Originally published on rikuq.com. Republished here for Dev.to's readers.

Most Indian mid-market entities I've worked with don't know how many SaaS subscriptions they have. The number lives across procurement, IT, individual BU credit cards, and "Sunil signed up for that one in 2023 for that project we never finished." When you actually count, the number is usually 60% higher than the CFO's initial estimate. And when you cross-reference what's being used, somewhere between 20-40% of those subscriptions are either redundant with another vendor in the stack or actively underutilised.

AI vendor consolidation is the procurement strategy of deliberately reducing this sprawl. It's been one of the dominant CFO procurement themes of 2026 — 68% of CIOs identify it as a top-3 priority, average enterprise reduced SaaS vendor count by 23% over the preceding 18 months. But consolidation done badly destroys more value than the savings it captures. This post walks through when it works, when it doesn't, and what changes when you're operating in the Indian tax context.

I'm Ravi. I run three production AI SaaS solo (Prism, Citare, BatchWise) and do advisory work on AI vendor portfolios via rikuq services. The framework below is what I take into vendor consolidation engagements.

TL;DR

What AI vendor consolidation actually means

AI vendor consolidation is the procurement strategy of deliberately reducing the count of AI / SaaS / cloud vendors an enterprise relies on, concentrating spend across a smaller number of strategically-chosen platforms.

It's the natural counterforce to SaaS sprawl — the cumulative count of subscriptions that builds up over time as individual teams add tools faster than anyone evaluates whether existing tools could meet the same need. SaaS sprawl is not an attribute of bad procurement; it's the default behaviour of distributed organisations adopting cloud and AI capabilities. Active consolidation is the only thing that pushes back against it.

The 2025-2026 wave of consolidation is specifically about AI-era vendors — foundation model providers, ML platforms, AI-feature SaaS, and the traditional SaaS that has added AI features competing with native-AI tools.

Why consolidation accelerated in 2025-2026

Three structural forces converged:

1. AI compound advantage at platform level. Platforms that ship AI capability in Q1 ship more in Q2, Q3, Q4. The capability gap widens against point solutions that struggle to keep AI pace. Buyers consolidating onto AI-native platforms now lock in compounding advantage over the next 2-3 years that fragmented vendor portfolios can't access. This is different from typical procurement consolidation — it's not just about negotiating leverage, it's about which side of the compound advantage curve you end up on.

2. Procurement leverage at scale. Concentrating spend with a smaller number of vendors enables better negotiation. Committed-use discounts, Most-Favoured-Nation pricing clauses, take-or-pay structures, multi-year pricing locks — all become accessible at scale that fragmented spend cannot achieve. A mid-market entity with ₹2 crore split across 8 AI vendors negotiates worse than the same entity with ₹2 crore concentrated on 3 vendors.

3. FinOps optimisation efficiency. AI-spend optimisation techniques like model tiering, prompt caching, batching, and request routing work best when applied across a small number of consolidated platforms. Industry analysis shows these techniques can cut GenAI costs by 40-70% — but only when applied to a manageable vendor surface. Try to do them across 12 providers and the engineering effort overwhelms the savings.

The consolidation playbook — patterns that actually work

A typical mid-market consolidation engagement identifies recurring patterns:

The pattern across all of these: real redundancy exists when the consumed capability is substantially the same. Apparent redundancy where the vendor branding is similar but the actual use is distinct is not real redundancy and consolidating there destroys value.

The trade-offs consolidation creates

Consolidation is not unambiguously positive. Three counterforces:

Concentration risk rises with consolidation. Single-vendor dependence creates outage exposure (if your foundation model provider has a bad day, half your product is offline), pricing-power asymmetry (your vendor knows your switching cost is high, so they negotiate harder on renewal), and switching difficulty (if the vendor changes terms unfavourably, your remediation path is months not days).

Capability ceiling — best-of-breed point solutions may genuinely outperform consolidated platform incumbents on specific use cases. A specialist transcription tool may be materially better than Anthropic's built-in capabilities for a workload that depends on transcription accuracy. Consolidating onto the incumbent at the cost of capability is a real risk that gets papered over in consolidation pitches.

Migration cost to consolidate is non-trivial. Data migration, integration rebuilds, team retraining, change management all consume engineering capacity that has opportunity cost. A consolidation that looks like ₹50 lakh in annual savings often costs ₹30 lakh in migration work and three months of engineering time that could have shipped something else.

The methodology that supports vendor consolidation analysis quantifies this trade-off explicitly: gross savings from consolidation minus capability degradation cost minus migration cost = net consolidation value. Not all theoretical consolidations are economically rational.

