AI sourcing implementation demo

Works with your current process

AI implementation for sourcing teams

Add secure AI to the sourcing process you already use.

James helps procurement teams reduce manual supplier discovery, benchmarking, rate-card normalization, and bid evaluation—without replacing their current RFP platform.

Works with existing sourcing tools Start with one live category Secure, human-reviewed workflow

Example supplier recommendationHuman review

Rank 01OrbitStack SoftwareEnterprise SaaS platform

89.4/100

0189.4

0287.9

0386.1

Recommendation changes when pricing reaches 47.5% of total weight.

Define requirementsComplete
Sourcing strategyGenerated
Build the RFPGenerated
Evaluate proposalsConsensus
Negotiate & awardRecommended
06
Implement & governIn review

What you hire James to implement

Keep your sourcing system. Add AI where the manual work remains.

James configures secure, category-specific AI around the tools and approval process your team already uses. SourceLens shows the kind of workflow, controls, and deliverables he can build with you.

Implementation service · Not replacement RFP software

01 · Your current process

Keep the tools your team knows

Ariba, Coupa, Ivalua, another e-sourcing platform, or Excel and email.

02 · AI accelerator configured by James

Automate the research and analysis

Find challengers, organize benchmark evidence, normalize rate cards, and evaluate supplier responses.

03 · Back into your workflow

Use familiar sourcing outputs

Bidder lists, evaluation workbooks, negotiation packs, approval summaries, and a decision audit trail.

Supplier discovery

Find credible non-incumbents

Research qualified challengers and document why each supplier should—or should not—be included.

Benchmarking & pricing

Normalize rates faster

Combine approved internal, licensed, supplier-provided, and public evidence into a comparable pricing view.

Secure bid evaluation

Produce an approval-ready decision

Separate pricing from non-pricing scores, preserve evidence, test award scenarios, and keep human approval.

AI does not invent benchmark rates. It organizes and compares authorized evidence with its sources, while confidential supplier data stays in the company-approved environment.

How a pilot works

Bring one upcoming sourcing event.

Start with one live category and measure time saved, additional supplier competition, and improvement in the quality of the analysis.

Explore the synthetic pilot Visit James’s portfolio

01
Select the eventChoose a category with meaningful manual effort.
02
Map the workflowReview current tools, data sources, controls, and outputs.
03
Configure and runBuild the secure AI workflow and run it alongside the current process.
04
Measure and hand overCompare results, train the team, and document governance.

SourceLens is a synthetic demonstration of the deliverables from an AI sourcing implementation—not a production RFP system.

Step 1 of 6

Define Requirements

Enter the sourcing need and constraints

Select an indirect sourcing category, then review its spend, scale, objectives, and constraints. The entire page updates with fictional Technology & Software data.

Step 2 of 6

Develop Sourcing Strategy

Find suppliers and set qualification rules

Use AI to build a bidder list, distinguish the incumbent from credible challengers, and define the mandatory qualifications that determine who advances.

Mandatory qualification gatesEvaluated before weighted scoring

4 pass · 1 review

Security certification

SOC 2 Type II and annual penetration-test evidence

Pass

Identity integration

SAML SSO plus SCIM provisioning

Pass

Data residency

Approved hosting regions and subprocessors

Pass

API capability

Documented APIs for all critical integrations

Pass

Exit support

Contractual data export and deletion commitments

Review

Bidder strategyAI-assisted supplier discovery · sourcing validation required

1 incumbent · 2 challengers

OrbitStack SoftwareEnterprise SaaS platform

Incumbent

Meridian Cloud SystemsCost-led SaaS challenger

Challenger

NovusOne TechnologiesProduct-led technology specialist

Challenger

Step 3 of 6

Build the RFP

Create the weighted scoring model and pricing schedule

AI replaces hours of manual Excel setup by separating non-pricing requirements from the quantitative pricing model and preparing consistent supplier response templates.

QualitativeNon-pricing evaluation75%

Capabilities, service, risk, implementation, coverage, data, and other category requirements scored on a defined 1–5 scale.

QuantitativePricing evaluation25%

Rate cards and cost schedules normalized to the same demand assumptions, then scored from the lowest evaluated price.

