RedlineAdvisory Group

Revenue Cycle · Coding · CDI · Quality · Audit · Risk Adjustment

27 years selling you the AI. Now I'm on your side of the table.

Independent evaluation, validation and negotiation support for hospitals and health systems buying AI for medical coding, clinical documentation integrity, quality, auditing and risk adjustment. From someone who has been in the vendor's chair since 1999 and knows exactly what you should know beyond the RFP and demo.

Pick your seat ↓ Start a conversation
Exhibit A · vendor proposalIllustrative

Up to 85% autonomous coding rate*

On which patient types? Your charts or theirs? *By whose definition?

ROI guaranteed in 12 months

Guaranteed how? Find the remedy. There isn't one.

97% coding accuracy

Precision or recall? Against what gold standard?

Not a real vendor's proposal. Every real one I've seen deserves the same pen.

Who's in the room?

Pick your seat. Each one sees a different deal. Here's what matters from yours, and what the vendor hopes you won't ask.

For the CFO

You're being sold a business case. Make it one you can bank.

Every vendor brings an ROI model, built on their assumptions, their definitions and their best customer. I built those models for a living. I'll rebuild yours on your data, then make the vendor sign up to it.

Coding & autonomous coding

What they won't volunteer. Autonomous rates are quoted on the vendor's definition and their best-fit chart mix. Two vendors quoting the same percentage are often measuring different things.

Clinical documentation integrity

What they won't volunteer. Before-and-after CMI comparisons rarely control for service-line changes, new physicians or seasonality. The lift you're shown may not be the lift you get.

Quality: HACs & PSIs

What they won't volunteer. The HAC Reduction Program cuts Medicare payments by 1% for the worst-performing quartile of hospitals. Some of what drives your score is documentation and present-on-admission coding, and that's where software claims get vague.

Auditing

What they won't volunteer. Audit tools are sold on how much they find. What matters is how much of it is real and recoverable.

Risk adjustment (HCC)

What they won't volunteer. Risk adjustment ROI is usually modeled on HCC capture alone. Unsupported codes found in a Risk Adjustment Data Validation (RADV) audit can be extrapolated across a whole contract, and that downside rarely shows up in the model.

Negotiate for outcomes, not price

  • Fees tied to measured outcomes, with the measurement method written into the contract
  • Performance guarantees with real remedies instead of best efforts
  • Caps on annual escalators, and per-chart pricing that adjusts when volume moves
  • Exit terms that don't hold your data or your workflow hostage

For the VP of Revenue Cycle

Your KPIs are the scoreboard. Don't let the vendor keep score.

DNFB, coder productivity, backlog, denials, turnover. Every vendor promises to move them. Few will commit to how they'll be measured or what your team gives up to get there. I've sat in those implementation meetings on the vendor side. I know where the hours go.

Coding & autonomous coding

What they won't volunteer. When the easy charts go autonomous, the charts left for your coders are harder. Per-coder productivity can drop even as total throughput rises. Staff for that.

Clinical documentation integrity

What they won't volunteer. Case prioritization drives a large share of CDI results, and it's usually the least demonstrated part of the demo.

Quality: HACs & PSIs

What they won't volunteer. Retrospective quality reporting tells you what went wrong. Concurrent review is where documentation can still be clarified.

Auditing

What they won't volunteer. Denial prevention results are usually drawn from customers whose payer mix looks nothing like yours.

Risk adjustment (HCC)

What they won't volunteer. Retrospective chart review finds HCCs after the fact. Capture before and during the visit is harder to implement, and it's usually where lasting results come from.

Negotiate for outcomes, not price

  • KPI baselines agreed before signature, measured the same way after go-live
  • Vendor implementation resources committed in writing
  • Remedies when go-live dates slip on the vendor's side
  • Pricing that doesn't penalize you for growing volume

For the HIM Director

Accuracy against what? That question decides everything.

You own coding quality and compliance, and you'll own the result long after the sales team moves on. Vendors will quote accuracy numbers. I'll help you find out what they measured, against whose gold standard and on which charts, then validate it with your team on live data.

Coding & autonomous coding

What they won't volunteer. A model can be very accurate on the codes it assigns and still miss the codes it never suggests. Precision without recall is half the story.

Clinical documentation integrity

What they won't volunteer. Aggressive query logic can raise capture rates and audit exposure at the same time.

Quality: HACs & PSIs

What they won't volunteer. A flag your team has to disprove is not free. False positives are a labor cost the demo never shows.

Auditing

What they won't volunteer. Some vendors sell the coding AI and the audit tool that checks it. Ask who is checking whom.

Risk adjustment (HCC)

What they won't volunteer. A tool that only finds additions creates one-way risk. Auditors look for unsupported codes, and an add-only program is hard to defend.

Negotiate for outcomes, not price

  • Accuracy definitions, sample sizes and audit methods written into the contract
  • Notice and validation rights before model changes reach production
  • Clear accountability when autonomous codes are wrong
  • Coder training and support commitments that outlast go-live

For the CIO

You're not buying software. You're buying a model that changes after you sign.

AI in the revenue cycle touches your EHR, your PHI and your security posture, and the model inside it is retrained on a schedule you don't control. I'll help you put the right technical and governance questions in front of vendors, and get the answers into the contract.

Coding & autonomous coding

What they won't volunteer. Many products now combine AI/NLP/CLI, machine learning and third-party language models. The architecture slide rarely shows every subcontractor.

Clinical documentation integrity

What they won't volunteer. Concurrent CDI runs on near-real-time data. Prioritization algorithms quietly turn it into retrospective review for missed opportunities.

Quality: HACs & PSIs

What they won't volunteer. Quality, CDI and coding tools often compute the same PSI logic three different ways, and you end up paying for all three.

