Transforming In-House Legal Service Delivery
We share some insights from our research, and our own experiences on transforming service delivery with technology and process.
Embedding new technology
Understanding the Challenges
Our research at the start of the year highlighted the challenges faced by in-house teams when it comes to making the most of new technology.
Understanding the Challenges
In January 2020, we published a report titled “The Future of In-House Legal,” which explored the challenges facing in-house teams and how they’re tackling them.
This report drew on 80 responses to an online survey, plus findings from over 30 in-depth interviews with:
- General counsels
- Heads of legal
- Other members of in-house teams.
By some distance, in-house teams felt they were least ready for the challenges they’ll face in the next five years when it came to:
- Technology (only 4% felt ready)
- Processes (only 6% felt ready).
Similarly, 38% of participants believed the use of technology by their in-house teams was the thing they most needed to change. And 19% of participants thought their in-house team’s processes were the thing that most needed to change. Technology and process were the first and second highest ranked overall.
COVID-19 has dramatically changed the way we all work. Many of the in-house lawyers we’ve spoken to over recent months have adopted new ways of working and new types of technology at a speed that wouldn’t have seemed possible back in January.
Our own experience was similar. In the space of a few days, we moved to a new way of working that would have otherwise taken months or years.
The initial focus was on home working, with the broader range of service delivery challenges remaining as before.
As the majority of us now work in geographically distributed teams, we have found that certain challenges such as team management have been magnified. We’re working hard to tackle these issues.
We’ve put together this whitepaper to share some insights from our research, and our own experiences on transforming service delivery with technology and process. We’d love to know if you find it helpful, so please don’t hesitate to get in touch.
Understand your drivers
Different in-house legal teams may have different drivers and priorities when it comes to transforming service delivery. Some may be chosen by the in-house team, and others may be part of wider organisational strategy. The first thing to agree is what your drivers are.
Consider how your priorities fit with your business’s wider goals. They can change, so it’s important to regularly assess your plans against your priorities.
What's driving your need to change?
- Reduce cost
- Shorten turnaround times
- Free up the legal team’s time for more strategic projects
- Improve quality of advice
- Make the in-house legal team more approachable
- Reduce the amount of time it takes the business to instruct legal
- Meet a growing range of business needs
- Be more agile in dealing with legislative/regulatory change
- Identify ways of managing new risks/ threats
- Something else.
Start with Processes –
Technology is not a magic bullet.
Start with Processes – Not Technology
Conversations around legal technology and service delivery tend to focus on what piece of technology the legal team should buy.
The hope is that the technology will dramatically solve a number of often undefined issues.
This isn’t helped by suppliers of legal technology promoting their products as magic bullets to solve lawyers’ problems. Unless you’re starting with a blank sheet of paper, technology should be seen as a way to improve an existing process, not a solution in its own right.
Our working lives are full of processes. Many are informal or undocumented, so we don’t recognise them as such.
For example, most teams have a filing system for documents, or an accepted way of responding to, and prioritising, enquiries from the business.
If you buy technology that doesn’t or can’t fit with your processes, then your users (both in the team and in the wider business) are likely to try to find workarounds, and may become irritated by technology that appears to make their life more difficult.
There are exceptions, such as where the technology dictates a process and it’s generally agreed that the dictated process is the optimal one. Even then, you’ll still need to think about how that process fits in with your wider processes, and how you can encourage users to adopt them.
The introduction of tech-led solutions without a proper analysis of the end-to-end process and consideration of the needs of end users can lead to an amplification of the problem that you’re trying to solve. It can also result in buying expensive products that are never used.
Operations Manager, Irwin Mitchell
Picking Your Improvement Candidates
Assess feasibility early on.
Picking Your Improvement Candidates
Some processes and technology will be easier to change than others.
Sometimes this will be obvious from the outset, but may need a little investigation as to whether they are feasible or not before you commit lots of time to re-engineering things.
- Is the legal team heavily dependent on another team, such as IT or sales, to successfully develop and implement the improvement? If so, does that team have the capacity?
- Who provides you with information, and who do you pass information on to? You need to understand these ‘data flows’
- Does the process or technology need to be followed or used by the business for it to be successful? Do you have the relationships and influence needed to get them to adopt the change? For example, will your sales team accept a different way of engaging the legal team?
Is the improvement likely to have sizeable cost implications outside your own budget? Do you have the ear of the board to get additional budget? Or can the cost be shared with your IT team or the team that’ll benefit?
Will organisational governance have any impact on what you can and can’t change? Will organisational governance make the required timescales infeasible?