When consolidation is the wrong move

Specific situations where holding is correct:

• The capability gap is large. If the specialist tool produces materially better outcomes on the workload that drives the business value, the consolidation savings rarely exceed the capability loss.
• Concentration risk is unacceptable. For mission-critical workloads where outage exposure or pricing-power asymmetry is genuinely catastrophic, the single-vendor risk premium isn't worth the consolidation savings.
• Migration cost exceeds multi-year savings. Some integrations are expensive enough to rebuild that the consolidation math works only over 5+ year horizons that nobody has visibility into.
• The apparent overlap isn't real overlap. When the workloads look similar in vendor categorisation but the actual use cases are distinct, consolidation is solving a problem that doesn't exist.
• Vendor strategic alignment matters. When the vendor is genuinely differentiated by its strategic roadmap that aligns with your direction, switching is worse than the consolidation math suggests.

The rule of thumb: consolidate aggressively where capability is commoditised, hold tightly where capability is genuinely differentiated.

India context — what changes for Indian entities

Indian mid-market enterprises typically have lower SaaS vendor counts than US peers (smaller workforce, less BU diversity, often single-jurisdiction operations). The Gartner 305-application enterprise average is a US-skewed number; Indian mid-market typically sits at half to two-thirds of that. But the rate of accretion at 10-15% year-over-year for growing companies is in line with global benchmarks.

What's distinctive in the Indian context — three tax overlays that materially change the consolidation math:

GST input credit recovery on consolidation savings. Lower aggregate spend on registered Indian SaaS vendors means lower GST cash-flow drain, subject to ITC time-limit constraints under Section 16(4) CGST. Consolidation that reduces domestic SaaS spend frees up working capital that was tied up in the ITC cycle.

Section 195 TDS simplification on foreign vendor consolidation. Consolidating from 10 foreign vendors to 5 means fewer Form 15CA/15CB filings, simpler TDS reconciliation, and reduced compliance overhead. The operational savings here often exceed the procurement savings themselves — depending on the entity's existing compliance burden.

Transfer pricing on cross-entity AI cost allocation. Consolidating shared AI/cloud spend at the group level rather than entity-by-entity simplifies TP documentation under Section 92 plus OECD TP Guidelines Chapter VII. Cleaner allocation methodology, fewer related-party transactions to document, easier benefit test substantiation. See the AI tax recovery framework for sub-domain 5 on TP.

The combined effect: the Indian tax overlay typically improves the consolidation math by 10-20% beyond what the gross procurement savings would suggest. This is rarely included in vendor consolidation business cases produced by global consultancies operating without the India context.

The six-step framework that survives contact with reality

Step 1: Build a complete vendor inventory across all BUs. Most entities don't actually know their full vendor list. Pull from procurement, accounting, individual BU credit cards, IT systems. The first run usually surfaces 30-50% more vendors than initial estimates.

Step 2: Classify by category. Foundation model providers, ML platforms, AI-feature SaaS, traditional SaaS with AI features, infrastructure tooling. Allows comparison within categories where overlap is actually possible.

Step 3: Identify overlap zones. Where multiple vendors serve substantially the same workload for the same teams. This is where consolidation candidates live.

Step 4: Quantify the consolidation math per overlap. Gross savings (procurement leverage + redundancy elimination), capability degradation cost (does the surviving vendor handle the work as well?), migration cost (engineering effort, integration rebuilds, change management), concentration risk (what's the dependency exposure?). Each overlap gets a net value.

Step 5: Execute the consolidations that survive the math. Prioritise overlaps where renewal is imminent (no extra exit cost) and where the consolidation math is most favourable. Sequence over 12-18 months; don't try to consolidate everything in one quarter.

Step 6: Build the procurement discipline that prevents re-sprawl. New vendor evaluation criteria (does this overlap with existing tools? what's the sunset plan if not used?), owner sign-off requirements, sunset criteria for unused tools, regular portfolio review cadence.

The discipline matters more than the one-time consolidation pass. Without it, sprawl reasserts itself within 18 months.

What to do this week

If you're at an Indian mid-market entity with material AI/SaaS spend and you haven't formally run vendor consolidation analysis:

• Pull the complete vendor inventory. Procurement, accounting, IT systems, BU credit cards. If your initial list is under 30 SaaS vendors, the inventory is incomplete.
• Classify by AI/non-AI and by category. This is the foundation for identifying overlap zones.
• Identify the 3-5 highest-overlap opportunities. The ones where multiple vendors clearly serve the same workload.
• Run the math on those specifically. Gross savings, capability degradation, migration cost, concentration risk. Many will turn out to be marginal; that's useful information.
• Audit your foreign vendor list for tax overlay opportunities. Section 195 simplification and GST ITC recovery often make consolidations economic that wouldn't be on procurement math alone.

If you want a written scope proposal for what a vendor consolidation engagement would look like for your specific situation — including the Indian tax overlay analysis that most procurement consultants don't run — the services page has both a Cal.com booking link and a Tally intake form.

What's next

This post focuses on vendor consolidation specifically. The adjacent posts that go deeper on related dimensions:

• AI Systems Review framework — Domain 2 of the six-domain framework covers vendor consolidation analysis
• AI Tax Recovery for Indian CFOs — the tax overlay that changes consolidation math
• AI/Cloud FinOps Providers in India — provider selection when external help is needed for consolidation analysis
• The FinOps Foundation Framework — where vendor consolidation sits in the FinOps Optimize phase