Comparison · Active 2026
Lift vs ConnectionAI
How Lift compares to ConnectionAI on detection breadth, integrations, customer outcomes, and what an estimator should buy in 2026.
01 · In short
ConnectionAI's flagship is the Steel Brain, an AI that predicts steel connections at LOD 400 in minutes. It is a detailing aid more than a takeoff product. That is the seam where it stops competing head-to-head with Lift. Plenty of teams run both.
This page lays out how the two line up on detection, maturity, integrations, usability, pricing clarity, and stack independence. Then it covers where each has the real edge and where each falls short. The one-paragraph call is at the bottom. Both are real options for a steel shop. Which one fits depends on the specifics below, not on a single headline number.
02 · Head-to-head score
Editorial scoring, 0 to 100, based on public claims, product surface area, field reports, and 2026 positioning. It is not a benchmark. Neither product leads on every axis, and where ConnectionAI is ahead, the bar shows it.
03 · Spec comparison
| Attribute | Lift | ConnectionAI |
|---|---|---|
| Vendor | SketchDeck.ai | ConnectionAI (connectionai.io) |
| Status | Generally available · paying customers since 2021 | Generally available · Active 2026 |
| Pricing | By quote · no per-month takeoff cap | Not disclosed |
| Detects | Beams, Columns, Braces, Joists, Plates | Connections (LOD 400) |
| Member attributes | Shape, Size, Length, Stud counts, Camber, Framing conditions, Moments, Copes, Holes | Not publicly itemized at this depth |
| Elevations | Yes, plans and elevations | Not applicable |
| Integrations | Tekla, Strumis, Fabtrol, E.J.E., Excel | Tekla (detailing workflow) |
| Headline metric | 95–99% accuracy · 50–80% time savings | Not publicly quantified |
| Named results |
King Steel · 50% bid takeoff time reduction FabArc Steel · 91% time reduction per sheet Metals Fabrication · 40% more bids completed |
Not published |
04 · Where each has the edge
In Lift's favour
- Different stage of the workflow. Lift owns the estimating phase, ConnectionAI the detailing phase.
- Lift exports into the same detailing platforms ConnectionAI lives inside (Tekla, Strumis, Fabtrol).
- For estimators, Lift returns a BOM with weights and revision-aware quantities. ConnectionAI returns connections.
In ConnectionAI's favour
- Purpose-built domain model for steel, not a fine-tuned generalist.
- LOD 400 connection output, framed as a teammate to detailers.
- Clean positioning: human in the loop, no over-claiming.
- Active in the same NASCC orbit as Lift, with a collaborative tone rather than an adversarial one.
05 · Limitations on both sides
Lift's limitations
- Priced at the higher end of the category, and quoted rather than published, which makes it harder to weigh against a public number.
- The interface is not the fastest in head-to-head demos. Newer entrants like Ferra switch pages with less lag.
- Column workflows are still maturing. Some rivals ship a dedicated column scanner today.
- Feature pace is steady rather than flashy. Parametric assemblies and base-plate schedules are on the roadmap, not shipped.
ConnectionAI's limitations
- A detailing-stage tool. It isn't designed to produce a bid-ready BOM from a PDF set.
- Tonnage and beam-level attributes sit out of scope, as does revision management.
- Smaller customer footprint than the takeoff-focused vendors.
The take
ConnectionAI and Lift mostly don't compete. ConnectionAI automates LOD 400 connections for detailers. Lift produces the estimating takeoff. Plenty of shops run both. If your bottleneck is detailing throughput, ConnectionAI is the better tool. If it's getting bids out the door, Lift is. Treat this as a workflow question, not a head-to-head.