Weighted scoring modelQualitative 75% + pricing 25% · total automatically normalizes when needed

100 total points

Qualitative · non-pricing75 points

91%

24%

86%

18%

83%

14%

85%

11%

79%

Quantitative · pricing25 points

88%

25%

Price score formula

Lowest evaluated price ÷ supplier evaluated price × 100

Rate-card normalization

Every bidder is evaluated against the same volumes, roles, usage, and commercial assumptions.

Total evaluated cost

Recurring charges, variable rates, one-time costs, discounts, credits, and risk adjustments are included.

Step 4 of 6

Evaluate Proposals

Score non-pricing responses and normalize supplier pricing

The evaluation keeps qualitative scoring and quantitative pricing distinct, then combines them using the approved weights. Select a supplier to review the underlying evidence.

Secure internal AI evaluation

Production deployment pattern: supplier responses and confidential rate cards stay inside the company-approved AI environment—not copied into a personal ChatGPT subscription. This public demonstration uses synthetic data.

Company controlled

Quantitative pricing comparisonAnnualized five-year total cost using forecast licenses and usage, implementation amortization, support, and committed discounts.

Lowest price ÷ supplier price × 100

Evaluated cost build-up	OrbitStack SoftwareIncumbent	Meridian Cloud SystemsChallenger	NovusOne TechnologiesChallenger
Subscription & support	$31.0M	$29.2M	$29.8M
Usage & add-ons	$4.0M	$3.6M	$4.1M
Implementation (annualized)	$2.6M	$2.1M	$2.4M
Credits & discounts	−$0.8M	−$0.8M	−$0.9M
Total evaluated price	$36.8M	$34.1M	$35.4M
Pricing score / 100	92.7	100.0	96.3

AI maps unlike supplier rate cards into one comparison, checks the arithmetic, flags missing assumptions, and preserves the source-to-score trail for human review.

OrbitStack Software · qualitative evaluationNon-pricing response evidence mapped to evaluator consensus scores

91% evidence confidence

Qualitative score88.3

Solution fit4.6

Security & risk4.3

Implementation4.2

Integration & data4.5

Roadmap4.4

Evidence notes

Scripted demos met 46 of 48 priority use cases without customization.
Security package includes current SOC 2, penetration testing, and a documented AI-data policy.
Commercial proposal needs stronger protections for unused licenses at renewal.

Secure AI summarized · evaluator verification required

Step 5 of 6

Negotiate & Award

Test options and prepare an award recommendation

See when the recommended supplier changes, identify negotiation priorities, and use AI to prepare a concise award summary.

Pricing-weight sensitivityTest pricing from 5% → 50% of the total evaluation

5%15%25%35%50%

Decision switch point

NovusOne Technologies overtakes OrbitStack Software when pricing reaches 47.5% of the total score.

OrbitStackMeridianNovusOne

Negotiation prioritiesValue capture before award

01
Protect against shelfwareAdd annual license rebalancing and volume-down rights without repricing.
02
Cap renewal increasesSet a firm multi-year ceiling and preserve benchmark rights.
03
Secure the exitDefine data formats, deletion evidence, transition support, and capped fees.

Executive summaryDrafted for steering committee

The evaluation supports a conditional award to OrbitStack Software (incumbent). Its combined score of 89.4/100 consists of a 88.3 qualitative score and 92.7 pricing score. It leads on functional fit, security, integration readiness, and product roadmap while delivering modeled savings of 13%. Approval should remain conditional on security remediation closure, license-volume protections, and capped data-extraction fees.

148 evidence references 3 conditions

Step 6 of 6

Implement & Govern

Review, approve, and record the sourcing decision

The sourcing team remains responsible for the decision. Human controls confirm the non-pricing evaluation, pricing analysis, data handling, incumbent bias review, and award rationale.

Approval checklist0 of 6 controls complete

Pending

Evaluation weights reviewed with stakeholders

Qualitative evidence and evaluator rationales verified

Rate-card normalization, totals, and pricing formula verified

Supplier data handled in the approved internal AI environment

Evaluator conflicts and incumbent bias disclosed

Transition risks accepted by the business owner

Complete all human controls to enable approval.

Decision audit trail4 recorded actions

Logged

09:26

Scoring importedSourcing Manager · 3 evaluator consensus scorecards

09:18

Evidence mappedSourceLens AI · 148 response citations · 3 data-quality flags

09:14

Model generatedSourceLens AI · 5 gates · 6 criteria · pricing schedule

09:12

Event createdSourcing Manager · Technology & Software