
- At least 3 partner-level endorsements of knowledge discovery value (qualitative)
Success Criteria:
- Research time per engagement kickoff reduced by ≥ 40% for Perplexity users (measured by analyst self-report)
- Internal knowledge referrals (consultants finding internal expertise via Glean) up ≥ 50% vs. pre-deployment baseline
- Zero security incidents involving client data in AI tools
Budget: $97K (tools) + $25K (IT implementation effort, SharePoint remediation, training) = $122K
Phase 2: Expansion & Optimization (Months 7–12)
Objective: Expand Glean firm-wide; run Harvey pilot; begin Copilot readiness preparation; integrate AI tools into proposal workflow.
Milestones:
- Month 7: Glean expanded to all 180 consultants; Denver and Atlanta onboarding sessions
- Month 7: Proposal generation workflow redesigned with AI-assist step (Perplexity for competitive intelligence, Glean for relevant past work retrieval, AI draft for structure)
- Month 8: Harvey PE pilot launched
- Month 9: FastTrack Copilot readiness assessment complete; SharePoint sensitivity labeling begins
- Month 9: Harvey pilot go/no-go data collected
- Month 10: Proposal workflow AI-assist measured: win rate trend, proposal cycle time
- Month 12: Copilot deployment decision; Harvey expansion decision; Year 1 ROI review against estimates
Success Criteria:
- Glean firm-wide active user rate ≥ 60%
- Proposal development cycle time reduced ≥ 25%
- Harvey PE pilot: data room review hours reduced ≥ 35% vs. pre-Harvey baseline
- AI tool total cost vs. recovered capacity ROI documented for partnership review
Budget: $180K–$220K (Glean expansion $117K annualized, Harvey pilot $7K, Copilot prep IT effort $40K–$60K, training $16K)