I recently had the opportunity to join a panel at the Intapp DealCloud Ambassador User Summit alongside fellow DealCloud delivery partners. With an audience of DealCloud firms in the room, the conversation got into the details that matter most on the implementation side: what separates successful deployments from ones that stall, where firms consistently leave value on the table, and how the best partners think about setting clients up for long-term success.
Below are the themes I kept coming back to.
Every Firm Thinks Their Process Is Unique. Most of Them Are Right at the Wrong Level.
This comes up in nearly every engagement, and it is worth unpacking. At the data architecture level, private capital firms are more alike than they tend to believe. The objects, their relationships, and how they broadly function are largely consistent across the industry. Where real uniqueness emerges is at the datapoint level.
- How is a firm scoring deals?
- What values are they attributing to them?
- How do those values feed into their calculations and reporting?
The more important question we push ourselves to ask is whether customization ties to a measurable business outcome. Adding a field or adjusting a dropdown list is trivial. But building a custom stage workflow deserves a harder look. Does it help the firm understand deal velocity? Does it surface which relationships are actually converting?
If those questions do not have clear answers, the customization is decoration, not infrastructure.
Perfect Data or Perfect Workflow on Day 1: The Answer Is Always Workflow.
A firm will never have perfect data on Day 1. A great workflow improves data quality every day after launch. Perfect data sitting behind a clunky workflow is a museum exhibit: beautiful, static, and ignored.
The impulse toward perfect data before go-live is usually a stalling mechanism. We have seen firms spend 18 months on data migration before launch and lose all momentum in the process. By the time they went live, the data they had cleaned was stale and the internal champions had moved on.
There is a minimum viable data layer that matters:
- The active deal pipeline
- The top 200 LP relationships
- Live mandates
Those need to be trustworthy on Day 1. The long tail of historical contacts and closed deals can be cleaned in flight. A useful rule of thumb is to launch with workflow at 95 percent and data at 70 percent. The workflow gets data to 90 percent within six months. The reverse is not true.
One more thing worth noting: do not get stuck on what can be migrated. Ask whether it should be. Moving six figures of documents into a new system is sometimes the right call. Often it is not.
The Single Biggest Predictor of Implementation Success Is Not What You Think.
It is not the quality of the data going in. It is not the sophistication of the workflows. It is whether there is a named, empowered business owner with executive support behind the project. Not the IT person. Someone who can say no to scope creep and yes to phase two.
Beyond that, the two most common places firms stumble before implementation even begins are:
- Skipping the “what does success look like” conversation. If a firm cannot define what success looks like at 90, 180, and 365 days, they will never feel like they got there. This conversation needs to happen before the first configuration decision is made, and it is the implementation partner’s job to lead it.
- Treating the data audit as a formality. A real data audit is not “we have a CRM, and the data is mostly fine.” It is sitting down and looking at what the top 200 LP records actually look like in the source system. That conversation reframes everything. We regularly find 50 to 100 fields that are not being used at all, and the question worth asking is what is keeping the team from using something that could genuinely help them?
Training Tells People What Buttons to Press. Workflow Design Tells Them Why They Would Want To.
Long-term adoption is where most implementations either prove themselves or quietly fall apart. A few approaches that consistently work:
- Measure adoption like a product manager. Track logins, record updates, and activity capture rates. Address gaps directly rather than scheduling another all-hands training session.
- Build a champions network. Identify one or two power users per team. They absorb the majority of everyday questions and serve as the voice of the user back to whoever is managing the platform.
- Make DealCloud the path of least resistance for things people already need. If the Monday pipeline meeting runs off DealCloud reports, the pipeline gets updated. If LP commitment dashboards live in DealCloud, the IR team keeps them current. Reporting drives behavior far more reliably than training does.
The hardest user to win is the senior partner who has been doing deals for 25 years and has an analyst maintaining their CRM. When leadership delegates their own usage, the system becomes a reporting tool rather than a working tool, and the data degrades within a quarter. Senior people set the tone, and their activity in the system matters just as much as anyone else’s.
This is also where the one-time deployment mindset does the most damage. I like to think of a CRM as a tailored suit. When you pick it up, it fits perfectly. But a year later, your organization has grown, your team has changed, and the deals you are chasing look different from what they did at go-live. The suit needs to go back to the tailor. And just like a suit, the longer you wait to make adjustments, the more obvious it becomes that something does not quite fit. The firms that treat DealCloud as a one-time deployment tend to be the ones calling us two years later, wondering why adoption has slipped and why the system no longer reflects how the business actually operates. The tailoring never really stops.
Getting Your Data Structure Ready for AI
This may be the most forward-looking part of the conversation, and it is one our team thinks about constantly.
Three data hygiene priorities that pay dividends when AI enters the picture:
- Standardize your taxonomies across sectors, deal stages, and relationship strength so that models can generalize across your data.
- Ruthlessly deduplicate contacts and entities.
- Capture activity with rich tagging, because relationship intelligence runs on who is talking to whom about what, not on static fields.
Resist the urge to custom-field everything. Custom fields are where AI models go to struggle. They do not generalize, they fragment your taxonomy, and every one of them is a small tax on future optionality.
At Monarch, we have been experimenting with practical automation that removes the steps teams find most painful. We have built workflows that allow a document to be forwarded to an internal address, use AI to parse it into DealCloud fields, and create companies, contacts, and deals with a couple of approval clicks. What used to take ten minutes of manual data entry is now two clicks. If a team is entering the same information repeatedly from a computer-readable source, there is a strong chance that the process can be automated with the right guardrails in place.
The firms that invest in data quality now are the ones who will be able to put AI to work meaningfully. The good news is that a lot of the most painful data capture steps can be removed entirely in the process of getting there.
It was a genuinely valuable panel, and the questions from the room reflected how seriously this community takes the work of getting CRM right. If any of this resonates with where your firm is right now, I am always glad to continue the conversation.
