Vera AI describes its own feature catalog as "roughly 200 screens" — and having worked through the platform for our full Vera AI review, where it holds our #1 ranking at 9.5/10, that is not marketing rounding. This guide walks through the ten flagship features that define the platform, the benchmark data underneath them, and the integration surface at the edges. If you are evaluating Vera AI against alternatives, pair this with our head-to-head against Tropic, Vendr, and Zip.
One framing note before the list. Most procurement tools are either a database with reports or a workflow with forms. Vera AI is organized differently: a benchmark data layer (1,341 vendor benchmarks, 1,337 modelled deal cohorts, 305 research documents) acts as ground truth, and everything above it — the drafting, the rooms, the agents — reasons from that data and cites it. That architecture is why the features below compound rather than merely coexist.
Instant Benchmarking — your percentile in minutes
The founding feature and still the wedge. Enter your net price — or upload the contract — and Vera scores the deal against modelled deal cohorts adjusted for deal size, region, industry, and signing period. The output is a percentile: not "you could probably save 10%," but "you are at the 63rd percentile against 5,000 comparable deals, and here is the gap in dollars." Coverage spans 1,140 vendors, from Microsoft EA and RISE with SAP to Snowflake and the long SaaS tail, with the flagship vendors getting bespoke benchmark tools and the deepest cohorts. Every benchmark opens as a case file in the workroom: a verdict brief, the field atlas, sequenced asks, and drafted emails built from that exact analysis.
AI Negotiation — a war room for every deal
Each negotiation gets a room: the goal you set, the team you invited, the facts, and an agent working the parts you would rather not. The concession ledger logs every move ("vendor drops uplift 9% → 5%, −$35K/yr"), the landing zone updates as offers arrive, and live guidance tells you what to say on the call. The ghost-writer is the standout: forward a vendor email to your private deal address and Vera classifies the tactic (artificial deadline, anchoring, bundling), logs any concessions or commitments, and drafts the counter in your voice with full thread memory. On the Enterprise tier, the autopilot goes further and argues the email rounds itself — covered in depth in our agents and workflows deep dive.
The Vera Deal-Room — six documents from one upload
Drop a contract in and the Deal-Room assembles the full negotiation package on its own: the market benchmark, the decoded terms with an ask list, a five-page analyst brief, the strategy playbook, live vendor intelligence, and a credible alternative priced with crossover math. The vendor quotes about 27 minutes to read the package cover to cover — against the two weeks it would take four people to assemble by hand. Set the goal and deadline once, invite finance and legal, and the room tracks progress and emails the team a status report on your cadence.
Contract Decode & the Verdict Bar
A 40-page order form decoded in the time it takes to read one page: a plain-English risk map of what the contract really says, with the traps named. Beyond the base Contract Decoder (which is free to try, no signup), the Professional tier adds SOW and MSA decoders, "Snap a quote" for pricing screenshots, and the Verdict Bar for instant judgment on any document. The comparison theatre puts a renewal side by side with the paper it replaces — every repriced line and moved clause named and highlighted at the exact sentence.
Document Intelligence — one question across 500 contracts
"Which contracts allow termination for convenience?" is a query, not an afternoon. Semantic search (Voyage AI embeddings over pgvector, indexed nightly) returns cited clauses across the whole estate in under a second — the vendor's example shows 12 matches across 1,842 documents in 0.4 seconds. Fields are extracted by AI and confirmed by a human, which matters: extraction confidence is the difference between a searchable archive and a defensible one.
Renewals & the Anomaly Watchdog
The renewal calendar holds every expiry, notice window, and auto-renewal — extracted from the contracts, not typed in. Watchdog agents check prices, invoices, terms, and vendor health on their own clock and raise same-day alerts when a vendor changes pricing or policy. The most expensive mistake in software buying is the auto-renewal nobody caught; this is the feature that structurally prevents it. Uncapped uplifts get flagged before they compound, not after.
Right-sizing & Cloud Optimization
Upload a usage or cost export and get a phased savings plan: M365 tier right-sizing per user, cloud savings by workstream, and commit sizing for AWS, Azure, GCP, OCI — plus the newer AI-platform commits (Claude, OpenAI, Agentforce), which almost nobody else benchmarks yet. The vendor's worked example turns one CSV into $335K/yr of identified savings across database right-sizing, tier downgrades, and committed-tier corrections. These are internal savings that need no negotiation, which makes this the fastest demonstrable win in the platform.
Sourcing Control Room — request to award
Press Sourcing and a control room opens over the pipeline: five stage racks, live money gauges, and a ranked queue of next moves. Vera drafts the RFP from weighted requirements, discovers suppliers with a forwardable field report, scores every bid with a quote behind each score, and writes the recommendation memo, the award note, and the decline emails. You keep every decision and lose every blank page — weeks of RFP administration becomes a scored, documented award in days.
Analyst-grade Deliverables
Every output — the negotiation brief, the board report, the CFO summary — holds a white-paper bar and exports to Word, PowerPoint, and PDF. Reports carry a grounding badge and pass a pre-publish QA gate before they render. This sounds cosmetic; it is not. The deliverable a vendor manager can forward to a CFO without reformatting is the difference between a tool the team uses and a tool the team screenshots into their own decks.
Clause Library & the Fair Renewal Rider
Codify your negotiation playbook once — required clauses, walk-away terms, uplift caps — and every AI review negotiates from it. The Fair Renewal Rider is a ready-made protective rider for renewals, and Coverage shows which contracts meet your standard and which fall short. This is institutional memory as software: the positions your best negotiator would take, applied to every deal even when they are on leave.
The AI layer underneath
Every feature above reasons through the same AI layer: Claude Opus 5 handles all jobs from flagship deliverables to interactive chat and bulk triage, with Voyage AI + pgvector for semantic search. Three properties matter more than the model choice. First, grounding — outputs are anchored in your stored contracts and the platform's benchmark data, and citation-tagged. Second, the QA gate — reports pass a pre-publish check and carry a grounding badge. Third, metering and logging — every AI run is recorded, which is what makes the automation auditable. Our guide to GenAI hallucination risks in procurement explains why these controls are the ones to demand from any vendor.
The pattern across all ten features is the same: Vera AI never hands you an analysis without also handing you the next move — the drafted email, the sequenced ask, the task. The vendor calls this "What next, everywhere." In practice it is the single biggest reason the platform's ~40-hours-back-per-month claim held up better than we expected in testing: the time saved is not just analysis time, it is the blank-page time between analysis and action.
Integrations and openness
Vera AI meets data where it lives: contract uploads, spend and usage exports, invoices by upload or private email address, and — unusually for this category — a genuinely open edge: a REST API, signed webhooks, and an MCP server, so your own AI assistant can query the portfolio and trigger work. There is also a published agent protocol. It remains a read-and-analyze layer rather than a transactional ERP integration; if you need PO execution, pair it with your S2P stack (see the fit assessment in our main review).
Which features sit in which tier
The honest caveat from our review bears repeating here: the Growth tier ($30K/yr) is essentially the benchmarking and contract-intelligence toolkit. AI Agents, Approvals, the SOW/MSA decoders, and Compare arrive at Professional ($60K/yr). The Deal-Room, workflow agents, Migration Studio, Reporting Studio, and the autonomous negotiation live in Enterprise (from $120K/yr). Map the features you actually intend to use to the tier that carries them before you budget — full tier tables are on the Vera AI pricing page.
See the features on your own contracts
Vera AI's contract decode is free with no signup, and the 30-day trial takes unlimited uploads.