Auditing

What they won't volunteer. If you can't explain a code to an auditor, the risk is yours, not the model's.

Risk adjustment (HCC)

What they won't volunteer. Suspecting conditions from outside data is powerful. It's also exactly the model behavior you'll have to explain if a payer or regulator asks.

Negotiate for outcomes, not price

  • Limits on whether and how the vendor trains on your data
  • Model change notification and rollback rights
  • Uptime SLAs, security attestations and data return on exit
  • Disclosure of every third-party model and subprocessor

The other side of the table

Most revenue cycle consultants came up on the provider side. They know how buyers think. I spent 27 years learning how vendors think, which is what you're negotiating against.

Typical provider-side consultantRedline · Tim Minnich
ProposalsReads themHas written them: RFP, RFI and RFQ responses, business cases and ROI models
DemosScores what was shownHas built and delivered thousands of them and knows what the provider should be focusing on
The AIEvaluates features and workflowHas explained how the AI, natural language processing (NLP) and clinical language intelligence (CLI) under those features reads a chart since 1999
QuestionsKnows what buyers usually askKnows what vendors hope you won't ask
NegotiationPushes on priceKnows where vendors actually have room: guarantees, scope, terms and timing
The marketSees the vendors their clients boughtHas competed against them in deal after deal
Tim Minnich, founder of Redline Advisory Group
Tim Minnich · Founder & Principal

The record

I sold computer-assisted coding before the category had a name, and I've sold it longer than anyone in the industry. Here's what that looks like.

27 yrs

Selling AI/NLP/CLI for IP/OP/Professional coding and clinical documentation integrity since 1999, longer than anyone else in the industry.

41 yrs

On the vendor side of healthcare information systems, starting as a programmer, transitioning into sales in 1993.

70+

Hospital clinical lab system implementations managed before I ever sold software

1985
Programmer, then implementations manager · Shared Medical SystemsBuilt and installed hospital laboratory systems on site. Later ran the team behind 70+ implementations. In 1993, moved into sales of clinical information systems.
1999
A-Life Medical · physician marketStarted selling AI/NLP/CLI that reads clinical narrative and assigns codes, 10 years before "AI in healthcare" was a budget line. Personally sold more than 80% of company business in my first six years.
2008
Computer-assisted coding comes to hospitalsFirst salesperson bringing CAC to hospitals and health networks. Built a substantial hospital pipeline in six months and the go-to-market from nothing, through the national launch at AHIMA.
2010
Optum acquires A-Life HospitalEnterprise CAC, CDI and professional coding inside a Fortune 5 organization. Moved into a team selling role supporting the entire sales force.
2021
Senior Solutions Consultant · mid-revenue cycleSolution lead on complex health system pursuits: RFP responses, executive demos and the business cases CFOs are asked to approve. AHIMA regional conference speaker on AI-enabled coding and CDI.
Now
IndependentSame knowledge. Different client. You.

What I'm not

I'm not a clinician, and I won't pretend to be. Your CDI specialists and coders know the chart better than I ever will.

I know the machine, the vendor and the deal. I'll tell your experts what to measure during live-data validation and why it matters.

Put us in the same room and the vendor has nowhere to hide.

The method

Most evaluations start with an RFP and a vendor's demo and end with a fight about price.

A disciplined one starts with your money and ends with a contract that holds the vendor and the provider mutually accountable for results.

  1. Frame and fund

    Define the problem, baseline your metrics and get the business case approved before any vendor sets the agenda.

  2. Remove the bias

    Write requirements around outcomes instead of a favorite vendor's feature list. Build the shortlist on evidence.

  3. Rethink Demo Requirements

    Define and demo best practices, not a replication of current workflows.

  4. Validate With Live Data

    Observe encounters and workflow in a live environment. Your team validates: precision, recall, DRG impact and autonomous rate by encounter type and your definition.

  5. Negotiate for outcomes

    Pricing tied to measured results, guarantees with remedies, data rights, model-change control and real exit terms.

  6. Hold them to it

    Measure after go-live against the same baseline, so you walk into the renewal with proof instead of anecdotes.

Engagements

Three ways in, depending on where you are in the cycle.

Start to signature

Full evaluation

The whole method, run with your team, for a new coding, autonomous coding, CDI, quality, audit or risk adjustment decision.

  • Business case and budget approval
  • Outcome-based requirements and shortlist
  • Best-practice demo requirements and live-data validation
  • Scorecard, selection and negotiation
For: a new purchase or a full replacement
Before you re-sign

Renewal leverage

Your renewal is the most leverage you'll have with this vendor until the next one. Use it.

  • Contract teardown
  • Performance against what was promised
  • Market alternatives, priced and credible
  • Strategy and support at the table
For: renewals 6 to 12 months out
A second opinion

Claim check

A focused review of one vendor's proposal, ROI model or accuracy claim before you commit to it.

  • What the numbers actually measure
  • What's missing from the contract
  • The questions to send back
For: a decision already on the table

Yes, I charge a premium. Here's the math.

An AI coding or CDI contract is a multi-year commitment that touches every inpatient and outpatient claim you send. One loosely defined metric, one uncapped escalator or one missing exit clause can cost more than my entire engagement.

You're paying for 27 years of knowing where the room is in the deal. I take no commissions, referral fees or resale margins from any vendor.

Before you sign. Call.

Thirty minutes. Tell me where you are in the evaluation process and I'll tell you what my first steps would be to support and drive an effective buy cycle.

Phone
610-587-7366
Email
Tim.Minnich@redlineadvisory.health
Web
RedLineAdvisory.health
Based in
Mohnton, Pennsylvania
Working nationwide