Introducing concepts such as compliance by design can help manage legal risks, while taking the burden away from the legal team being the business’s police.
These might be hurdles you’re prepared to jump if the potential benefit is big enough, but it’s best to plan for them from the start.
Has your appetite to invest in legal tech in the next 12 months changed as a result of the COVID-19 pandemic?
- Greater appetite
- Less appetite
- No change
Has your organisation's appetite to invest in legal tech in the next 12 months changed as a result of the COVID-19 pandemic?
- Greater appetite
- Less appetite
- No change
Documenting the Process
How should you go about it?
Documenting The Process
Before you put pen to paper to document your new or existing process, there’s a few things to consider.
Who’s the target audience for the documentation of your process?
People familiar with the process may just need reminding once in a while. This audience would probably prefer something high-level that they can quickly cast an eye over.
For people not familiar with the process, such as your IT team, you’ll need to add more detail and context. Lawyers often forget that their ways of working aren’t necessarily obvious to people who’ve not interacted with or been embedded in legal functions.
Don’t jump to producing a process map. Consider what works best for your target audience. Is a contract negotiation playbook or checklist preferable? Could it be worthwhile to produce different versions for different audiences?
What different stakeholders are involved in the process?
Do some or all of them need to be involved? Do they need to be consulted when putting together the process to make sure it considers their pain points?
What makes life easy for one stakeholder can make it significantly more difficult for another. This is particularly important if one of your priorities is to improve the experience of others, such as your internal clients. For example, introducing self-help tools to cut down the workload of the legal team may increase the workload of the rest of the business, which may not be a good outcome if your main goal was to improve the experience of your internal clients.
Can you break the process down into bitesize chunks that could be implemented in their own right?
This approach means you put simple fixes in place quickly to achieve benefits much earlier. It also reduces the risk of the goalposts moving while you’re still documenting your process. Simple fixes done well can help.
Bear in mind that you might not have to begin at the start of the process. There might be benefits in addressing issues that arise in the middle or at the end first.
And don’t forget that sometimes a simple checklist could be all that you need to become more effective.
When documenting processes, it’s important to get the right people on board. You’ll need input from everyone who plays a role in the process, so the first step is to identify those people.
You may also want to consider representation from teams or functions that may not play a direct role in the process, but still interact with it. For example, your process may produce certain triggers for your finance department. Without the right people, you’ll make assumptions about how the process should run, which is almost always different to the reality.
You don’t necessarily need specialist software to map processes. Post-it notes, marker pens and a blank wall make for an interactive and engaging method of capturing process. In a world of remote working, there are some free virtual tools available online.
If you’re not an expert, keep the shapes simple: rectangles for tasks, circles for events, diamonds for decision points. There’s plenty of basic guidance available online on employing a consistent process modelling language. Business Process Model and Notation (BPMN) is my preferred option.
Process Improvement Manager, Irwin Mitchell
What Tech Should
Assess candidates carefully.
What Tech Should You Target?
We’d generally recommend you look at:
- Fine-tuning or developing processes
- Monitoring the quality of the data that your processes are producing. Downstream data quality will tell you whether process changes are actual improvements or not
- Testing them for a while to make sure they work
- Identifying which processes would work significantly better if augmented with technology, or where ‘eyeballs’ remain an important step
- Assessing what technology is available, how difficult it would be to put in place, and how much it would cost to buy and implement. Your IT colleagues may help you navigate the increasing number of ‘lawtech’ products available.
The first two points and the last are fairly self-explanatory. But how do you identify which processes will work significantly better augmented with technology?
Here’s a few things to think about:
- You’ll often get a better return on investment for processes, or part-processes, which are relatively simple but are repeated often or take up a large amount of time. When testing your processes, consider collecting some data, such as how long it takes a person to execute that part of the process.
- If one of your objectives is to improve reporting on key performance indicators (KPIs), which processes could generate useful data if digitised?
- Are there particular parts of a process that don’t take long in terms of resource, but often cause delay? For example, does your contract governance process require signature from your CFO, who is only in the office twice a week?
- Which of your processes involve analysing lots of data and would benefit by using analytic tools?
For legal teams, the simple and time-consuming processes are often at the beginning or end of a matter lifecycle, including:
- Taking instructions
- Producing initial drafts of documents
- Executing and storing a contract.
You might find a few opportunities further into the ‘delivery’ part of your processes too. If you had better playbooks for negotiation of a particular type of contract that regularly comes across your desk, could a paralegal do the work instead of a senior solicitor?
What tech do you already have?
Don’t forget about technology that you already have in your business when assessing what’s available.
