AI's impact on value pricing in law

Firms who are value pricing are already pioneering how most legal services will be delivered in the future, but we require smarter ways to price work and more reliable ways to manage scope. Rather than pontificate about value pricing, I'm going to share the recent progress we've made on this and offer some emerging concepts and technology that you can use to value price more effectively right now.?

Why value pricing, why now?

Value pricing is better for both clients and firms (I'll leave that sit as an axiom, but here is why I think this is true).

Value pricing isn't a new idea in law, but it's about to become a mainstream practice. This is because those who have adopted it up to now have been driven by opportunity rather than necessity. These early adopters were proactively motivated by the prospect of creating more innovative legal services, price-certainty for clients, and more profitable businesses. The next tranche of firms to transition to value pricing will be less idealistic and instead driven by the need to incorporate technology into their services. As the adoption curve marches on, the late majority and laggards will eventually follow the herd through sheer necessity to compete with those who are providing better services to clients more profitably.?

At adieu.ai we've already seen a trend emerging in which AI acts as a trigger for firms to switch to value pricing and I expect this to increase over the next few years. Using AI forces your hand because it emphatically breaks the time-based billing model. If you're only fiddling with AI around the edges you can turn a blind eye, but this stops being viable as you get more ambitious.

In short, you can't time-bill if you use AI. That leaves you with two options:

1. Value price everything (most of our firms do this)
2. Value price the AI-based work and time-bill the rest (some of our firms do this)

Either way, if you want to benefit from AI, you need to value price at least some of your services.

The challenge with value pricing

Being a good idea doesn't make value pricing easy however. Value pricing is a big leap and in my experience, no one gets it right first time. Family law firms are still relatively early in the value pricing journey and have had to feel their way forward by trial and error. The current methods and rudimentary tooling got us to where we are today, but won't take us where we need to go. Firms need smarter ways to price work and more reliable ways to manage scope. Less voodoo, more systems thinking.

I've experienced most of the challenges with value pricing first hand. One of the biggest challenges with building a social enterprise is that you need to find a viable business model - and fast - or it's game over. I've also seen it up-close with our firms across one and a half thousand matters.

The challenge with value pricing in family law at this moment in time is that the instruments are too blunt. Firms lack the tooling to properly measure value and wind up with models which at worst attempt to charge the same price for every matter, or at best just cater for rudimentary small/medium/large or simple/complex options. This is further exacerbated by a tendency for the pricing itself to simply be a de facto hourly rate multiplied by an anticipated number of hours - what academic and consultant John Chisholm refers to as "time billing in drag".

This is a natural part of the evolution of pricing in family law and these firms are still net ahead - but they have the opportunity to pull away from traditional firms as technology starts to play more of a role in the delivery of legal services. The exciting thing about good ideas that are early in their arc is the sheer unrealised potential they hold, meaning that even small improvements can result in massive gains.

Why is pricing so damn hard?

Here are some things I've either personally experienced in developing value pricing for adieu.ai, or observed in supporting our firms who are value pricing for their own clients.

It's not just about pricing

It's easy to focus on the 'pricing' part of value pricing and forget about the 'value' part. Value pricing is as much about value as it is about pricing. To get the best outcomes for your clients and your firm, you need to be moving both levers. There is a flavour of 90's motivational-cassette-tape value pricing which is little more than Jedi mind tricks to dupe clients into paying more than they should. Good value pricing requires an holistic approach; it's not about working out clever ways to up the price on the same old slop, it's about deeply understanding what your clients value, delivering that in spade loads, and charging a fair price for doing so.

You should first maximise value, and then work out how that value should be shared - and that is a much taller ask than merely picking some numbers.

It's highly emotional

Asking for money can be terrifying. You're really putting yourself out there, and the chance of rejection is palpable. You'll likely experience some emotional cocktail of hope, doubt, pride, guilt, terror, imposter syndrome, excitement, tall-poppy syndrome, relief, and everything in between. It's not just business, it's putting a number on your worth and the devil on your shoulder will remind you that handing the client a time-based bill after the fact would just be a whole lot easier.