We wouldn’t recommend shoehorning a process into a technology that was designed for something else, as this can lead to problems further down the line. But your business may already have something that could easily be used to improve a process.
Perhaps the most common are those included in Office 365 enterprise plans, such as:
- Microsoft Planner and To-Do for task management
- Microsoft Power Apps, Power Automate, Power Virtual Agents and AI Builder for automation/digitisation of simple processes
- Microsoft SharePoint Online for instruction portals, self-help guidance, templates, precedents, organisational and business knowledge, collaboration and contract/document management.
You may need some initial help from your IT team to get you set up, but once you are a lot of these tools are relatively easy to configure. This makes you less reliant on other teams, or your suppliers, to get your change through.
Never believe the sales hype – nothing is a silver bullet. Trying to cut corners or increase pace by not really understanding the problem you’re trying to solve first never ends well. Even agile delivery and fast iterative working to clarify ambiguous requirements still requires you to know where you’re going in the first place.
Chief Technology Officer, Irwin Mitchell
Implementing Your Solution – Who Should Be on the Team?
Collaboration with other teams and frequent involvement of subject matter experts can make your project a success.
Implementing Your Solution – Who Should Be on the Team?
Don’t be tempted to ‘drop requirements’ with another team, and leave them to interpret them.
The best results are typically achieved through collaboration, with frequent involvement of the necessary subject matter experts.
This doesn’t mean everyone needs to be on the project full-time. Initially, you should try and get everyone together to workshop through the problem. Then regularly keep in touch with subject matter experts to clarify any points the core team’s unsure about, and validate work that’s been done. This means that you can quickly correct it if it’s going wrong.
It’s important for everyone to come to this process with an open mind. While you might have a good idea of a possible solution, present the problem to the group first, and listen to the ideas they come up with.
More often than not, the best solutions are a mash-up of ideas. This works particularly well when those involved come from different industries or backgrounds, as they can really challenge the status quo.
The specific skills you need will vary depending on what kind of solution you’re putting together, and what your priorities are. But don’t rule out the need for a particular skill too quickly. For instance, you might assume that someone with a background in data or user experience can only bring something to a technology solution.
They could also help you figure out ways of generating useable data from a manual process, or make your contract negotiation playbook easier and quicker to use.
Finally, don’t forget about your end users, whether they’re an end user of technology or a manual process.
This might be an internal client of the legal team, or the type of user that’s going to use your solution. It could be the sales person who will navigate your contract lifecycle management solution, the procurement professional who will use your self-help toolkit, or the paralegal who’ll use your playbook.
Test, learn, and iterate
It’s unlikely you’ll get your processes or your technology right first time. It can be tempting to move onto something else very quickly, but it’s worth spending time to consider whether there’s anything else you can do to create marginal improvements, as these can quickly stack up.
There are a few continuous improvement methodologies out there – Lean Six Sigma (DMAIC), Plan Do Check Act (PDCA) etc – but they generally follow similar principles.
With a little training, these methodologies can be adopted by any team, enabling them to make their own small, incremental improvements. But the key is in the culture. Continuous improvement has to become part of the day job if it’s to be successful.
Process Improvement Manager, Irwin Mitchell
Get the grips with the tech terms you need to know.
Many in-house counsel contacts told us there wasn’t enough easily digestible information available on legal tech options and what they can be used for.
We’ve put together a short glossary of the less self-explanatory legal tech terms, and how different types of technology might be used in a legal context.
Artificial Intelligence (AI)
AI is an umbrella term used to describe machine learning, natural language processing, and/or expert systems. Any given AI product probably uses a combination of these.
Our AI glossary provides more detail on what those terms mean, and what other AI terms mean.
AI in legal is predominantly used to review large volumes of data, such as single or multiple contracts, and either do something with the output, like reporting on it, or make decisions based on the output, like do or don’t sign this contract.
For some more common use cases, see Document Review and Litigation Analytics.
The challenge with AI is typically the amount of training required, where the system essentially learns from a manual human review. This can mean you need some volume to make the exercise justifiable, although AI products increasingly have pre-trained models you can use.
To a degree, this is what it says on the tin: extremely large sets of data.
The term is typically used to talk about being able to do things with that big set of data. It used to be extremely difficult to do anything useful with big sets of unstructured data. But with improvements in AI and computing hardware (e.g. more power for less money), it’s becoming possible to get insights from those kinds of data sets.