It's not science (and it never will be)

To value price you have to quantify things that are inherently qualitative. No matter how scientific you try to make it, regardless of the number of data points and analytical models you may have, at some point you have to leap across a gap from an intuition to a number. You have to reduce a concept as fuzzy and ephemeral as 'value' down to a handful of digits.

Even things that appear methodical and data-driven are only predictions. Regardless of how big CoreLogic's dataset is, ultimately the value of a house is whatever someone is willing to pay for it, not what we predict they should. There is no silver bullet for this - the price can never be more than an approximation of something qualitative and you'll feel a lot better when you accept this.

The shadow of the hourly rate

I think one of the biggest barriers to value pricing is overcoming the sense that you're 'just making it up' while the seemingly safe and serene waters of the hourly rate beckon you home.

This couldn't be further from the truth. Ironically, almost everything that makes value pricing difficult applies equally to time-based billing. Setting an hourly rate is not fundamentally different to pricing an engagement, it just provides the illusion of being a smaller bet. This is an illusion because this seemingly small number is in fact a multiple, so every dollar in an hourly rate has a big impact on the ultimate price. In fact, your hourly rate was almost certainly just copied from the people around you (which was in turn copied from the people around them). This lemming-pricing is disconnected from any meaningful measure of value, constrains innovation, and broadcasts to the world that legal work is a commodity in which lawyers and firms are interchangeable with each other.?

Don't mistake the drag net of the billable hour for a safety net.

Where AI meets value pricing

I'm not a value pricing zealot. To be honest, we stumbled into value pricing after tripping over time-based billing. As a startup with a mission to enable separating families to reach resolution 10x sooner, we started out by building autonomous AI paralegals who could offer the massive leaps in speed and quality that are required to achieve this. Because work performed by AI happens orders of magnitude more quickly than the equivalent work done by humans, the only viable option for us and our firms was to value price this work. Building a model to value price work done by AI has been a multi-year research and development effort and was something we had to solve just so we (and our firms) could get paid. For us, actually doing the work is the main game, pricing it is just a necessary sideshow.

We didn't set out to build a value pricing tool; it emerged as a side effect of building autonomous AI paralegals that can define, scope and price legal work. It turns out that having a sophisticated model for pricing work done by an AI is really useful for determining what the price should be before the work is done. If you combine this model with the ability to predict scope based on historical matters, you suddenly have the ability to predict, scope, price and manage work in ways that simply weren't possible before. Value pricing in a box if you will.

When one of our firms first engages a client, their autonomous AI paralegal can predict and define the scope of the work based on historical matters, prepare a detailed scope document for the client, and value price the work based on a number of factors including firm profitability. Once the client accepts the fee proposal, the AI then goes ahead and performs the work as scoped, usually within a day or two.

This means that firms who are already value pricing can now provide fee proposals that are hyper-specific to each client, have much tighter control on their internal costs, and be more profitable. Firms who are not yet value pricing will now have an out-of-the-box way to start which has a set up time of less than 60 seconds and is more advanced than most firms who have been value pricing for years.

What we're playing with and what we've learned

To value price well, you have to define the scope correctly, accurately predict the value, and understand the costs. Here is what we've built for this step by step.

I ask my AI paralegal (Chris in this example, although each of our firms has their own AI paralegal) to prepare a value-priced fee proposal for my client by telling him what I know about the matter:

Chris then reviews the last thousand matters done through adieu.ai and predicts what the scope of work required for this matter is going to be.

Once he has determined a scope, he then prepares a detailed stage-based scope document for the work to be performed.?

Finally, Chris value prices the work for the client using our value pricing framework, ensuring that it achieves whatever target profit margin that the firm has defined.

Once the client has accepted the fee proposal and engaged the firm, Chris then goes ahead and performs the work defined, for exactly the price quoted. Every time.