Blockchain is the fundamental technology underlying things like bitcoin. At its most basic level, it’s a method of storing information. But it attracts attention as it typically means:
- Data is distributed – every participant (or “node”) has a copy of the whole dataset
- It’s decentralised – no one person has control over the data, in theory making it “trustless” (you don’t need to trust anyone to know the data is valid)
- Each bit of data added to it is immutable i.e. it can’t be reversed or amended.
How it achieves this is beyond the scope of this glossary. The debate around blockchain is typically on whether or not it’s necessary for a particular use case. Blockchain may be over-engineering, and a normal database of some form may be more appropriate, unless you:
- Need to share data
- Have multiple people contribute to the data
- Don’t need data to be amendable
- Don’t have a central body or set of bodies that can be trusted to be a custodian or source of truth
- Need the data to be tamperproof/highly auditable.
The most common use cases talked about in legal are for registries of one form or another, such as Land Registry, and smart contracts.
Technology that can extract data from a bank of documents, such as renewal dates, which you might then import into your contract management system.
Content extraction software uses several techniques – some AI, and quite often some old fashioned “regex,” which looks for specific patterns of characters.
Contract lifecycle management
Contract lifecycle management software is in essence a contract database for your business, combined with:
- Configurable reporting and analysis
- Powerful search
- Document automation, including negotiation playbooks
- Document approval processes
- Document e-signature.
Contract review automation
Typically the use of AI, and all natural language processing, machine learning and expert systems, is to review contracts for key risks, and report on them.
At the time of writing, this is predominantly used as either:
- An initial sift or where otherwise the contract wouldn’t be looked at all
- To facilitate a guided review by a human, such as a junior lawyer using a playbook, but being pointed to the right part of the agreement.
At its most basic level, document automation means to generate a pre-populated template document based on the answers to a questionnaire.
But document automation software is now going beyond this. For instance, often allowing you to:
- Produce an initial draft
- Negotiate that draft, and replace clauses with pre-defined alternatives and/or raising escalations where approvals are needed to change something (e.g. a liability cap)
- Track negotiations against the original version.
Document review is a catch-all term for things like content extraction, contract review or due diligence automation.
Due diligence automation
This is technology that fully or semi-automates due diligence on a large volume of documents.
For example, it might use content extraction to identify which or how many customer contracts have:
- A change of control clause in them
- A restriction on assignment in them.
It would then provide a report on the results.
Typically used in litigation, but also potentially useful for data subject access requests (DSARs) and other compliance activities, e-discovery automation software typically offers a variety of features to make the job of sorting through information easier, for instance:
- Automatically removing duplication, such as email threads that have gone off in multiple directions
- Using AI to perform an initial sift for documents which are potentially disclosable, either in litigation, or to satisfy a DSAR
- Showing you information in different ways, like turning multiple email chains into timeline view.
Robotic process automation (RPA)
Robotic process automation software allows you to turn a manual “human” process carried out on a computer into an automated one.
This involves teaching a software robot to carry out a task a human would ordinarily complete using a computer, and then scheduling it to carry out that task at certain times or on certain triggers.
The robot typically uses the same interface as a human. It will open programs, copy and paste, and so on. It’s most often used when it’s not possible to carry out that process programmatically – or in other words where there’s no means for two computer programmes to talk to each other directly.
You might use it to synchronise documents between two systems or to run input data into a report package every night so there’s an updated report ready for the morning.
Smart contracts are self-executing contracts, written in code or quasi code and typically stored using blockchain. The smart contract doesn’t in itself constitute a contract. It’s better thought of as the means through which a contract is automatically performed.
While the term smart contract is usually associated with the use of blockchain, a more simple example might be a fully automated e-commerce system. The vendor and consumer agree to sell or buy a product, and then payment is automatically taken and goods are automatically shipped.
Despite being commonly associated with blockchain, there's much debate around whether blockchain is necessarily needed in all smart contract use cases.
An expert system is a type of AI. It can make decisions based on data input into it. It’s typically based on the logic a human has created (think of it as a series of “if this, do that’s”).
If a legal AI product makes some form of assessment, such as recommending a course of action or flagging the degree of risk, expert systems are likely involved.
This is simply a technique used in AI products to recognise patterns. It’s the fundamental building block of how AI products learn from the people training it (active learning), and/or those using it (passive learning).
This is the end result of training an AI product. Think of it as a set of criteria the system applies against the documents you feed into it, based on what you taught it.
Natural language processing
Natural language processing is a type of AI, which essentially turns words into meaning.
So instead of only looking for the word “dog,” because that’s what the text you used to train the product, it might also look for “hound,” “cocker spaniel,” “golden retriever,” or “poodle.”