Our framework for value pricing AI-based paraprofessional work is informed by real-time predictions of what the alternative cost would be for a human to do the work, as well as considering multiple value-axis, such as speed, convenience, quality and price-certainty.

The proposal and pricing provided by Chris currently only covers the AI-based work, and can be added to the other value priced work performed by the rest of the team to create a complete client fee proposal. It's conceivable that we could extend this approach to cover the entire matter if firms found that valuable - but whether we do that or not will depend on the feedback we get.

Managing scope: Micro-scope

Being able to accurately predict, scope and price work is immensely valuable and gets us much of the way to scalable value pricing - but that doesn't solve one of the other key challenges I mentioned above: managing scope.

Dealing with scope change is probably the biggest single barrier to value pricing, because even if you manage to get the scope right up front, the scope will change. And then it will change again. Family law matters are unpredictable and evolve quickly - that's the nature of the space we're in. Any model that we use to value price needs to be robust enough to handle this.

In traditional fixed price engagements (like the engineering ones I'm all too familiar with), the game is to define the scope within an inch of its life, and then ruthlessly manage that scope. When the scope looks like it is going to change, you draw up a variation order and negotiate a price for the variation. This model does NOT WORK in family law. Even if you had the ability to manage the scope like this (which you don't - these are full time roles in engineering projects), your client doesn't have a project manager on their side to negotiate each variation with you, and they sure as hell don't have the time, patience or expertise to do it themselves.

Another pragmatic but less-than-ideal model I've seen is to simply let it go - management by karma. Some matters will go over and you'll lose a little, some will come under and you'll be ahead, but it all evens out in the end. As far as I can tell, this is actually pretty common among those who value price. Although this might sound like horrendous business practice to some, it's not all that different from the traditional practice of writing down bills (which is not talked about a lot but is prevalent). In my view this simply isn't good enough.

Traditional fixed-fee scopes are too rigid to work in family law and management by karma is a wasteful response. This leads us to an alternative approach that we've been building out: micro-scoping. Micro-scoping involves deconstructing a matter into various modular components and creating a value price formula for each. Not just staged pricing, but a level lower, into the components that make up each stage, and then a level lower than that again, into the components that make up those components. These modules are then combined to create a model for each type of engagement (in our case a family law property matter), which provides the glue between a scope and key value axis to generate a value price.

This enables you to do four things:

1. Explicitly account for the factors that most represent value to your client and adjust them on a client-by-client basis
2. Move from a scope to a value price quickly, consistently and with confidence
3. Build multiple options and pathways into a value priced engagement - including repeatable components which have a price per cycle
4. Provide the price and definition of additional scope upfront, which can simply be activated as needed on a per-item basis with no need to re-scope or re-price anything later on

This moves us away from the one-number-for-everything, spray-and-pray approach and towards a model that requires minimal scope management and no surprises.

Where to from here

Although we've hammered this out from first principles, I'd be surprised if there weren't others out there doing similar things - this pattern makes intuitive sense and solves a real problem. I know our friend Clarissa Rayward 's 'Choose your own adventure' concept has repeatable elements and alternative pathways.

There are some interesting open questions to work though including:

• How do we ensure that this approach doesn't become too prescriptive?
• How granular can / should you go with scope?
• Will too much granularity lead to a reductive model which fails to account for value that can't be easily represented at lower levels?
• Are we using the right value-axis?
• Are we sufficiently capturing the uniqueness of each firm?
• To what degree will AI be able to manage scope day to day?

We're super-excited to work with our beta firms over the coming months to keep refining this model. If your firm is currently value pricing or looking to make the leap, register for our adieu.ai private beta and if we think you're a good fit we'll invite you to join (subject to capacity - Australian firms only). We're looking for early-adopters with the right mindset - tech-savviness is not a requirement.

I'd love to hear your thoughts and questions below. Make sure you follow me (or connect) on LinkedIn where I'll be sharing more of what's becoming possible with AI and value pricing, and what we're doing and learning at adieu.ai.